Porti & ambiente — le notizie raccolte

The European Commission is eyeing a massive expansion of its emissions trading system to cover all departing flights, a move that could reshape airline economics and carbon markets alike. Share The European Union is preparing to do something it tried once before, and got its hand slapped for. The European Commission plans to extend its Emissions Trading System to cover flights departing EU airports for international destinations, not just the intra-European routes currently subject to carbon pricing. Right now, the EU ETS only applies to flights within the European Economic Area. Think Paris to Berlin, not Paris to New York. The proposed expansion would capture emissions from all outgoing flights, which account for a significant majority of aviation emissions tied to EU airports. A formal assessment is scheduled for July 2026, at which point the Commission will evaluate whether CORSIA, the International Civil Aviation Organization’s carbon offsetting scheme, is doing enough to justify keeping international flights outside the EU’s pricing regime. Legislative proposals could follow, with expanded coverage potentially taking effect by 2027. The EU has already been tightening the screws on aviation emissions through its Fit for 55 climate package. Free carbon allowances for airlines have been shrinking: a 25% reduction in 2024, 50% in 2025, with full auctioning expected by 2026. The scale of what’s at stake is significant. Flights from European airports emitted approximately 195 million tons of CO2 in 2025, surpassing pre-pandemic levels from 2019. The ETS expansion could bring an additional roughly 107 million tons of CO2 annually under the pricing umbrella, generating substantial revenue at current allowance prices ranging between 70 and 100 euros per tonne. In 2012, the Commission attempted to include all international flights in the ETS. The response was swift and hostile. China threatened to block Airbus orders. The US passed legislation prohibiting its airlines from complying. India and Russia joined the opposition chorus. The international flight provisions were suspended, and the ETS was scaled back to cover only intra-EEA routes. The current proposal would likely exempt incoming flights while capturing emissions from all departing ones. Airline executives are already signaling their displeasure, with the core concern being that carbon pricing increases operational costs, which get passed to passengers through higher ticket prices. Environmental advocates, including organizations like Transport & Environment, argue that comprehensive emissions pricing is not just appropriate but overdue, and call for a full expansion of the ETS to all departing flights by 2027 to align with the polluter-pays principle. The revenue stream it creates is essential for financing the transition to sustainable aviation fuels and other clean energy technologies. For EU-based airlines, full auctioning of allowances by 2026 means emissions become a direct line-item cost. If the ETS then expands to cover international departures, the cost base grows further. Higher ticket prices could dampen leisure travel, which tends to be more price-sensitive than business travel. Cargo shipping rates could also be affected. For carbon market participants, adding roughly 107 million tons of annual emissions to the ETS pool means significantly more demand for carbon allowances. If supply doesn’t expand proportionally, allowance prices could push toward the upper end of the 70-100 euro per tonne range or beyond. The geopolitical risk remains the outstanding variable from 2012: unilateral climate measures affecting international aviation invited retaliation from China, the US, India, and Russia. The EU is betting that CORSIA’s limitations are now more widely acknowledged and that the pushback will be more manageable this time.

Jefe de sección Actualizado el 09/06/2026 21:23 CEST El Barça perdió en su ‘destierro’ de Blanes (1-2) Y puso fin a una temporada decepcionante con su eliminación en semifinales de la OK Liga. Será el Igualada Rigat el que jugará a partir del viernes la final ante el Liceo. Gane o pierda, firmará su mejor clasificación desde 2004 y cerrará una campaña de éxito absoluto. Todo lo contrario para el Barça, que ha firmado una campaña muy por debajo de lo esperado. El campeón de 11 de las 12 últimas Ligas, entre ellas las tres últimas, se va de vacaciones con un título menor, la Supercopa, sin pisar la final de la Liga ni la Copa del Rey, y con la derrota en la final de la Champions ante el Porto como único éxito relativo. Muy poco para un equipo que el año pasado despidió a su anterior técnico, David Càceres, tras ganar la Liga. El quinto partido de la semifinal se alejó del Palau Blaugrana por la coincidencia con el playoff de la Liga de basket y el Barça dio la impresión de jugar como visitante ante una colonia más numerosa y ruidosa de seguidores del Igualada. Pero culpar al ambiente por la eliminación blaugrana sería faltar a la realidad. El Barça no estará en la final porque dejó escapar un partido en el Palau, no remató la serie después de lograr una victoria en el terer partido en Les Comes, y porque a lo largo de la serie no encontró soluciones ofensivas más allá de la bola parada ante la roca que es este Igualada, posiblemente la mejor defensa de la Liga y el resultado de un excelente trabajo en la pista y en los despachos de un club modesto pero lleno de orgullo. En el Barça toca reflexión y seguramente cambios importantes en una plantilla que ayer dijo adiós a Sergi Fernández y a Pablo Álvarez, el máximo goleador histórico de la Liga, que hubiera preferido una despedida mejor del hockey patines. De poco le sirvió al Barça salir con intensidad máxima en una primera parte de ritmo asfixiante. Llegó más que el Igualada, pero falló siempre sus mejores ocasiones ante el meta Arnau Martínez, uno de los grandes protagonistas de esta serie. Y el Igualada, con menos presencia en el área rival, también tuvo sus opciones y supo esperar hasta que a un minuto del descanso Cañadillas remató a gol tras una incursión por la derecha del área que sorprendió a Sergi Fernández, hasta entonces perfecto. El Barça tampoco acertó al inicio de la segunda mitad con la falta directa por la 10ª falta del Igualada. Alabart se enredó con la bola ante Marquès, el meta suplente que salió para detener el lanzamiento. Con aún más pressing defensivo, los jugadores de Ricard Ares consiguieron reducir las posesiones del Igualada, pero llegó otro error fatal: tarjeta azul a Ferran Font, y en la consiguiente inferioridad numérica, Llanes marcó el 0-2 con un zambombazo. Barça: Sergi Fernández, Barroso, Font, Alabart (1), Marc Grau -inicial-, Eloi Cervera, Llorca, Aragonès y Pablo Álvarez​Igualada Rigat: Arnau Martínez, Carol, Ruano, Pascual, Marc González -inicial-, Llanes (1), Cardil, Cañadillas (1), Bars y Marquès (p.s.)​Goles: 0-1 Cañadillas 24'; 0-2 Llanes 42'; 1-2 Alabart (f.d.) 47'​Árbitros: Miguel Días e Iván González. Tareta azul a Ferran Font (2). Faltas: Barça 13, Igualada 15​Incidencias: 1.500 espectadores en la Ciutat Esportiva de Blanes, com amplia presencia de aficionados del Igualada Las prisas se adueñaron del Barça, que al menos rompió el muro con una falta directa de Alabart a tres minutos del final para el 1-2. Y hasta allí llegó el Barça. Será el Igualada el que luche por su sexta Liga ante el Liceo, 29 años después de la última.

Anthropic released its latest model Tuesday afternoon, heralding the public’s first access to the AI company’s most powerful class of AI systems. Subscribe to read this story ad-free Get unlimited access to ad-free articles and exclusive content. The company says the model, termed Fable 5, is the first publicly available product in the same family as Anthropic’s powerful Mythos models, which sentshock waves through the cybersecurity worldearlier this year for their superhuman ability to find and exploit cyber vulnerabilities. “Fable’s capabilities exceed those of any model we’ve ever made generally available,” the company said in a blog post announcing the model’s release. “It is state-of-the-art on nearly all tested benchmarks of AI capability, showing exceptional performance in software engineering, knowledge work, vision, scientific research, and many other areas.” According to the company, Fable 5 uses the same tier of technology as Mythos but is safe for use by the public because of safeguards and limits placed on the technology. Anthropic’s Mythos Preview model, which lacks similar safeguards, was able to find thousands of critical and severe cyber vulnerabilities, including bugs and exploits in all major operating systems and web browsers. Many AI researchers worry that increasingly powerful AI systems could help bad actors carry out cyberattacks on banks, power grids or other critical infrastructure. Others hypothesize that increasingly intelligent AI systems could help terrorists design and deploy bioweapons. “The same queries that are beneficial in the hands of cybersecurity professionals and biology researchers could be dangerous if available to malicious actors,” the company wrote in its blog. To address these potential threats, Anthropic has said Fable 5 is being deployed with guardrails that block many of its responses to queries regarding potentially dangerous topics. For those user prompts, Anthropic will steer answers to an earlier, less powerful model called Opus 4.8 — which was the highest-performing publicly available model until Tuesday. “Fable 5’s capabilities in areas like cybersecurity, biology and chemistry are advanced enough that we’re taking a deliberately conservative approach for these topics at launch,” Anthropic wrote in reply to a question from NBC News. “To enable general availability of other Mythos level capabilities, we’ve decided to deploy safeguards that err on the side of caution, applying broad restrictions to these topics for now.” Anthropic said the two-pronged approach, diverting sensitive questions from Fable 5 to the older Opus 4.8, would allow users to still obtain helpful answers to questions when Fable 5’s capabilities could prove too dangerous. Anthropic said it will make Fable 5 available to all users on its Pro, Max, Team and seat-based Enterprise plans at no additional cost until June 22. After that, users will be able to access the model only by buying and using extra computing credits. Anthropic said it aimed to incorporate Fable 5 into usual subscription plans “as quickly as we can.” Anthropic has recently struggled to keep up with soaring demand for its AI systems, and Fable 5 is likely to add additional stress to its limited computational resources. Anthropic also announced that the trusted partners who had previously been able to access Mythos Preview will now be able to access an upgraded Mythos model, called Mythos 5. The company hadmade Mythos Preview available to over 150 organizationsaround the world to help financial institutions, software companies and healthcare networks shore up weaknesses in their cyber defenses prior to a wider public release of a Mythos-class model. Anthropic said Mythos 5’s capabilities shattered existing performance records across a range of other domains, including drug design and molecular biology. Anthropic said Mythos 5 is the company’s “first model to consistently produce novel, compelling scientific hypotheses.” President Donald Trump signed an executive orderlast week that aims to establish a voluntary mechanism for AI companies to share their systems with the government for safety testing before they are publicly released and to shore up the government’s own cyberdefenses. Anthropic signaled that the government tested Fable 5 ahead of its release, telling NBC News that it has offered the government early access to all its models for years.

This 2016 ATC Quest 24’ enclosed trailer was purchased new by the seller’s company to haul accessories to various automotive events. The trailer is constructed of aluminum sheet metal over a rectangular-tube frame and is finished in black with polished accents and manufacturer decals. It features LED lighting, a manually operated tongue jack, a right-side access door, a diamond-plate lower nose panel, and six-spoke 15″ wheels. Inside, upper and lower metal equipment cabinets are accompanied by a television, tie-down rails, recessed floor-mounted anchor tracks, and electrical outlets. This ATC trailer is now offered at no reserve by the seller on behalf of the owner with an awning and a clean Michigan title in the name of the owner’s company. The aluminum body is finished in black with various manufacturer decals and a diamond-plate front lower nose panel. The trailer measures approximately 24′ long and has a right-side access door, a manually operated jack, a 7-pin plug, a shore power connection, bright trim, and LED lighting. The tandem-axle trailer rides on 15″ alloy wheels mounted with 225/75 Goodyear tires. Braking is provided by electric drums. The tail panel folds down to form a compound loading ramp. The interior is outfitted with overhead lighting, tie-down rails, and recessed floor-mounted anchor tracks. The forward interior area houses black-finished upper and lower metal equipment cabinets, a work surface, a wall-mounted television, a power distribution panel, and electrical outlets. An awning is included in the sale. The winning bid does not include shipping. It is the buyer's responsibility to arrange the details of any shipping or delivery, and to pay any taxes, duties, or charges associated with shipping or delivery.View our third-party shipper recommendations. We need to confirm your billing address in order to appropriately charge fees and taxes should you win an auction. Please provide your billing address below. Congratulations! You're the high bidder. Your bid has been posted in the comment flow on the listing, and you can see other bids there as they happen. Good luck! Please confirm if the following details are aligned with your current contact information. If not, pleaseupdate your profile. Bidding will advance immediately to $. The BaT Service Fee is 5% of the bid, with a minimum of $250 up to a maximum of $7,500.VAT on Service Fee is charged in USD If you win the auction, your card will be charged for the service fee and you pay the seller directly for the vehicle. If you don't win, your existing pre-authorization will be released. When you bid we pre-authorize your credit card for the service fee(this helps prevent fraud). If you win the auction, your card will be charged for the service fee and you pay the seller directly for the vehicle. If you don't win, the pre-authorization will be released. *Exchange Rates You are bidding for this item in USD. This means, if you have the winning bid, you will need to make your payment to the seller in USD. It is your responsibility to check the conversion rate, and you should also note that exchange rates may fluctuate between now and the due date of your payment after the end of the auction. Taxation If you are the highest bidder, you will also need to pay the seller any applicable taxes/VAT. Your bid may not be inclusive of these amounts. Relevant details are included in the listing, so please ensure you have read and understood this information before placing your bid. Note that, if you will need to import the vehicle to your country, you may be responsible for import-related taxes. For more info,read about our auctionsoremail uswith any questions. By clicking on “Place a Bid” below, I acknowledge that theright to cancelservice will not apply once the bid has been placed, as the service will be provided immediately and agree to Bring a Trailer’sTerms of Use. Your bid of $is $more that the current high bid of $. Are you sure you want to proceed?

Retrofitting a port berth with shore power can take anywhere from three to seven years of permitting, construction, and grid upgrades. Now, a UK company has developed a floating hydrogen-powered platform and can make that wait disappear without having to move a single brick. The Hydrogen Power Hub, a modular floating platform developed by a UK-led consortium headed by Elire Maritime, has cleared six months of engineering trials, removing the last technical barriers before commercial deployment. The system can dispatch up to 5 MW of continuous clean power directly to a docked vessel while it sits at berth – no grid connection, no port construction required. The platform is built from three hexagonal modules that together cover around 1,200 sq m (12,917 sq ft). At full capacity, it can supply 91 MWh of energy per week, enough to serve mid-size cruise ships. The core of the system is a set of 1.3-MW modular hydrogen fuel cells – essentially electrochemical devices that convert hydrogen gas into electricity through a chemical reaction, with water as the only byproduct. Those fuel cells run continuously, consuming between 7,500 and 8,000 kg (16,535 and 17,637 lb) of hydrogen per week, slowly charging a 45-MWh onboard battery bank. When a ship pulls up, that stored energy can be discharged rapidly like a giant power bank. An onboard solar array generates up to 146 kW of additional power, giving the platform some autonomy between hydrogen resupply visits, which happen roughly twice a week by support vessel. Docked ships are some of the dirtiest neighbors a port city has. Their diesel auxiliary engines keep running, burning fuel and pushing exhaust over the surrounding city just to keep the onboard systems alive. The Hydrogen Power Hub cuts port emissions by 77% compared to those conventional diesel generators, saving an estimated 47 tonnes of CO2 per ship per week and eliminating the particulate pollution that drifts over surrounding cities. One of the more novel technical bets is the hydrogen storage system developed by Rux Energy UK, which uses nanoporous materials – materials riddled with microscopic pores that trap hydrogen molecules – to store the gas compactly and at low pressure. That's a meaningful safety and logistics advantage over high-pressure tank alternatives. The University of Strathclyde stress-tested the designs in wave tanks to verify structural integrity and inter-module connectivity under storm conditions. Schneider Electric and Ricardo plc independently verified that the electrical architecture can operate fully off-grid and thathydrogenintegration meets safety standards. Engineers found no technical barriers to full construction. Its main downside is price. According to Elire, hydrogen-generated electricity from this platform is estimated at £0.25–0.50 per kWh (around US$0.33 to $0.67), compared to £0.15–0.25 (~$0.20 to $0.33) for grid power or diesel – roughly two to three times more expensive. But the platform's proponents argue that speed and flexibility change the equation. It can be assembled, deployed, and relocated as shipping routes shift, avoiding the risk of expensive fixed infrastructure becoming stranded. "Ports are under increasing pressure to decarbonize while facing major infrastructure constraints," said Luke Jenkinson, founder and CEO of Elire Maritime. "We have validated a practical, scalable, and deployable system capable of delivering clean power directly where it is needed most." The consortium, funded through the UKRI Clean Maritime Demonstrator Competition Round 6, has entered early stage engagement with ports in London, Singapore, Hamburg, Brisbane, and Riga – ports already under regulatory pressure to cut emissions but unable to pause operations for years of construction. Source:Elire Maritime

pip install taosmdCopy PIP instructions Released:Jun 9, 2026 Framework-agnostic, local-first AI memory system. 97.0% end-to-end Judge accuracy on LongMemEval-S, measured on a low-end reference stack. Framework-agnostic AI memory system. 97.0% end-to-end Judge accuracy on LongMemEval-S. End-to-end Judge accuracy means retrieve → generate → LLM-grade against the reference answer. Runs offline on anything with 8 GB+ RAM and Python 3.10+: a Pi 4B, an old laptop, an Intel mini PC, a Mac mini, or a 16 GB Orange Pi 5 Plus (our reference low-end). Zero cloud dependencies. Part of thetaOSecosystem. Methodology and comparison notes indocs/benchmarks.md. Most memory benchmarks run on hosted models (GPT-4o-mini, Claude, Gemini) behind an API key, fine for prototypes, not fine if you're: The 97.0% on LongMemEval-S was measured on our reference low-end stack (Orange Pi 5 Plus, 16 GB RAM). The same code runs on a Pi 4B, an Intel mini PC, a Mac mini, an old laptop, or a workstation with a GPU, seeHardware Tested. No hidden hosted model doing the heavy lifting off-camera. Zero cloud dependencies. Zero API keys. NPU if you have one, CPU if you don't, cluster if you want the throughput. The cleanest way to install taosmd is to ask your agent to do it. Paste this message into Claude Code, Cursor, your OpenClaw shell, whatever: Please install taosmd as my memory system. The repo is github.com/jaylfc/taosmd. Don't summarise the repo or paraphrase the rules. Copy them verbatim, the wording is the contract. The agent will pull the repo, run the install, register itself, append the per-turn rules block to its own instruction file, and verify everything works. After that, every turn it runs it'll check the librarian when it's uncertain, seetaosmd/docs/agent-rules.mdfor the rules block it installs (also available viataosmd.agent_rules()). Multiple agents in one framework?Same install message works. The agent will ask you to name it before registering, so each agent gets its own shelf. The taosmd service stays one process with one shared set of stores; per-agent isolation is enforced by anagenttag on every row, not by separate files. Seedocs/multi-agent.mdfor the full naming convention, project-scoped and cross-agent memory, migration scenarios, and a five-agent worked example. Inside taOS?Don't use this. taOS provisions taosmd automatically when you deploy an agent, and the rules block is baked into the agent template. This install path is for standalone framework users. Note:A v0.3.0 PyPI release is prepared and will be published shortly; until it is live,pip install taosmdwill not resolve, so install from source:git clonethenpip install -e .(add the MCP server withpip install -e ".[mcp]"), verified on a clean environment. The one-line bootstrap (which additionally installs Ollama and downloads the embedding and LLM models) is newer and still being validated across clean machines, so please report issues. This will: Always install taOSmd into a virtual environment, and never withsudointo the system Python. A system-wide copy under/usr/lib/python3/dist-packages/taosmd(left by an earliersudo pip install) will shadow a venv editable install, soimport taosmdresolves to the stale system copy instead of your checkout. To check what is actually being imported and remove a stale system copy: Runningtaosmd serveas a systemd unit? Point the unit at your venv's interpreter (ExecStart=/path/to/.venv/bin/python -m taosmd serve ...); a PyPI or venv install needs noWorkingDirectoryand no repo checkout to start. If you're using Claude Code, OpenClaw, Cursor, or any AI coding agent, paste this into your conversation or CLAUDE.md / AGENTS.md: Set up taOSmd as my persistent memory system. Clonehttps://github.com/jaylfc/taosmdand runscripts/setup.sh, it installs everything (embedding model, Qwen3-4B LLM, data stores, daily archive compression cron). After setup, integrate taOSmd into your workflow: Store facts you learn about me: Archive every conversation turn (zero-loss: never summarise, never delete): Search my history when I ask about something: Extract facts automatically from our conversations: Before answering, check what you know about me: The archive is append-only and kept forever. Every conversation, tool call, decision, and error should be recorded. Old archives are compressed to gzip daily at 3 AM. The knowledge graph tracks structured facts with temporal validity, update facts when they change, don't delete them. Production leader recipefor LoCoMo-class workloads:qwen3.5:9bgenerator +--retrieval-top-k 20 --adjacent-turns 2 --llm-query-expansion --fusion mem0_additive --gen-temp 0.2. Reproduces 0.68 Overall / 0.55 Single-hop on full 1540 QAs undergemma4:e2bjudge, seedocs/benchmarks.md. All flags available on master as of PR #69. 97.0% end-to-end Judge accuracy on LongMemEval-S(500 questions, standard test set). Harness:benchmarks/longmemeval_runner.py. These are our own reproducible measurements, not a third-party-audited landscape ranking. Every number pins its generator, judge, dataset, and commit so you can re-run it on your own hardware, and we deliberately report under a strict local judge (qwen3:4b) alongside the lenient frontier-judge number that other systems publish, so the comparison is honest about what's being measured. See thejudge-sensitivity analysis. Seedocs/benchmarks.mdfor the full LongMemEval-S breakdown, the LoCoMo (1540-QA multi-session) measurements with retrieval-architecture ablations, methodology, and per-hardware-tier configuration recommendations (12 / 8 / 4 GB GPU, Orange Pi NPU, RPi 4). The Librarian adds LLM-assisted query expansion on top of the vector + cross-encoder stack. We measure its effect with a purpose-built three-axis harness on long-horizon sessions (60 turns, fact buried at turn 5). Axis C, vocabulary-gap coherence(2026-04-15, gemma4:e2b 5B, Fedora host): +15.4% on the vocabulary-gap axis.The cross-encoder alone adds nothing when the target fact is excluded from its candidate pool, only the Librarian's expansion bridges category→specific-name gaps (e.g. query:"code editor", fact:"Neovim lua config done"). These are preliminary results on one class of retrieval failure; we're actively working on tougher benchmarks to stress-test staleness detection and multi-store routing before drawing composite conclusions. LoCoMo-10 is a harder dataset than LongMemEval-S: 1540 QAs across multi-session conversations (50+ sessions, 400 to 700 turns), four categories, more pressure on the retrieval architecture. We run it on the smaller generators we actually target so the numbers reflect the hardware tier our users run on, not gpt-4o-mini. MaxSim + reranking is the current leader on the full 1540-QA test set, first across all three of our judges.It runs abge-v2-m3cross-encoder doing late-interaction (MaxSim) scoring over a wider k=50 candidate pool: MaxSim+rerank wins clearly under the lenient and qwen3-instruct judges (about +0.025 each) and is first, by a near-tie margin, under the strict llama judge. The ordering MaxSim > RRF > mem0_additive is identical across all three judges. Published numbers from Mem0/EMem/Zep use a lenient frontier judge (gpt-4o-mini), so thegemma4:e2b0.748 is the more apples-to-apples comparison with their headlines. Judge note.Our earlier strict judge (qwen3:4b) is a thinking model whose current build no longer emits clean YES/NO verdicts, so the strict column now uses two non-thinking judges:llama3.1:8bandqwen3:4b-instruct-2507. Legacy strict figures from earlier tables (RRF 0.557, mem0 0.540) came from an olderqwen3:4bbuild and are not directly comparable to the new judges. Seedocs/benchmarks.mdfor the full multi-judge analysis. Default recipe (tier-gated).Where the cross-encoder reranker is affordable (a GPU box, or any tier with headroom), MaxSim+rerank is the recommended default. On constrained tiers such as a Pi 4B on CPU, drop the reranker and use a lighter recipe (RRF or mem0_additive), trading roughly 0.02 to 0.06 for much lower latency. Both RRF and mem0_additive beat the older mem0-only guidance on every judge at full 1540, so any earlier "mem0_additive is the default" wording is superseded by this. Subset 200 ≠ full 1540 for every recipe.Earlier versions of this table reported subset-200 numbers for some rows. Validating those at full 1540 found the leader recipe (qwen+mem0+temp 0.2) generalises within −0.01 SH, but the previously-listed "Best Single-hop" pick (llama3.1:8b + RRF + temp 0.2) regressed by −0.16 SH at full scale and has been removed below. All ranks shown here are now validated at full 1540 before promotion; the asterisked rows are explicit about which scale they were measured at. Recommended generators at the 12 GB GPU tier(leader recipe, dual-judge scored, temperature-tuned): † Subset-200 measurement; the top two rows are validated at full 1540. The llama3.1:8b + RRF row from earlier versions of this table was removed after its full-1540 Single-hop measured at 0.49 (vs 0.65 on subset 200), failing the validation threshold. Other measured 12 GB-tier generators (Overall under gemma4:e2b judge, leader recipe):gemma4:e4b0.60 / 0.65 (best at temp 0.5),gemma4:e2b0.60 (best at temp 0.5),granite4:tiny-h0.56,phi4-reasoningtimeout. Per-generator temp sweet spots(matters more than we expected):qwen3.5:9bpeaks at temp 0.2;llama3.1:8bprefers fully-greedy (0.0);gemma4:e4bprefers 0.5 for Overall;gemma4:e2bprefers 0.0 for Single-hop. There's no universal sampling temperature for our local-tier stack, the right temp interacts with the model's training distribution. Seedocs/benchmarks.mdfor the per-generator × per-temp breakdown. About the dual judge.We score every new cell under two LLM judges:qwen3:4b(locally-runnable, deliberately strict, never refuses) andgemma4:e2b(lenient, calibrated closer to gpt-4o-mini). Thesame predictionsscore 0.28 vs 0.53 Single-hop respectively, judge strictness is a load-bearing variable in any LLM-judged benchmark. Reporting both is the honest middle ground between under-claiming under our strict judge and over-claiming under a frontier-API judge we can't afford to run on every cell. Hardware-tier defaults: 4-6 GB VRAM systems should keepqwen3:4bas their judge (fits comfortably, fast), 8 GB+ can rungemma4:e2b(7.2 GB) for matched-with-paper-SOTA comparison numbers. Full table, the 9B quant cliff (8 quants from Q2 through Q6, including the 8 GB-tier IQ4_XS at 0.55), the answer-prompt-variants and ENGRAM-typed-retrieval negative results, judge-sensitivity analysis, and per-hardware-tier configurations indocs/benchmarks.md. Tier breakouts for 4 GB / 8 GB / 16 GB Pi NPU tiers will land as those benches dual-rescore. taOSmd is a standalone library. Platform features like job scheduling, worker management, gaming detection, and mesh sync live in the host platform (e.g.,taOS). For multi-turn data where surrounding turns add context, populate an integer position field at ingest and askretrieve()for ±N positional neighbours. Worth +0.089 on LoCoMo same-tier, seedocs/benchmarks.md. Score retrieval by sign-bit Hamming similarity instead of full-precision cosine. Each vector collapses to1 bit per dimension (32× smaller)with integer-friendly distance, a footprint and CPU-speed win for memory-constrained or SBC deployments. It'srecall-neutral: −0.001 / +0.005 across both judges on the full 1540-QA LoCoMo set (seedocs/benchmarks.md). Off by default; standalone behaviour is unchanged unless you enable it. Use it when the vector-store footprint or CPU distance cost is your binding constraint rather than recall, e.g. an Orange Pi / Rock 5 holding a large memory. Keep it off on a GPU box where full-precision cosine is effectively free. Arecipeis a named, declared config bundle (retrieval + ingest + generator + librarian settings) that carries its own benchmark scores, target hardware tier, and pros/cons. Instead of leaving the retrieval levers at their defaults, taOSmd ships a small registry of recipes we have actually measured (for example themaxsim-rerank-9bleader for a 12 GB GPU and alite-pino-LLM-ingest profile for an Orange Pi / CPU), and afresh install auto-detects your hardware and applies the best affordable recipe on first use, so you run a benchmarked configuration rather than unconfigured defaults. No taOS or network is required: the hardware probe is local, and the reranker model (when a recipe asks for one) downloads on first use with progress and degrades gracefully if it is not yet present. taosmd.recipe_schema()returns the recipe bundle as a JSON Schema, so a host UI can render any recipe generically. The default is tier-gated: the reranking leader where the cross-encoder is affordable, a lighter recipe (RRF, or the lite profile) on constrained tiers. Seedocs/benchmarks.mdfor the per-recipe scores. taOSmd can expose its memory over theModel Context Protocolso any MCP-capable agent (Claude Desktop, Cursor, Codex, OpenWebUI, …) can read and write memory directly, no custom integration. The server is local-first and offline: it speaks thestdiotransport (the standard for desktop MCP clients), with no network listener and no cloud dependency. The MCP SDK is anoptional dependency, so the core install stays lean. From a source checkout: Run the server over stdio: Point an MCP client at it by spawning that command. For example, in a Claude Desktop / Cursor MCP config: Tools exposed (memory tools each take anagentargument, honouring the same per-agent isolation as the Python API): It reuses the same shared service layer as thelocal HTTP/REST server, so behaviour matches the Python API and CLI exactly. The MCP server is additive and opt-in, it only runs when you start it; standalone use is untouched, andimport taosmdworks whether or not themcpSDK is installed. By default taOSmd reads and writes a local data store on the same machine. When you want a thin client (a laptop, a desktop agent) to use memory that lives on a more capable machine (a Pi, a GPU box, a shared server), point the client at the server's HTTP URL and every data call routes there transparently. On the server machine, install taOSmd and start it as a background service, binding to all interfaces so remote clients can reach it: The helper scriptsscripts/install-server.sh(bash) andscripts/install-server.ps1(PowerShell) automate the install, health-check, and Tailscale setup guidance. On each client machine: The helper scriptsscripts/install-client.sh(bash) andscripts/install-client.ps1(PowerShell) automate the package install, URL configuration, skill install, and health check. You can also set the URL for a single invocation without touching the config: Or export it for a shell session: Resolution order (first non-empty wins):TAOSMD_SERVER_URLenv var, thenserver_urlin~/.taosmd/config.json, then local mode. By default there is no authentication. If the server is on a trusted private network (Tailscale, a home LAN), the network boundary is the access control. For defence in depth, you can require a bearer token: The token is sent asAuthorization: Bearer <token>on every request.GET /healthand the web inspector (GET /) are always public so monitoring probes keep working. Never commit the token to version control. TheRemoteClientclass (taosmd.remote) mirrors the same async interface as the local service module (taosmd.service). The CLI, Python API, and MCP server all checkTAOSMD_SERVER_URL(andconfig.json) at startup and delegate toRemoteClientwhen a URL is configured. From the caller's perspective nothing changes: the sametaosmd.ingest(),taosmd.search(), and A2A calls work in both modes. The ONNX model (models/minilm-onnx/model.onnx) is not included in this repo due to size (90MB). Download it: Or download directly fromHuggingFace. Minimum: 8 GB RAM and Python 3.10+. NPU/GPU optional, they speed up LLM tasks but aren't required. This is the author's primary deployment and the exact stack the 97.0% benchmark was measured on. Other tiers (Pi 4B, Intel mini, Mac mini, GPU box) run the same code, they swap the runtime (Ollama instead of rkllama, CPU/GPU instead of NPU) but keep the same models and the same architecture. Everything in this reference stack runs on the Pi itself; no external server needed for this tier.The Qwen3-4B handles both fact extraction and question answering on the NPU. The ONNX embedding model runs in-process on the CPU. An optional GPU worker (e.g. Fedora with RTX 3060) can accelerate LLM tasks ~10x but is not required, the Pi is fully self-contained. The default Manual Install path. Ollama serves the LLM, ONNX Runtime serves embeddings on the CPU. No platform-specific steps beyond the standard install. Not required for any tier, the LLM runs locally on whatever you've got. A GPU worker accelerates LLM tasks ~10x if you want to offload from a smaller node: taosmd servestarts a local HTTP/REST server (default127.0.0.1:7900, stdlib only, no new dependencies). It is a thin JSON shell over the same service layer as the Python API and CLI, so behaviour is identical across surfaces. Every endpoint that takes anagentparameter forwards it to the service layer, honouring the same per-agent isolation as the Python API. Security note:the server binds127.0.0.1by default, no auth is needed because only local processes can reach it. If you pass--host 0.0.0.0to expose it on a LAN, there is no authentication; put it behind your own network controls. Each hit in/searchresults has the agent-rules contract shape:{text, source, timestamp, confidence, metadata}. taosmd servealso exposes a lightweight message bus for agent-to-agent communication on the same port: Messages are stored as append-only archive events, so they inherit the archive's durability and automatic secret redaction.threaddefaults to"general"when omitted from/a2a/send. Each message object has shape{id, ts, from, body, thread, reply_to}. Three CLI commands consume the bus for monitoring (all share an id-based, exactly-once cursor and a--exclude <sender>filter):taosmd a2a-pollfetches only-new messages and updates a small state file (cron-friendly, durable across sessions);taosmd a2a-watchholds the SSE stream and prints one line per new message for instant in-session pickup; andtaosmd a2a-bridge --trigger <cmd>runs a command per new message with the message JSON on stdin, which can wake a dormant agent on arrival. The bundledtaosmd-a2askill (taosmd install-skill) walks through joining a channel and setting up durable or realtime monitoring. GET /andGET /uiserve a read-only local web dashboard, a bundled React single-page app with three views: memory search, the pending-review queue, and a live A2A channel monitor (it lists channels, then backfills a channel's history and live-updates over the SSE stream). It is served entirely from local bundled assets (no CDN, works fully offline); if the dashboard assets haven't been built, the server falls back to a minimal self-contained stdlib inspector page. Read-only, it exposes no write or destructive actions; the JSON endpoints above are the integration surface. To run the server as a background service (systemd on Linux, launchd on macOS, or a Scheduled Task on Windows), use--install-service: Seedocs/serve-service.mdfor platform-specific details, log locations, and how to change host/port/data-dir after installation. A crash between the archive write and the vector-store write iningest()can leave a conversation turn in the archive but absent from vector recall. The archive is the source of truth;reconciledetects and repairs that gap: --checkexits non-zero when any turn is missing, so you can run it from a cron health-check. The repair path re-adds only turns that are genuinely absent; it never resurrects turns that were intentionally superseded by a correction. Safe to run after a crash or periodically via cron. Copies the bundledtaosmd-a2aagent-setup skill into~/.claude/skills/taosmd-a2a/so it is available across all Claude Code projects. Pass--forceto overwrite an existing installation. All benchmark numbers indocs/benchmarks.mdpin the commit they were measured on. If taOSmd is useful to you: MIT Core taOSmd (the 97.0% benchmark) is fully self-contained, it uses only standard packages (SQLite, numpy, ONNX Runtime) plus the MiniLM embedding model. No external servers or forked repos needed. Optional integrations for the full taOS stack: Built byjaylfc. Part of thetaOSecosystem. Benchmark dataset:LongMemEval(ICLR 2025) Embedding model:all-MiniLM-L6-v2 0.3.0 Jun 9, 2026 Download the file for your platform. If you're not sure which to choose, learn more aboutinstalling packages. UploadedJun 9, 2026Source Filter files by name, interpreter, ABI, and platform. If you're not sure about the file name format, learn more aboutwheel file names. Copy a direct link to the current filtersCopy UploadedJun 9, 2026Python 3 Details for the filetaosmd-0.3.0.tar.gz. See more details on using hashes here. The following attestation bundles were made fortaosmd-0.3.0.tar.gz: Publisher:publish.ymlon jaylfc/taosmd Details for the filetaosmd-0.3.0-py3-none-any.whl. See more details on using hashes here. The following attestation bundles were made fortaosmd-0.3.0-py3-none-any.whl: Publisher:publish.ymlon jaylfc/taosmd Supported by

Il governo britannico sta preparando una mossa drastica nella protezione dei minori online che prende di mira Google, Apple e le altre aziende tecnologiche, le quali potrebbero presto essere obbligate per legge a bloccare l’accesso a immagini sessualmente esplicite su smartphone e altri dispositivi utilizzati da utenti con meno di 18 anni. Una proposta legislativa che alza l’asticella rispetto alle normative esistenti e che potrebbe avere ripercussioni ben oltre i confini del Regno Unito. Il primo ministro Keir Starmer ha voluto sottolineare la fattibilità della richiesta, respingendo in anticipo le obiezioni che potrebbero arrivare dalle big tech. “Non si tratta di una sfida impossibile. Queste sono alcune delle aziende più innovative al mondo e credo che possano risolverla“, ha dichiarato il capo del governo britannico, lasciando intendere che Londra non accetterà scuse tecniche o tentennamenti. Facciamo chiarezza. Segui TuttoAndroid su Google Discover Multe e responsabilità penale per chi non si adegua Le aziende produttrici che non rispetteranno la futura normativa potrebbero trovarsi a fare i conti con sanzioni economiche significative e, nei casi più gravi, con una vera e propria responsabilità penale per scoraggiare eventuali resistenze da parte dei colossi tech, storicamente poco inclini ad accettare imposizioni governative sul funzionamento dei propri dispositivi. L’obiettivo della proposta è integrare funzionalità specifiche nei dispositivi, anche attraverso aggiornamenti software per quelli già in commercio, in modo da impedire ai minori di scattare, inviare o visualizzare immagini sessualmente esplicite sui loro telefoni e tablet. Un intervento che andrebbe a incidere direttamente sul funzionamento della fotocamera e delle applicazioni di messaggistica. Segui Google su Telegram, ricevi news e offerte per primo Il problema dei contenuti autoprodotti dai minori Il Regno Unito ha già adottato l’Online Safety Act, ma ora punta ad alzare ulteriormente il livello di protezione. La ragione di questa accelerazione emerge chiaramente dai dati. Il 91% delle segnalazioni di abusi sessuali sui minori online registrate nel 2024 è stato legato a contenuti autoprodotti dai minori stessi. Un dato impressionante che evidenzia come il problema non riguardi soltanto la protezione dai contenuti esterni, ma anche la prevenzione di comportamenti rischiosi da parte degli stessi giovani utenti. Nel frattempo, il governo britannico sta lavorando anche su un’altra misura controversa ma che negli ultimi mesi ha preso sempre più piede in diverse nazioni nel mondo, ovvero un blocco all’accesso ai social media per i minori; il che andrebbe a configurare un approccio particolarmente restrittivo alla presenza online dei più giovani. Le prime reazioni delle aziende coinvolte La BBC ha raccolto una prima dichiarazione di Google, che sulla questione ha commentato in tono collaborativo. “Stiamo collaborando in modo costruttivo con i partner del Regno Unito per trovare soluzioni efficaci e rispettose della privacy che scoraggino la diffusione di contenuti dannosi, garantendo al contempo un ambiente digitale sicuro per i giovani“, ha fatto sapere l’azienda di Mountain View. Apple, invece, ha scelto per il momento di non commentare in alcun modo la proposta del governo britannico. Un silenzio che potrebbe indicare una fase di valutazione interna, o una strategia attendista in attesa di capire come si evolverà il percorso legislativo. Sebbene l’idea alla base sia lodevole e meritevole di attenzioni, resta ancora da capire come queste funzionalità potrebbero essere implementate tecnicamente senza compromettere la privacy degli utenti adulti e senza creare sistemi di scansione dei contenuti che potrebbero essere utilizzati per altri scopi. Ci troviamo di fronte al solito equilibrio delicato tra protezione e libertà; una situazione a cui le aziende coinvolte saranno chiamate a prendere una decisione, sempre che la proposta di legge arrivi effettivamente in porto. Vi terremo aggiornati in tal senso.

Fairfield, NJ, June 09, 2026 (GLOBE NEWSWIRE) -- A 2026 consumer overview examining Arctic Air Ice Jet X3 cooling technology claims, Arctic Air cooler cooling efficiency, return policy requirements and what product information publicly available across the Arctic Air portfolio. Where terms such as "under investigation" and "review" appear within this content, they refer to a structured overview of publicly available product information and company-published policies — not a third-party analysis, independent testing, legal proceeding, or external evaluation. All details should be verified on the official website. This overview contains affiliate links. If a purchase is made through these links, a commission may be earned at no additional cost to the buyer. This content is informational and does not constitute medical, health, or fitness advice. All product details described below are stated as presented by the company and should be verified directly on the official website before any purchasing decision. Arctic Air states that its product line - which includes Arctic Air Ice Jet X3 and Arctic Air Cooler - is designed to provide a portable and personal cooling experience all summer long. The company states features included Jet- Fan Technology, Ice-Powered cooling system, Triple-speed power, Whisper Quiet Operation, Compact and Portable, and LED Nightlight. Search activity around Arctic Air has increased as consumers look for more information about product features, pricing and transparency across the portable cooling systems. Interest in cooling devices continues to rise as consumers seek for an efficient and affordable alternative to keep home cold on the hottest summer night. This overview is structured to provide more clarity on product features, pricing, publicly available information across the Arctic Air Lineup. It walks through how product details and policies are presented across the company’s official materials with more emphasis on Arctic Ice Jet X3, the latest release by Arctic Air so readers can review how product details and policies are presented across the company’s official materials. View the current Arctic Air Ice Jet X3 offer (official Website) Why Search Interest Around Arctic Air Has Increased Enormously In 2026 Obviously, consumer awareness around Arctic Air has grown. As the brand has expanded from single product to multiple cooling systems, including the newest arrival, The Arctic Air Ice Jet X3, more people are looking into what is new and how the brand presents its product lineup within the cooling ecosystem. The types of questions that come up most often tend to fall into a few predictable categories: Does this actually work? Is the brand legitimate? How does it compare with others? What are the obvious limitations? What does the return policy look like in practice? This is a common step in the decision-making process and those are exactly the kind of questions this overview tries to address. One thing worth noting upfront: Arctic Air's official website displays customer ratings directly on its own product pages - an average 4.9 out of 5 rating for Arctic Air Ice Jet X3 out of thousands of users. They also claimed that 94% of users will likely recommend Arctic Air Ice Jet X3 this summer. These figures appear on the company’s own storefront, if independently verified third-party ratings matter to you, searching for products on external consumer platforms can give you a broader picture. Arctic Air Ice Jet X3 Product Details and Features The Arctic Air Ice Jet X3 is a small, portable personal air cooler manufactured and sold directly by Arctic Air. At its core, it is a personal cooling unit designed not to cool an entire home or office space, but to cool you — the individual sitting in front of it. Unlike traditional air conditioning units, which require professional installation, refrigerant systems, ductwork, and significant electrical infrastructure, Arctic Air claims that this is a plug-and-play device. It sits comfortably on a desk, nightstand, kitchen counter, or any flat surface, and requires nothing more than a power outlet (or USB port) and a supply of ice to operate. From What Arctic Air Claims, the concept looks simple: you fill the internal cooling chamber with ice, switch the unit on, and within seconds it begins blasting cold, refreshing air directly at you. They claim that there are no technicians to call, no installation appointments to schedule, no $3,000 upfront equipment costs, and no dramatic monthly electricity spikes to dread. Special Offer -Visit The Official Product Sales Page For More Details About Arctic Air Ice Jet X3 The Science and Technology Behind It To understand whether the Ice Jet X3 truly delivers, it helps to understand the physics at work. The device operates on what Arctic Air called evaporative cooling combined with forced convection — a combination that has been used in various forms for thousands of years, from ancient Egyptian wet-cloth cooling techniques to modern industrial evaporative coolers. Here is the basic science:when warm air passes over or through a cold medium (in this case, ice), heat energy is transferred from the air to the ice. The ice absorbs that thermal energy as it melts, and the air that exits the other side of the process is significantly cooler. The Arctic Air Ice Jet X3 accelerates and amplifies this process using what Arctic Air calls Jet-Fan Technology. Jet-Fan Technology, according to Arctic Air Reviews, is described as being inspired by the airflow dynamics of actual jet engines. The fan draws warm ambient air in through the back of the unit at high velocity, forces it through the ice-filled cooling chamber at close range, and then expels the cooled air outward in a directed, concentrated stream. The result is not a gentle misting effect or a vaguely cool breeze — it is a focused blast of cold air that can be felt immediately and directly. If true, then it is the key distinction between a basic fan and the Arctic Air Ice Jet X3. A standard desk fan simply moves existing room-temperature air around, it creates the sensation of cooling through evaporation of sweat on your skin, but it does not actually reduce the temperature of the air itself. The Arctic Air Ice Jet X3, by contrast, genuinely lowers the temperature of the air it outputs if the working technology is true as the paper said. Basically, The ice does real thermodynamic work, and the Jet-Fan delivery system ensures that cooled air reaches you efficiently rather than dispersing into the room.. quoted by the Manufacturer, Arctic Air. Obviously, the effectiveness of the unit is naturally dependent on how much ice is loaded and ambient room conditions. In a small, enclosed space with direct positioning, the cooling effect is most pronounced. In a large open area, the cold air will dissipate more quickly. This is why Arctic Air classified it as a personal cooler. It is designed to cool a person, not a room. Used correctly, with the unit positioned within a few feet and pointed directly at the user, Arctic Air Ice Jet X3 claims to perform exactly as advertised. Key Features The Arctic Air Ice Jet X3 comes equipped with a range of features that make it practical, versatile, and genuinely pleasant to use day-to-day: Jet-Fan Technology:The proprietary fan system draws hot air in and pushes cold air out at speed, delivering a concentrated stream of chilled air rather than a soft, diffuse breeze. This is the engineering heart of the device and the primary reason it outperforms standard desk fans for cooling comfort. Ice-Powered Cooling Chamber: The internal compartment is specifically designed to hold ice efficiently, maximizing contact between the ice and incoming air. The chamber is insulated to extend ice life, keeping the air output cold for hours rather than minutes. Three Fan Speed Settings:Arctic Air Claims Users can choose between Low, Medium, and High settings, giving full control over airflow intensity. Low is ideal for quiet overnight use or close-range cooling; High delivers maximum cold air output for peak heat conditions. Whisper-Quiet Operation:One of the most frequently praised aspects of the Arctic Air Ice Jet X3 is its remarkably low noise profile. The unit is engineered to run quietly enough that it can be used in a bedroom throughout the night without disrupting sleep — something that cannot always be said for traditional window AC units or even portable air conditioners. Built-In LED Nightlight:The Arctic Air Ice Jet X3 includes an integrated LED nightlight with seven color options. This soft, ambient glow makes it particularly well-suited for bedroom use, providing gentle illumination without the harshness of overhead lighting. Portable and Lightweight Design:The compact form factor makes the Arctic Air Ice Jet X3 easy to move from room to room, or even take outside the home. It is well-suited for dorm rooms, home offices, garages, RVs, campers, and any space where personal cooling is needed without the ability to install permanent air conditioning. Universal Power Compatibility:The unit connects via a standard US power cable that plugs into any wall outlet or compatible USB port, making it accessible without any special electrical requirements. No Filters, No Maintenance:Unlike traditional air conditioning systems that require periodic filter replacements, refrigerant checks, and professional servicing, the Arctic Air Ice Jet X3 requires no ongoing maintenance beyond cleaning and restocking with ice. Pricing Thecompany’s product pagelists the single unit of Arctic Air Ice Jet X3 at $49.99, two units at $99.99, three units at $139.99 and four units at $179.99. The company didn’t disclose anything like shipping and handling fees so any additional fees should be displayed at checkout. Who The Arctic Air Product Line May Be Right For? These products may align well with people who: Remote workers and home office users who spend long hours at a desk and want targeted personal cooling without running whole-home AC continuously. People in hot climates who want to reduce their reliance on central air conditioning without sacrificing comfort. Students in dorms or shared apartments where central AC is limited or shared. Light sleepers who want nighttime cooling without the noise of traditional AC units. RV, camper, and van-life enthusiasts who need portable, low-power cooling on the road. Garage and workshop users who work in spaces that are not connected to the home's HVAC system. Budget-conscious households looking to lower summer electricity costs. Anyone in a rental property where installing permanent air conditioning is not possible or practical. Visit Official Website to view current price, return policy and other information regarding purchase What Consumers Are Looking For When Searching For A Portable Cooling Device In 2026 Product transparency:Obviously, researchers want to see the real company behind the product, their address, history etc which help in building trust. Verified technology:Is the science behind its operations legit or just an engineering word to draw attention. Researchers always want to go deeper to get more confidence. Real user experience:Real user experience is definitely a factor to consider when evaluating such products. Most people are looking at firsthand information which is okay for anyone looking to try something but doesn’t want risk. Build Quality:Researchers are looking at durability of the product, evaluating whether it can stand the test of time or just use and thrash device. Return policy clarity.The company describes a 30-day money-back guarantee for the Arctic Air Ice Jet x3. Before relying on this as a safety net, contacting the support team to confirm current return requirements — including whether return shipping is your responsibility, whether original packaging is needed, and how long refund processing takes Questions Consumers Commonly Ask About Arctic Air Is Arctic Air Legitimate Brand? Arctic Air is a commercial operating business based in FairField. The company provides customer support via phone and email during published business hours. As with other products, reviewing product details, return terms and company’s publish policies before ordering is a sound approach. Are The Customer Ratings On The Official Website Independently Verified? All ratings displayed on Arctic Air Website appear on the company’s own product pages. Whether these ratings are aggregated from an independent third-party platform is not clearly stated on the website. Searching for Arctic Air product information on external consumer guides provides an additional, helpful insight. What’s The Difference Between Arctic Air Ice Jet X3 and Arctic Air Ice Jet? From theofficial website, both products are the same. Just spelling differences. Is Arctic Air Cooler On Sale Now? Currently, Arctic Air Cooler is not available, the company only advertise The latest version called the Arctic Air Ice Jet X3 Can These Replace A Home Central Cooling System? No, even with multiple units. The website advertised it as a personal cooling system. Ideally, a single unit for one person. What Are The Limitations Of This Brand Of Cooling System? Transparency requires acknowledging what the Arctic Air brand does not do. Obviously, the Arctic Air Ice Jet X3is not a whole-room or whole-home cooler. If your goal is to lower the ambient temperature of a large room, this device will underperform. Effectiveness decreases in extreme heat without regular ice replenishment. On very hot days, ice will melt faster, requiring more frequent refills to maintain optimal cooling output which is an obvious drawback to know if you are looking to invest in this type of air cooler. It requires a steady supply of ice. While this is a minor inconvenience for most households, it is worth noting that ongoing operation has a soft cost in the form of ice consumption. It will not dehumidify though some reviews claim that it does but it really doesn’t. Traditional air conditioning also removes humidity from the air, which contributes significantly to perceived comfort in humid climates. Limited range. Its cooling effect is most powerful within a few feet of the unit. It is not effective for cooling people on the other side of the room. Complaints From Verified Users In The United States Obviously, Some users express frustration with the ongoing need to refill ice, particularly on very hot days when ice melts quickly. For people who do not have a convenient ice supply at home, this can become a minor but recurring inconvenience. A small number of buyers who purchased it as a whole-room or whole-home cooling have been disappointed — though this reflects an expectation mismatch rather than a product failure. Most marketing materials from some platforms clearly position it as a personal cooler. A few users in very large, open spaces (such as large garages or outdoor patios) noted that the cold air dispersed too quickly to provide meaningful relief without positioning the unit very close to the body. Verdict The Arctic Air is just a brand offering a range of portable air coolers though we concentrate on Arctic Air Ice Jet X3. Will hope this work addresses some of you query and we are sorry if it doesn’t. Complete product details, current pricing and published terms are available at the link below: View the current Arctic Air Ice Jet X3 offer (official Website)

Delray Beach, FL, June 09, 2026 (GLOBE NEWSWIRE) -- According to MarketsandMarkets™, theFacility Management Marketis projected to grow from USD 61.08 billion in 2025 to USD 138.50 billion by 2030, expanding at a CAGR of 17.8% during the forecast period. Browse 383 market data Tables and 56 Figures spread through 318 Pages and in-depth TOC on "Facility ManagementMarket - Global Forecast to 2030" Facility Management Market Size & Forecast: Facility Management Market Trends & Insights: Download PDF Brochure @https://www.marketsandmarkets.com/pdfdownloadNew.asp?id=1030 This growth reflects a rising demand among organizations to streamline complex building operations and ensure that physical assets, energy systems, and workplace environments function in sync to support broader business goals. As organizations increasingly prioritize occupant well-being, operational efficiency, and sustainability, several market forces are actively reshaping the facility management landscape. Technological advancements, particularly the integration of Integrated Workplace Management Systems (IWMS), IoT-based asset monitoring, AI-enabled predictive maintenance, and building information modeling, are playing a critical role in driving this transformation. Simultaneously, stricter regulations around energy consumption, carbon emissions, and occupational health and safety are pushing businesses to modernize and automate their facility operations. These developments are enabling organizations to optimize resource usage, enhance security protocols, and achieve higher levels of regulatory compliance. As a result, facility management is shifting from a traditionally cost-centric function to a strategic enabler that contributes directly to improved organizational performance, reduced expenses, and long-term sustainability outcomes. Based on offerings, the solutions segment is expected to hold the largest market size during the forecast period The solutions segment within facility management encompasses a comprehensive suite of integrated technologies and services designed to streamline operations, enhance efficiency, and ensure compliance across various facility types. This segment includes offerings such as integrated workplace management systems (IWMS), building information modeling (BIM), and advanced security systems, which collectively address the multifaceted needs of modern facilities. The increasing complexity of building operations, driven by sustainability mandates, regulatory requirements, and the need for real-time data analytics, has led organizations to adopt these holistic solutions. Gasunie, the Netherlands’ main natural gas transportation network, had 90 different systems for facility processes. To improve efficiency and reduce costs, they consolidated these systems by implementing Planon's IWMS, which included modules for work order management, service desk, move management, project management, and maintenance management. This streamlined their processes down to just one system. King's College London, managing 4.3 million square feet across 130 buildings, used Planon's IWMS platform to gain control and transparency, reducing its maintenance backlog by 50% in three months and shifting from reactive to preventive maintenance. Request Sample Pages @https://www.marketsandmarkets.com/requestsampleNew.asp?id=1030 Based on services, the managed services segment is expected to witness a higher growth rate during the forecast period Managed services in facility management refer to the outsourcing of specific operational functions to specialized third-party providers who assume responsibility for the day-to-day management and maintenance of these services. This approach allows organizations to focus on their core competencies while ensuring that facility operations are handled by experts with the necessary skills and resources. The growing complexity of building systems and the increasing need for compliance with stringent regulations and standards have made managed services an attractive option for many organizations. For instance, Maastricht University Medical Center (MUMC) integrated its facilities data using Planon's IWMS, resulting in improved operational efficiency and a significant reduction in customer queries. The hospital's Facilities Unit, comprising 340 employees and managing a USD 38.9 million budget, streamlined operations across engineering, logistics, catering, and security departments. The implementation of managed services enabled MUMC to eliminate ambiguity in its processes and gain better insight and transparency, enhancing overall service delivery. This shift toward managed services is driven by the need for specialized expertise, cost efficiency, and the ability to leverage advanced technologies without the burden of in-house management.Asia Pacific is expected to witness the highest growth rate during the forecast period The Asia Pacific region is experiencing rapid urbanization, industrialization, and infrastructure development, all of which contribute to the accelerated demand for facility management services. China and India are witnessing significant growth in their real estate sectors, leading to an increased need for efficient facility operations and management. The expansion of commercial real estate, manufacturing, healthcare, education, retail, and hospitality sectors in the region further drives the demand for comprehensive facility management solutions. Additionally, the growing emphasis on sustainability and energy efficiency is prompting organizations to adopt green building practices and technologies, which require specialized facility management expertise. For instance, Fortescue Metals Group (FMG), based in Western Australia, has adopted System Information Modeling (SIM) for all its projects built since 2010, including large-scale projects like the Solomon Iron Ore project and the expansion of its export port facility. FMG acknowledges that using SIM on these projects resulted in significant savings and more efficient project execution. The adoption of SIM technology in the Asia Pacific region is indicative of the broader trend toward digital transformation in facility management, driven by the need for improved efficiency, cost savings, and sustainability. As these trends continue to unfold, the Asia Pacific region is poised to lead the facility management market in terms of growth rate during the forecast period. Inquire Before Buying:https://www.marketsandmarkets.com/Enquiry_Before_BuyingNew.asp?id=1030 As the facility management market moves beyond traditional maintenance to become a critical enabler of organizational performance, the real opportunity lies in convergence, where digitalization, sustainability, and service integration meet. Players that can shift from offering siloed solutions to delivering intelligent, outcome-based platforms will define the competitive frontier. The success of Planon’s IWMS in reducing maintenance backlogs by 50 percent and Gasunie’s consolidation of 90 systems into one operational ecosystem proves that enterprises are no longer just buying services, they are investing in transformation. Similarly, Asia Pacific’s accelerated adoption of technologies such as SIM, as seen with Fortescue Metals, highlights a regional appetite for scalable, high-impact solutions that align with both growth and green imperatives. To capitalize on this momentum, vendors must transition from solution providers to long-term transformation partners by offering modular platforms, data-driven services, and agile delivery models tailored to specific verticals and regulatory landscapes. The winners will be those who do not just respond to demand but shape it by turning facilities into levers of resilience, efficiency, and strategic value creation. Top Companies in Facility Management Market: TheTop Companies in Facility Management Marketinclude Caterpillar (US), ABB (Switzerland), Cisco (US), Sandvik AB (Sweden), Epiroc (Sweden), Hexagon AB (Sweden), Komatsu (Japan), Rockwell Automation (US), Metso (Finland), and Schneider Electric (France).

Searah, a new 50/50 independent joint venture between Eni and Petronas combining key businesses across Indonesia and Malaysia, has been created as Southeast Asia’s new independent integrated energy company, with a portfolio of 19 gas-producing and development assets. Searah wasestablishedjust seven months after the signing of the investment agreement between Eni and Petronas on November 3, 2025, and 16 months after the memorandum of understanding announced in February 2025. The joint venture brings together the complementary portfolios, capabilities, and regional expertise to deliver long-term value creation and operational excellence across Indonesia and Malaysia. Claudio Descalzi, CEO of Eni, commented:“Searah reflects our proven satellite strategy that aims at building focused, high-quality businesses that can combine scale, efficiency, and growth, and that are driven by our excellence in exploration and project execution, and our continued focus on technology and innovation. “Searah is a strong new entity in Southeast Asia—the first and largest of its kind in the region—combining our expertise with that of Petronas to support the development of energy resources in Indonesia and Malaysia, with a strong commitment to environmental protection and local growth.” All required regulatory, governmental, and partner approvals in both Malaysia and Indonesia have now been obtained, and all conditions precedent have been met. The firm’s gas-producing and development portfolio consists of 14 assets in Indonesia and five in Malaysia. The company will start with an initial production base of over 300,000 barrels of oil equivalent (boe) per day, aiming to exceed 500,000 boe/d of sustainable production within the next three years. Tengku Muhammad Taufik, President and Group CEO of Petronas, emphasized:“The establishment of Searah aligns with Petronas’ intensified focus on exercising greater discipline in developing resources coupled with more agile capital deployment as well as stronger emphasis on sustained value creation across the gas value chain. “Leveraging the complementary portfolios and capabilities of both Petronas and Eni, Searah is envisaged to bring the operational depth, financial resilience and growth capacity of both partners in addressing the region’s growing energy needs reliably and responsibly, even as it contributes towards the long-term security of supply in Indonesia and Malaysia.” A $6 billion revolving credit facility has been secured, reflecting the strong confidence of the financial markets to fund Searah’s growth plans, which include a pipeline of expected investment for over $20 billion over the next five years to support the development of more than 3 billion boe of discovered resources and unlock multi-billion boe of additional exploration potential. All members of Eni Indonesia and Petronas Indonesia staff have transitioned to Searah, alongside the establishment of Searah Malaysia, a dedicated entity created to manage Malaysian assets. The launch of the joint venture follows shortly after thefinal investment decisions (FIDs)for theGendaloandGandangfields (South Hub), andGeng NorthandGehemfields (North Hub). This content is available after accepting the cookies. Two deepwater gas hubs on the Asian energy horizon as Eni unveils FIDs These projects are said to hold nearly 10 trillion cubic feet (tcf) of gas initially in place (GIIP) and approximately 550 million barrels of associated condensate, with production expected to start in 2028 and reach a plateau of 2 bscfd of gas and 90,000 bpd of condensate by 2029. Eni alsorecently announcedthe giantGeliga-1gas discovery in the Ganal block in the Kutei basin, estimated to contain around 5 tcf of gas and 300 million barrels of condensate in place. The well has demonstrated excellent reservoir quality, capable of producing approximately 200 million scfd of gas and 10,000 bpd of condensate. The establishment of Searah is perceived to support the deployment of capital and resources required to achieve a short-term production target of 500,000 boe/d equity, while advancing further development opportunities stemming from the success of the Geliga exploration well. Take the spotlight and anchor your brand in the heart of the offshore world! Join us for a bigger impact and amplify your presence at the core hub of the offshore energy community!

Acoustic communication is crucial for survival across the animal kingdom, with acoustic signals being shaped by the interaction of producer and receiver selective pressures. While spectral features’ variation reflects species-specific selection, the evolutionary history of acoustic communication rhythms, i.e., the rhythmic modulations of acoustic signals, remains unknown. Using data from 98 species spanning primarily mammals and birds, with additional representation from amphibians, reptiles, fishes, and insects, we investigate the origins of acoustic communication rhythms, notably whether they are shaped by the producer’s anatomical characteristics, environmental constraints, or social complexity. Regression models which controlled for phylogenetic relatedness did not support an influence of these species-specific selective forces; instead, explicit phylogenetic models of trait evolution showed that most species’ rhythms are conserved around an evolutionary optimum of 2.7 Hz that falls within the neural delta range (1–4 Hz) and predates mammalian divergence. Given the known conserved brain oscillations across species and delta involvement in active sensing, we propose that, unlike spectral features, acoustic rhythm could be governed by a universal neural mechanism facilitating effective intra and interspecific communication via a shared channel that has persisted through evolutionary times. Citation:Piette T, Cathcart C, Barbieri C, Martin Ming K, Grandjean D, Bickel B, et al. (2026) Animal acoustic communication has a conserved optimal rhythm within the neural delta range. PLoS Biol 24(6): e3003798. https://doi.org/10.1371/journal.pbio.3003798 Academic Editor:Gail L. Patricelli, University of California Davis, UNITED STATES OF AMERICA Received:July 8, 2025;Accepted:April 28, 2026;Published:June 9, 2026 Copyright:© 2026 Piette et al. This is an open access article distributed under the terms of theCreative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability:All data and code used for the analysis in this study are available in a public Zenodo repository. This repository contains the raw data, scripts for data processing, analysis, and visualization, as well as detailed instructions for reproducing the results. The repository can be accessed athttps://zenodo.org/records/19816728. Funding:The NCCR Evolving Language, Swiss National Science Foundation Agreement Nr. #51NF40_180888, funded this work (https://evolvinglanguage.ch/). This work has also received support from the French National Research Agency (https://www.info.gouv.fr/grand-dossier/france-2030) under the France 2030 program (ANR-23-IAHU-0003), the Fondation pour l’Audition (FPA IDA11 to ALG,https://www.fondationpourlaudition.org/). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing interests:The authors have declared that no competing interests exist. Abbreviations:ANOVA, analysis of variance; BM, Brownian Motion; CI, credible intervals; ELPD, expected log pointwise density; IEI, inter-element interval; MCC, maximum clade credibility; MCMC, Markov chain Monte Carlo; PCM, phylogenetic comparative methods; SNR, signal-to-noise ratio Acoustic signals enable effective and instantaneous communication between individuals, even over considerable distances. For these signals to be functional, their structure must convey adaptive information. The importance of spectral features, such as frequency content and pitch, in achieving this has been well established [1–5]. However, acoustic signals are structured not only spectrally but also temporally. Rhythms, low-pass temporal modulations of the acoustic content, also known as the sound envelope, play equally important communicative roles. This is particularly well exemplified by human speech, where speech slow modulations are sufficient for comprehension [6,7]. In most languages, syllable production rate ranges between 4 and 9 syllables per second [8]. This range corresponds to theta neural oscillations, which flexibly adapt to the syllabic rhythm during speech perception, such that modifying this flow impedes this process and reduces speech intelligibility. Rhythmic patterns allow the identification of syllables, but also words, and sentences. At a slower timescale, delta oscillation (1–4 Hz) aligns with prosodic features, and can help convey meaning, intonation, and emotional state [9,10]. In animal calls, such temporal features are no less important, for instance, in vocal recognition [11], mating behavior [12], and predator avoidance [13]. Thus, the temporal patterning of acoustic sequences carries significant communicative functions and is not specific to speech. However, questions remain regarding the factors that influence the evolution of rhythm across the animal kingdom. In this study, we explore acoustic communication rhythms across diverse animal clades, test leading hypotheses on the evolution of signal structure, and model the most plausible evolutionary scenarios. Taking a macroevolutionary perspective, we examine whether large-scale patterns of rhythm variation across species can be explained by shared evolutionary constraints. We consider four major selective forces that may drive the evolution of rhythm. First, the presence of theta rhythms in the vocal productions and mouth movements of non-human primates suggests that such rhythms may originate from the natural oscillatory movements of the articulators, inherited directly from mastication [14–16]. Second, in vocalizing animals, morphological and physiological factors, such as breathing rate, heart rate, or metabolism, may constrain rhythmic range in a manner analogous to constraints on spectral features [17]. Third, for acoustic signals to be effective, they must reach receivers despite environmental constraints, which could shape not only spectral but also temporal features of communication [18,19]. Additionally, the social complexity hypothesis proposes a positive relationship between the complexity of an animal’s social environment and the complexity of its vocal repertoire [20]. Since rhythm determines the rate at which information can be transmitted, social complexity could also positively influence acoustic tempo, as species with more complex social systems may need to convey more information within a given time frame. Finally, it is possible that rhythm evolution is not shaped by species-specific selective pressures, but rather by phylogenetic processes of conservation and diversification that are either shared across, or vary between, lineages. To test these different hypotheses, we quantified rhythm in acoustic sequences primarily from birds and mammals, with additional, more limited data from amphibians, insects, reptiles, and fish. Using Bayesian multilevel models for phylogenetic regression [21], we controlled for phylogenetic relationships and evaluated, in accordance with the previously stated hypotheses, whether weight (as a proxy for breathing rate, heart rate, and metabolism), mastication status, sociality level or ecological characteristics could account for differences in rhythm. Finally, we compared phylogenetic models capturing contrasting evolutionary scenarios, asking whether rhythm shows signatures of convergence toward a shared optimum, as expected under an Ornstein–Uhlenbeck (OU) evolutionary process, or whether it has diversified through unconstrained, Brownian-like divergence across lineages. To explore the evolution of communication rhythm in animals, we analyzed acoustic sequences from 98 species (58 birds, 28 mammals, 4 amphibians, 4 insects, 1 reptile, and 1 fish). We calculated rhythm by analyzing the low-pass variations in the signal amplitude, allowing broad applicability across species (Fig 1a–1d). Validation against conventional methods showed consistent results (Fig 1e:F2.171= 0.33,p= 0.72), confirming the robustness of our approach. Control of the impact of signal-to-noise ratio (SNR) and sequence length revealed no significant relationship with rhythm (S1aandS1b Fig). Additional investigation of the allometric relationship between animal weight and the dominant frequency of acoustic signals (S2 Fig), and of the effect of context on rhythm (S4a Fig) further confirmed the validity of our acoustic database. a)Oscillogram of one acoustic sequence of polar skua call, green line shows the computed signal envelope.b)Power spectrum and time-frequency representation of the envelope of the previous sequence.c)Power spectra of the envelopes of all polar skua acoustic sequences.d)Number of sequences per species (ordered by alphabetical order).e)Rhythm computation using sequences of 10 randomly selected species, employing, from left to right, number of elements per second (Count), inter-element intervals (IEI), and wavelet method (Wavelet). The data underlying this Figure can be found inhttps://zenodo.org/records/19816728. a)Oscillogram of one acoustic sequence of polar skua call, green line shows the computed signal envelope.b)Power spectrum and time-frequency representation of the envelope of the previous sequence.c)Power spectra of the envelopes of all polar skua acoustic sequences.d)Number of sequences per species (ordered by alphabetical order).e)Rhythm computation using sequences of 10 randomly selected species, employing, from left to right, number of elements per second (Count), inter-element intervals (IEI), and wavelet method (Wavelet). The data underlying this Figure can be found inhttps://zenodo.org/records/19816728. https://doi.org/10.1371/journal.pbio.3003798.g001 To investigate which factors influence the evolution of rhythm, we fitted two phylogenetic regressions in a Bayesian multilevel framework. A full model investigating the impact of weight, mastication status, and living environment while controlling for phylogenetic relatedness, and a null model controlling for phylogenetic relatedness only. Because insects, fish and anurans are underrepresented in our dataset, we restricted the comparative evolutionary analysis to birds and mammals to ensure robust phylogenetic inference. We extended the analyses to the full dataset in the supplementary materials as an exploratory approach, providing an indication of potential generalization. We compared models via their leave-one-out expected log pointwise density (ELPD) and stacking weight. Due to heteroskedasticity, distributional (scale-location) models better fit the data, leveraging over 90% of the stacking weights (S3 Fig). Including the predictors did not improve predictive performance (Fig 2a), and the null model leveraged the highest stacking weight (Fig 2b). The posterior distribution of the regression coefficients of the full model revealed that none of our predictors had any decisive effect on rhythm, given that their 95% credible intervals (CI) all contained zero. Taken together, these results indicate that including these predictors does not add decisive explanatory value to phylogenetic history. The supplementary analyses including all species yielded comparable results, suggesting that this pattern may generalize to more distantly related taxa (S4 Fig). As vocal complexity, our proxy for social complexity in vocalizing species, and beak morphology in birds were not available for all species, these parameters could not be included in the main phylogenetic model. We therefore conducted separate linear mixed model analyses, regressing rhythm on each factor, with phylogenetic class and order as random effects to account for evolutionary influence. These analyses revealed no significant relationship between rhythm and either vocal complexity (S5d Fig:t= −0.75,p= 0.46) or beak morphology (S5e–S5g Fig:t= −1.4, 0.98, 0.35;p= 0.17, 0.33, 0.73). a)Leave-one-out expected log pointwise density difference (ELPD) between the null and the full distributional (scale-location) models.b)Stacking weight of the models.c)Posterior credible intervals (95% and 85%) of the full distributional model (W*M = Weight*Mastication).d)Rhythm plotted as a function of log-transformed weight with predicted slopes from the full distributional model and their standard error.e)Distribution of raw median rhythm (in Hz) across birds and mammals. Each vertical black line represents the median rhythm of a single species. The data underlying this Figure can be found inhttps://zenodo.org/records/19816728. a)Leave-one-out expected log pointwise density difference (ELPD) between the null and the full distributional (scale-location) models.b)Stacking weight of the models.c)Posterior credible intervals (95% and 85%) of the full distributional model (W*M = Weight*Mastication).d)Rhythm plotted as a function of log-transformed weight with predicted slopes from the full distributional model and their standard error.e)Distribution of raw median rhythm (in Hz) across birds and mammals. Each vertical black line represents the median rhythm of a single species. The data underlying this Figure can be found inhttps://zenodo.org/records/19816728. https://doi.org/10.1371/journal.pbio.3003798.g002 Visual inspection of the median rhythm for our tested species shows that in every class, acoustic rhythm spans mainly the lower rates <5 Hz (Fig 2d). To understand if the produced rhythms have randomly evolved within this range, favoring species-specific rhythm, or have been maintained around an optimum value, we fitted models of the evolution of vocal rhythm under Brownian Motion (BM) and OU processes [10], representing each evolutionary scenario, respectively. Comparisons of ELPD values and model stackings show that the OU model best fits the data (Fig 3a). The median posterior estimate of rhythm is 2.7 Hz, with 95% CI of [0.45, 4.99] Hz (Fig 3b). This represents both the optimum to which the OU process reverts over time and the likely state at the root of the phylogeny. The posterior distribution of sigma represents the stochastic volatility (Fig 3c) and has a median of 0.68 and a 95% CI of [0.35,2.37]. The posterior distribution of alpha represents the strength of attraction (Fig 3d) and has a median of 1.95 and a 95% CI of [0.42,6.01]. The proportion of posterior half-life estimates lower than 1 (height of the tree) is 95% (Fig 3e), supporting strong selection with fast reversals to the optimum. Consistent with this, rhythm values close to the optimum are also reconstructed for most interior nodes of the phylogeny (Fig 3f). In summary, the model suggests that there is a phylogeny-wide evolutionary pressure towards an optimal rhythm, to which species that would deviate quickly revert to. This result is further supported by models fitted separately for mammals and birds, where the OU model consistently received the highest stacking weight (birds: wOU = 1, mammals: wOU = 0.925) and showed a similar pattern of slow rhythm conservation (S6 Fig). In addition, a supplementary model that included anuran, insect, and fish species yielded comparable outcomes (S7 Fig), suggesting that the conservation of slow acoustic rhythm may have an even older evolutionary origin. a)Stacking weight of the OU and BM distributional models.b)Posterior distribution of rhythm in the OU distributional model.c)Posterior distribution of sigma, the scale of the drift process.d)Posterior distribution of alpha, the strength of selection.e)Posterior distribution of the phylogenetic half-life.f)Visualization of reconstructed median posterior values using a maximum clade credibility (MCC) tree. The data underlying this Figure can be found inhttps://zenodo.org/records/19816728. a)Stacking weight of the OU and BM distributional models.b)Posterior distribution of rhythm in the OU distributional model.c)Posterior distribution of sigma, the scale of the drift process.d)Posterior distribution of alpha, the strength of selection.e)Posterior distribution of the phylogenetic half-life.f)Visualization of reconstructed median posterior values using a maximum clade credibility (MCC) tree. The data underlying this Figure can be found inhttps://zenodo.org/records/19816728. https://doi.org/10.1371/journal.pbio.3003798.g003 In this exploratory study, we report an optimum rhythm for acoustic communication in birds and mammals with credible generalization to additional clades (insects, anurans, fish), challenging the idea that it would be shaped by biomechanical constraints. A biomechanical influence would predict distinct rhythms between masticating and non-masticating species, along with a strong negative allometric relationship between rhythm and weight in masticating species. Instead, our analyses show that rhythm evolution cannot be explained by weight, environmental pressures, or social complexity (Figs 2candS5d). This surprising absence of effect indicates that biomechanical constraints exert only a limited influence on acoustic communication rhythms at the macroevolutionary scale. By contrast, control analyses on the signals’ dominant frequency (S2 Fig) reconfirmed that spectral characteristics primarily depend on the specifics of the production organ, which chiefly vary with the animal’s weight [11]. Given the minimal contribution of physical characteristics, the lack of explanatory power from environmental and social variables, and the diversity of auditory structures and production mechanisms across taxa, it is unlikely that this shared rhythm results from similarities in anatomy alone. A key insight comes from comparing potential evolutionary scenarios for rhythm. The phylogenetic analyses clearly favored an OU model over a BM model, indicating strong stabilizing selection rather than unconstrained divergence (Fig 3). This result strengthens the view that species-specific factors, although capable of introducing local variation, do not account for the large-scale conservation of rhythm across distant taxa. With regard to its evolutionary origins, the presence of a similar rhythm across birds and mammals suggests that it was already present in their last common ancestor roughly 340 million years ago (Fig 3). Although insects, amphibians, and fishes are represented by fewer species in our dataset, the presence of slow rhythms in these taxa may hint at an even earlier emergence (S7 Fig). While other factors such as temperature or reverberation are also known to influence communication rhythms in ectotherms, such short-term and context-specific effects are unlikely to generate the macroevolutionary pattern observed here. These findings prompt the consideration of alternative explanations, potentially rooted in neural rather than anatomical, environmental or social factors. As neural mechanisms are essential for survival, they tend to be highly conserved across species and may offer a more plausible basis for a widespread optimal rhythm in communication. Although the size of neurons varies across species, their conductance is kept relatively constant through compensatory variations in ion channel density [22]. This results in conserved time constants for neural responses, including oscillatory phenomena [23]. The conserved rhythm found here around 2.8 Hz (95% interval 0.45–4.99 Hz) best matches delta brain oscillations (1–4 Hz). Interestingly, delta-range oscillations have been described across a wide range of taxa, including mammals, birds, reptiles and insects. In mammals, conserved oscillatory patterns have been reported from humans, primates, carnivores and rodents, with delta among the most stable across species [23]. In reptiles, such as turtles, delta activity dominates brain rhythms and supports ancient functions such as selective attention [24]. Even in birds and insects, delta oscillations have been observed during sleep or quiescent states, indicating that slow oscillatory activity is widespread throughout the animal realm [25]. Delta oscillations support slow perceptual integration and are engaged during active sensing behaviors such as sniffing or blinking [26], suggesting that they provide a fundamental temporal framework through which organisms sample and organize incoming sensory information. Because acoustic signals contain information distributed across both slow and fast temporal scales, as shown in our control analyses on context (S4 Fig), a conserved slow rhythm implies the need for a complementary faster mechanism capable of resolving finer temporal details. Within this neural framework, slow and fast timescales likely serve complementary functions. Delta oscillations would provide a long integration window that supports the identification of acoustic structure, while faster processes (most likely low-gamma) would enable rapid detection and fine temporal discrimination. This division of labor echoes the Fourier uncertainty principle, which states that long temporal windows allow precise frequency resolution at the expense of temporal precision, whereas short windows offer the opposite trade-off [27]. In this view, delta rhythms provide the temporal window best suited for integrating the broad acoustic patterns that carry communicative meaning, while faster oscillations handle local, transient features. Because the present study relies on large-scale comparative acoustic data, it naturally captures the slow integrative timescale rather than the fast detection range. Understanding how these slow and fast processes interact to shape acoustic communication will require dedicated neurophysiological work in future studies. The independence and parallel functioning of slow and fast scales acoustic processing is attested by human psychophysics findings. When time-compressed unintelligible speech is chunked and repackaged by interspersing silent gaps to restore regular syllable and word rates, speech becomes intelligible again. These results show that speech comprehension is limited by the packaging of information at slow rhythm, rather than by the absolute capacity to decode accelerated segmental (phonemic) cues [6,28]. This emphasizes the relevance of slow rhythms in maintaining acoustic signal decodability. Human speech operates at a mean rate that lies at the higher end of the median universal rhythm identified in this study, clustering around 5 Hz across all languages [8]. This reflects humans’ tendency to convey more information per unit of time compared to other species. Dogs, for instance, despite millennia of domestication and a demonstrated ability to understand many aspects of human language [29], process speech at a slower rhythm, typically between 1 and 3 Hz, than the human average of 5 Hz [30]. While they do not track the syllabic rhythm of speech, dogs tend to process it more globally, at the word rate, which aligns with their own vocal production rhythm (e.g., bark rate). These comparative findings highlight species-specific variations in preferred communication rhythms, while also underscoring a broader pattern: a conserved and effective temporal resolution for communication that enables interspecies sensitivity to acoustic signals. In sum, the conserved production rhythm identified here aligns closely with an equally conserved neural rhythm that facilitates the assignment of meaning to acoustic input, ultimately supporting an effective and evolutionarily robust communicative system. In conclusion, the maintenance throughout evolution of a common slow rhythm for acoustic signal production and perception de facto results in a common communication channel across coexisting species, offering possibilities for interspecies signaling (e.g., a common danger) and/or eavesdropping, and thus conferring evolutionary advantages. Spectral (frequency) and temporal (rhythm) features, both key to effective communication, seem to have followed different trajectories. While spectral traits diversify with the production and hearing organ anatomy, acoustic rhythms might be shaped by basic neural factors essential for survival that have led to a slow conserved optimum. Whether these neural factors were maintained because they amounted to an optimal communication design, or whether this design arose from recycling vital and highly conserved neural traits (neural rhythmicity via maintained neuronal conductance) remains to be elucidated. A further interesting avenue will be to probe for an optimum that combines fast and slow timescales in acoustic communication, alongside testing the delta wave hypothesis by assessing whether interspecific variation in rhythm is accompanied by corresponding variation in neural timescales. To perform an extensive phylogenetic comparison on a balanced phylogeny, we collected acoustic and biological data for at least one species per infra-order of tetrapods, when data were available, as well as a few species of insects and fishes, to obtain a good representation of rhythm throughout the phylogeny. Acoustic sequences were gathered from public and private databases (Fonozoo, Cisro, Berlin Museum fur Naturkunde and Xeno-canto), online videos platform (Youtube, Dailymotion), and from different research groups that kindly shared audio files. Following a cross-species literature survey [31–34], we defined a sequence as a recording of an acoustic display produced by a single individual, containing more than two calls separated by less than two seconds of silence. For computational purposes, we only retained sequences lasting more than one second. Data collection then followed a strict and standardized protocol. For each order, we first screened public databases and ranked species by the number of available sequences. We then selected the first species fulfilling the following criteria: at least five clean sequences (non-noisy, with a single individual vocalizing at a time) originating from different individuals. Once such a species was identified, we proceeded to the next infraorder. When no species meeting these criteria was available, we extended the search to private databases and directly contacted researchers. If no data could be obtained, that infraorder was excluded. Only recordings with minimal background noise and a clearly identifiable focal individual were retained; any containing rhythmic background noise, overlapping individuals, or ambiguous sound sources were excluded. This careful filtering minimized the risk of confounding effects and ensured that the extracted acoustic envelopes reliably reflected the focal signal across different sources. The final dataset comprised 98 species: 58 birds, 28 mammals, 4 amphibians, 4 insects, 1 reptile, and 1 fish. Species with fewer than five sequences were included only when they represented the sole available species for their infraorder. Weight presents an allometric relationship with morphological features and physiological processes involved in acoustic productions across various species. Thus, we collected biological data on the mean weight for each species, to serve as a proxy for heart rate, breathing rate, and metabolism. This involved calculating the average weight by considering both the minimum and maximum weights recorded for each species, irrespective of sex. These data were primarily obtained from the handbook of mammals of the world [35], and the handbook of birds of the world [36]. When data were unavailable from these sources, we looked for reference articles. Just as masticatory abilities may have influenced rhythm in mammals, a similar proposition could be made regarding beak morphology in birds. Unlike other animals, the morphological traits of a bird’s beak do not consistently adhere to an allometric relationship with its weight [37]. Thus we collected information on beak length, width and depth of our species. As these measures were not available for all species, and only applicable to birds, we could not include them in the phylogenetic model. We therefore built an additional linear mixed model investigating the variation of rhythm, including phylogenetic class and order as random effects to account for evolutionary influence, and beak length, width, depth, and their interactions as fixed effects. Data are available in the study github. As environmental conditions can impact vocal communication [38], we also collected data on the typical habitat of each species. We used a five level categorical classification, with habitats being either classified as closed (defined as habitats with heavy tree coverage), semi-closed (defined as habitats with light three coverage or human cities), open (defined as fully open habitat with no three coverage or obstacle), shore (for species living near a significant amount of water such as lakes, rivers or seas) or water (for species living below the water surface). These data were primarily obtained from the handbook of mammals of the world, and the handbook of birds of the world. When data were unavailable from these sources, we looked for reference articles. Data and linked references are available in the study github. As some have proposed that rhythmic communication in vocalizing animals may be linked to mastication regime [14–16], we also classified each species according to their mastication status (yes or no). As social complexity increases, individuals may need to communicate more information in a given time, and therefore speed up their communication. As communication signals have been linked to social complexity [20], we also gathered species vocal repertoire complexity (number of distinct calls in the species vocal repertoire) when this information was available. As this measure was not available for all species, we could not include it in the phylogenetic model. We therefore built an additional linear mixed model investigating the variation of rhythm, including phylogenetic class and order as random effects to account for evolutionary influence, and vocal complexity as fixed effect. These data were primarily obtained from the handbook of birds of the world, and reference articles. Data and linked references are available in the study github. To quantify rhythm in these acoustic sequences, we decided to adapt the method developed by Tilsen and colleagues to compute rhythm in human speech production [39]. This method uses the signal amplitude to automatically compute the rhythmic component of a sequence, without making any assumption on the components’ size, and is thus widely applicable across all species regardless of variations in unit size or spectral characteristics. First, we denoised the sequences using a first-order Butterworth filter, with a bandpass filter between a minimal frequency (minF), defined as 200 Hz below the minimum frequency of the animal call, and a maximum frequency (maxF), defined as 200 Hz above the maximum frequency of the animal acoustic signal obtained from reference articles. When this information was not available, we applied a large range filter with a 100 Hz minF and 10,000 Hz maxF. We then computed the normalized envelope of the denoised sequences using the Hilbert Transform. Next, we low-pass-filtered this envelope with a fourth-order Butterworth filter with a 20 Hz cut-off frequency to obtain the slow changes in acoustic energy. Before further analysis, we downsampled the resulting signal at 150 Hz for computational purposes, and then applied a continuous wavelet transform using the Morlet wavelet to obtain a time-frequency representation of the amplitude envelope. We replaced the Fourier transform with a wavelet transform, to allow for more flexibility with regards to the variation in sequence length present in our dataset. We finally analyzed that representation’s power spectrum to extract the five frequency peaks of highest amplitude in the power spectrum and used the time-frequency representation to select the main rhythmic component conserved across the entire sequence. To assess the validity of the proposed methodology, we conducted a comparative analysis between the calculated vocal rates of a subset comprising 10% of our database and those derived from two widely accepted conventional approaches: (1) by counting the number of elements per second (Count) and (2) by computing the inter-element interval (IEI). The three methods gave sensibly similar results (Fig 1d:F2.171= 0.33,p= 0.72), hence validating our rhythm quantification method. As further control analyses, we quantified recording durations in seconds and SNR in decibels. To control for the effect of both factors and their interaction on rhythm, we build a linear mixed model investigating the variation of rhythm including group and order as random effects, and SNR, length and their interaction as fixed effects. To determine dominant frequencies, which unlike the fundamental frequency are measurable in all types of communicative signals [40], we isolated the first acoustic unit in each denoised sequence. We then applied a single discrete Fourier transform to compute the power spectrum of these units, and extract the peak of highest amplitude. The obtained results were also visually controlled in Praat, to make sure that the extracted dominant frequency matched the acoustic energy present in the unit. If the first unit had poor SNR leading to inaccurate computation of the dominant frequency, we selected the next unit in the sequence. While the existing literature highlights the importance of context and its correlated arousal levels on acoustic signal rate [1], for most species we were not able to obtain these data. Nevertheless, whenever possible we selected recordings of different call types for each species. Further, we performed separated analysis of variance (ANOVA) to control for the effect of call type on rhythm in three species: one avian and two mammalian, from which we could obtain different call types, including contact calls, alarm calls, songs, agonistic and antagonistic vocal displays. To test our hypotheses of interest we used phylogenetic comparative methods (PCM), a broad family of methodological tools for characterizing and controlling for the evolutionary dynamics thought to give rise to the data under study. PCMs require a representation of the relatedness of the taxa under study in the form of a phylogenetic tree sample. To represent the tree topology of the species we performed a phylogenetic analysis based on comparable genetic sequences, using a Bayesian framework to infer a posterior tree sample. We first matched each species in the sample with their closest genetic proxies in GenBank [41], and extracted mitochondrial DNA for the corresponding species. For 54 of the 98 species, matching mitochondrial genomes were available from literature and deposited in GenBank. For the 44 remaining species we chose proxies from another closely related species. We took a species within the same genus when possible; if this was not available, we chose species within the family of the target species, after confirming that no more than one species per family was included in the original list. Only for four target species—Correlophus ciliates(Squamata),Galbula ruficauda(Aves),Leptosomus discolor(Aves),Phaethon rubricauda(Aves) —we did not find proxies within the family and we had to find a proxy within the order. To choose the best mtDNA proxy with those deposited in GenBank, we considered completeness of the available mitochondrial sequences and comparable average size, weight, and environment of the target species. Maximum missing data is 100 base pairs, for an average size of 16706 base pairs. MtDNA genomes were aligned with MAFFT software [42] and standard settings. The alignment was manually screened in BioEdit (version 7.2,https://thalljiscience.github.io/) for spotting irregularities and potential outlier sequences. Sequences were then cut to keep only the coding region, which is more conserved across species, using theHomo sapienssequence as a reference. The final alignment consisted of 21860 base pairs, which include large INDELs sections to accommodate alignment between the most divergent species (e.g.,Apis mellifera). We used BEAST2 to generate the trees, running 10′000′000 iterations of Markov chain Monte Carlo (MCMC) with a thinning interval of 1,000. We used the following settings to approximate the broad evolutionary range of the species considered: assuming an HKY substitution model, a strict clock (Uniform rates across branches), and a Birth-Death tree prior with a Yule birth rate. This resulted in 10′000 trees, of which we use 50 for phylogenetic comparative analyses. To assess the impact of several predictors of interest on different properties of variation in vocal rhythm and dominant frequency, we used phylogenetic regression modeling, a comparative method that assesses the effect of predictors on a response while controlling for the phylogenetic relatedness of the taxa. Due to heterogeneity in the number of datapoints and individuals in each species, we employed both non-distributional regression models, which model the mean of the response variable as a function of predictors, and distributional (scale-location) regression models, which model both the mean and standard deviation of the response variable as a function of predictors. We control for species-level idiosyncrasies in both the median and (in some cases) standard deviation of rhythm via phylogenetic random intercepts and slopes (phylogenetic random effects are similar to the standard random effects used in hierarchical regression modeling, but are generated by a Gaussian Process with a covariance kernel that is a function of the phylogenetic patristic distances between species under study rather than independently and identically distributed with diagonal variance). We fitted four phylogenetic regression models using brms [21] for each response variable (vocal rhythm, dominant frequency), resulting in eight models. Due to the constraints of bayesian multilevel modeling, particularly its sensitivity to missing data across hierarchical levels, we restricted the main model to predictors available for the majority of species. Specifically, we included average species weight (log-transformed, centered, and standardized) [43], mastication status, and species’ living environment. We also modeled the interaction between mastication status and weight. Additional predictors that were only available for a limited subset of species were analyzed separately in independent linear mixed models to avoid excessive data imputation or loss of statistical power. Our distributional models have the following basic generative process (below,is shorthand for all model predictors, including fixed and random effects): Non-distributional models have the following structure: We employ the default priors of brms. The first of the four models was a full distributional one that modeled both the expected median and variance of the response variable as a function of these predictors, while controlling for species-level idiosyncrasies in both the median and variance of rhythm via phylogenetic random intercepts and slopes. The second was a full non-distributional model that treated only the expected median rhythm as a function of the predictor variables as well as phylogenetic random intercepts and slopes. The third of these was a null distributional model that included only phylogenetic random intercepts and slopes for the expected median and variance. The final model was a null non-distributional that included only phylogenetic random intercepts and slopes for mean rhythm. We ran each of these models for 4,000 iterations of the no U-turn sampler over 4 chains with a log-normal link function and discarded the first half of samples, aggregating posterior samples across the retained sampled trees. Models in brms have the following formulae: Full, distributional Full, non-distributional Null, distributional Null non-distributional We compared fitted models via their leave-one-out ELPD values [44] and stacking [45], which average predictive distributions of different models to generate weights representing their relative predictive power. We used the function loo_compare to measure differences in ELPD across models. Finally, we inspected posterior distributions of regression coefficients of the full distributional model to assess the effects of predictors of interest. We further investigate the properties of rhythm across species using two Gaussian Process models of continuous trait evolution, asking specifically whether the evolution of vocal rhythm is characterized by a random process of drift (characterized by BM) or whether selective forces draw rhythm values toward an optimal value over time (a mean-reverting scenario characterized by an OU process). Under BM, the displacement of a continuous trait at timeshas a variance proportional to the amount of time elapsed over the course of displacement (below denoted ast), whererepresents the scale of the drift process: Under an OU process, the displacement of a character has the following formula: In the first component of the sum,represents the strength of selection to the optimal value. The second component represents a process of BM, withrepresenting the scale of drift. Thus, the OU process allows for both selective and random forces in character evolution. An standard way to interpretis to transform it to the phylogenetic half-life, ln 2/α[45]. This is interpreted as the average time for a trait to evolve halfway from an ancestral state toward a new optimum, indicating how long it will take before adaptation to a new regime is more influential than constraints from the ancestral state. If half-life values are greater than the height of the phylogeny (1 in our case, as the tree length is scaled to unit height), the process increasingly resembles BM and involves a slower adaptation speed. As above, we employ a distributional approach, allowing species-level mean rhythm values and species-level standard deviations of rhythm values to evolve over the phylogeny according to BM or OU processes. The distributional BM process has the following generative process: represents the trait value at the root of the tree, whileis a matrix of the shared history (the time between the root age of the tree and the most recent common ancestor) of each pair of nodes in the tree andis the positive scale of drift. Conventions are as above. The distributional OU process has the following generative process: represents the trait value at the root of the tree,is the positive scale of drift,αis the positive strength of selection, and Δ is a matrix of pairwise cophenetic distances between species in the phylogeny, scaled to a maximum distance of 1. Conventions are as above. We placeNormal(0,1)priors over unconstrained parameters andGamma(1,1)priors over positive parameters. In addition to running these models on all species in our sample, we validate results by running models on bird and mammal species alone. Rhythm values were reconstructed to internal nodes of the maximum clade credibility (MCC) tree of the phylogeny, using ggtree package [46] by drawing 10 draws from each of the posterior distributions inferred from the 50 different trees in the tree sample and sampling values at internal nodes of the tree from the normal distribution parameterized by the OU process, conditioned at the expected tip values (Fig 3f). All analyses and visualization were done using Stan and R version 4.1.2 (2021-11-01) with the following packages Seewave [43], Soundgen [47], DoBy [48], Lme4 [49], MuMYn [50], brms [21], ggplot2 [51]. https://doi.org/10.1371/journal.pbio.3003798.s001 (DOCX) a)Rhythm (Hz) as a function of signal-to-noise ratio (SNR) showing the absence of relationship between SNR and Rhythm (t= −0.21,p= 0.84,R2= 0.02).b)Rhythm (Hz) as a function of recording length showing the absence of relationship between sequence length and rhythm (t= −1.45,p= 0.15,R2= 0.02). The data underlying this Figure can be found inhttps://zenodo.org/records/19816728. https://doi.org/10.1371/journal.pbio.3003798.s002 (TIF) a)Leave-one-out expected log pointwise density difference (ELPD) between the full and the null distributional models.b)Stacking weight of the models.c)Posterior credible interval (95% and 85%) of the full model.d)Dominant frequency plotted on a logarithmic scale as a function of log-transformed weight with predicted slopes from the full distributional model and their standard error. The data underlying this Figure can be found inhttps://zenodo.org/records/19816728. https://doi.org/10.1371/journal.pbio.3003798.s003 (TIF) a)Leave-one-out expected log pointwise density difference (ELPD) between the null distributional (“dist”) model and the others (“ndist” for “non-distributional, modeling only the mean”).b)Stacking weight of the models. The data underlying this Figure can be found inhttps://zenodo.org/records/19816728. https://doi.org/10.1371/journal.pbio.3003798.s004 (TIF) a)Leave-one-out expected log pointwise density difference (ELPD) between the null and the full distributional (scale-location) models.b)Stacking weight of the models.c)Posterior credible intervals (95% and 85%) of the full distributional model (W*M = Weight*Mastication).d)Rhythm plotted as a function of log-transformed weight with predicted slopes from the full distributional model and their standard error for masticating and non-masticating species.e)Distribution of raw median rhythm (in Hz) across birds and mammals. Each vertical black line represents the median rhythm of a single species. The data underlying this Figure can be found inhttps://zenodo.org/records/19816728. https://doi.org/10.1371/journal.pbio.3003798.s005 (TIF) a)Rhythm in sequences of different call types (1 = affiliative grunt, 2 = scream, 3 = threat grunt) in olive baboons(Papio anubis).b)Rhythm in sequences of different call types (1 = bark, 2 = growl, 3 = howl, 4 = snarl, 5 = whine) in dogs (Canis lupus familiaris).c)Rhythm in sequences of different call types (1 = alarm call, 2 = flight call, 3 = song) in Eurasian stone-curlews (Burhinus oedicnemus).d)Rhythm plotted as a function of vocal repertoire size.e)Rhythm plotted as a function of beak length in birds.f)Rhythm plotted as a function of beak depth in birds.g)Rhythm plotted as a function of beak width in birds. The data underlying this Figure can be found inhttps://zenodo.org/records/19816728. https://doi.org/10.1371/journal.pbio.3003798.s006 (TIF) a,f) Stacking weight of the OU and BM distributional models. b,g) Posterior distribution of rhythm in the OU distributional model. c,h) Posterior distribution of sigma, the scale of the drift process. d,i) Posterior distribution of alpha, the strength of selection. e) Posterior distribution of the phylogenetic half-life. e,j) Visualization of reconstructed median posterior values using a maximum clade credibility (MCC) tree. The data underlying this Figure can be found inhttps://zenodo.org/records/19816728. https://doi.org/10.1371/journal.pbio.3003798.s007 (TIF) a)Stacking weight of the OU and BM distributional models.b)Posterior distribution of rhythm in the OU distributional model.c)Posterior distribution of sigma, the scale of the drift process.d)Posterior distribution of alpha, the strength of selection.e)Posterior distribution of the phylogenetic half-life.f)Visualization of reconstructed median posterior values using a maximum clade credibility (MCC) tree. The data underlying this Figure can be found inhttps://zenodo.org/records/19816728. https://doi.org/10.1371/journal.pbio.3003798.s008 (TIF) The authors are grateful to the CSIRO Australian National Wildlife Collection, (https://ror.org/059mabc80), Fonoteca Zoologica (https://www.fonozoo.com/), the Museum für Naturkunde of Berlin (https://www.museumfuernaturkunde.berlin) and the Xeno-Canto Foundation for Nature Sounds (https://xeno-canto.org) for their assistance in undertaking this research. The authors thank all the researchers from the NCCR Evolving Language as well as Pascal Belin and Steffen R.Hage for their insights on the project. The authors would like to thank Aaron Bauer, Andrew Spencer, Alex Kwet, Alex Rohtla, Camila Ferrara, Daniel Blumstein, Élodie Briefer, Émilie Genty, Frank Lambert, Isabelle Charrier, Gerald Carter, Hans Schneider, Marc D. Hauser, Nikola Falk, Peter Boesman and Robert Seyfarth for providing us with recordings for this study.

Noble Corporation, a U.S.-based offshore drilling giant, has lined up a new multi-well drilling assignment for one of its semi-submersible rigs with the UK-headquartered energy giant BP on the UK Continental Shelf (UKCS). Noble has secured a three‑well drilling contract with BP for the 2016-builtNoble GreatWhiteharsh environmentsemi-submersible rig, which has been renamedNoble Claus Bachmann. As a result, the rig will embark on a drilling campaign off the coast of the United Kingdom next year. The deal, which has an estimated duration of 150 to 210 days, comes with a day rate of $320,000, plus a $5 million mobilization fee. With a maximum drilling depth of 35,000 feet (10,670 meters), the semi-sub is capable of working in water depths of 9,840 feet (3,000 meters). The rig owner explains that the work is expected to begin in Q2 2027, directly preceding the firm’supcoming contractwith Aker BP in Norway, supporting continuity of operations in the region. Blake Denton, Noble’s SVP of Marketing & Contracts, commented:“This contract represents expansion of our scope with a key client ahead of operations in Norway with Aker BP. “We are equally proud to honor Claus Bachmann’s enduring impact on Noble by renaming the Noble GreatWhite to the Noble Claus Bachmann before operations commence. This recognition underscores a tenure defined by consistency, trust, and sustained performance.” Following a series of deals totaling approximately $565 million, Noble’scontract backlogon April 27, 2026, was $7.5 billion, excluding mobilization and demobilization revenue. Take the spotlight and anchor your brand in the heart of the offshore world! Join us for a bigger impact and amplify your presence at the core hub of the offshore energy community!

Canarias Ahora El puerto de Las Palmas realizó en la noche de este lunes el primer suministro de gas natural licuado (GNL) alfast ferry Mercedes Pintode Baleària Canarias, que se estrenó el pasado viernes en la ruta entre Tenerife, Gran Canaria y Fuerteventura para reforzar la conectividad en el archipiélago. La operativa supone todo un hito para el transporte marítimo en las islas, ya que se trata del primer suministro de GNL que se realiza en el Puerto de Las Palmas y el primero con cisternas que Baleària realiza a un buque en Canarias. El abastecimiento se llevó a cabo en el muelle Juan Sebastián Elcano mediante el sistema MTTS (Multi Truck To Ship), que permite el suministro simultáneo desde varios camiones cisterna hasta los tanques del buque. En concreto, la operación se realizó con dos cisternas y un suministro de 600MWh, bombeando ambas unidades de forma simultánea. El caudal de bombeo fue de 600 l/min. por cisterna, consiguiendo un caudal de bombeo total de 1.200 l/min. Según ha informado la Autoridad Portuaria de Las Palmas, esta operativa refuerza la capacidad técnica y logística del Puerto para acoger suministros de combustibles más ecoeficientes, y en esta línea, Baleària Canarias da un paso más en su compromiso con la ecoeficiencia, la innovación tecnológica y la mejora continua de sus servicios en el archipiélago. El director general de Baleària, Georges Bassoul, señaló que “con esta operación de suministro, Baleària Canarias da un primer paso para impulsar el desarrollo y suministro regular de estos combustibles sostenibles en el archipiélago, con el objetivo de seguir avanzando hacia una movilidad marítima más eficiente y ambientalmente responsable” La presidenta de la Autoridad Portuaria de Las Palmas, Beatriz Calzada, agradeció a Baleària su confianza en el Puerto de Las Palmas y subrayó: “Para nosotros, esta operación supone el pistoletazo de salida en suministro de combustibles menos contaminantes. Se trata de una actividad que hemos declarado estratégica para nuestro puerto y debemos poder suministrar cualquier tipo de combustible que los barcos necesiten y demande el mercado”. Con una inversión de 128 millones de euros, elMercedes Pintose posiciona como un referente en ecoeficiencia dentro del transporte marítimo de alta velocidad. El buque está equipado con motores duales que permiten el uso de gas natural (GNL) y biogás, un combustible neutro en emisiones de CO₂. Cabe recordar que el grupo Baleària mantiene desde hace años una apuesta firme por el gas natural como combustible de transición en el transporte marítimo, ya que permite reducir las emisiones de CO2 y NOx, así como eliminar las emisiones de azufre y partículas, avanzando hacia una movilidad marítima más eficiente y ambientalmente responsable. Además, ha empezado a usar en determinadas rutas en biogás, un combustible que permite operar con cero emisiones netas, en el marco de su estrategia de descarbonización.

La Autoridad Portuaria de Las Palmas (APLP) y la empresa de transporte marítimo Baleària Canarias han completado con éxito el primer suministro de gas natural licuado (GNL) a un buque en el archipiélago canario. Esta operación pionera se ha desarrollado en las instalaciones del muelle Juan Sebastián Elcano, situado en el Puerto de Las Palmas, y ha tenido como destinatario al fast ferry Mercedes Pinto, un moderno barco de la citada naviera. Según ha detallado la propia institución portuaria en un comunicado oficial emitido este martes, este hito representa el primer abastecimiento de GNL con cisternas que se lleva a cabo dentro del recinto portuario de la capital grancanaria. Asimismo, se trata de la primera operación de estas características que la compañía naviera ejecuta en las islas, consolidando su apuesta por la innovación tecnológica e infraestructural. Para realizar este complejo procedimiento, los técnicos han empleado el avanzado sistema logístico conocido como Multi Truck To Ship (MTTS). Este método de vanguardia permite el suministro simultáneo de combustible desde varios camiones cisterna directamente hasta los tanques de almacenamiento de la embarcación, lo que optimiza sensiblemente los tiempos de recarga y mejora la eficiencia operativa de los muelles. En concreto, el proceso de repostaje marítimo se ha llevado a cabo conectando dos unidades de transporte por carretera y ha alcanzado un suministro total de 600 MWh de energía. El bombeo desde ambas cisternas se realizó de manera paralela con un flujo de 600 litros por minuto en cada unidad. Gracias a esta sincronización, los operarios lograron alcanzar un caudal de bombeo de 1.200 litros por minuto, lo que demuestra la alta capacidad de las infraestructuras empleadas. La presidenta de la Autoridad Portuaria de Las Palmas, Beatriz Calzada, ha expresado su agradecimiento por la confianza depositada por la naviera y ha asegurado que esta exitosa operación supone "el pistoletazo de salida en suministro de combustibles menos contaminantes". De igual modo, ha querido reivindicar el papel de las instalaciones al recordar que el puerto es líder en servicios de bunkering en el Atlántico Medio y ocupa el segundo lugar en todo el sistema portuario español. Según la presidenta, se trata de una actividad que han calificado de "estratégica" y en la que continúan trabajando con la meta de almacenar y suministrar GNL de forma habitual. Por su parte, el director general de Baleària, Georges Bassoul, ha subrayado que con este repostaje la empresa da "un primer paso para impulsar el desarrollo y suministro regular de estos combustibles sostenibles en el archipiélago". En la misma línea, el directivo ha dejado claro que el gran objetivo de la naviera pasa por "seguir avanzando hacia una movilidad marítima más eficiente y ambientalmente responsable", adaptándose así a las exigencias del mercado y los nuevos retos del sector comercial. El buque protagonista de esta jornada, el Mercedes Pinto, ha requerido una inversión de 128 millones de euros para su construcción y puesta en servicio. Esta embarcación se ha incorporado recientemente a la flota para cubrir la ruta que conecta las islas de Tenerife, Gran Canaria y Fuerteventura. Entre sus especificaciones técnicas más destacadas, el buque está equipado con motores duales de última generación que permiten tanto el uso de GNL como de biogás, un recurso calificado como un combustible neutro en emisiones de dióxido de carbono que refuerza el compromiso empresarial con el medio natural y el mercado libre.

While visiting family in St. Petersburg, Florida, in November 2024, I found myself walking down a quiet residential street in Shore Acres, a low-lying, bayfront neighborhood not far from where I grew up. Two months earlier,Hurricane Helenehad sent several feet of water into homes here, even though the center of the storm had stayed far offshore. Just days after Helene,Milton made landfallnearby as a major hurricane, inflicting substantial wind damage. What I saw on that autumn morning was a scene of starkly unequal neighborhood recovery: Dozens of older homes, most built during the area’s postwar building boom, were in a state of shocking disrepair. Shattered drywall, warped kitchen cabinets, broken glass – entire interiors poured out into the street in piles, at times as high as I am tall. On the same street, I also saw pristine newer homes that looked untouched. Raised on posts far above their neighbors – in line with newer building codes – there was no sign that a major storm had recently clawed through the neighborhood. As the sound of buzz saws and hammers rang in my ears, I noticed “for sale” signs in front of many storm-damaged homes. Building back after a storm is a trying business, and it appeared some had called it quits. For over a decadeI’ve been researchinghow property markets adapt to the changing financial realities of climate change in Florida, the Netherlands and beyond. On this street, I saw a community slowly unraveling as climate shocks – and the subsequent market responses to them – have reshaped the cost of living. These costs are driven by more than major disasters. Soaringproperty insurance ratesare repricing life in the Sunshine State, and this is likely to worsen as hurricanes intensify in the coming years. But the current way we manage these risks and costs isn’t the only option. In Shore Acres and elsewhere, the climate crisis becomes tangible when you receive your annual homeowners insurance bill. Sustained, year-over-year price increases have beenwell documentedin Florida and other states. It isn’t uncommon tohear local news storiesabout neighbors who are seriously struggling to keep up with insurance payments. Nor is it surprising to learn that a neighbor has been dropped by yet another insurer, or that they’re being asked to replace their roof if they want coverage from a new carrier. These seemingly mundane experiences reflect the structural importance of insurance in ourhousing finance and climate risk management systemsin the United Statesandbeyond. In Florida, where individual and collective fortune is built on property value and housing markets,insurance markets are particularly vital. Without insurance, you and I wouldn’t be able to get – or keep – a mortgage. Without mortgages, buyers would lose access to a market, causing home prices to fall. The real estate market, and all the jobs wound up in it, would stutter. The property tax base would fall, and with it, local governments’ budgets would drop. A downward financial spiral ensues. Even those who can afford to stay suffer losses in home equity and lifestyle as the community around them disintegrates or disappears. This played out during the2008 subprime mortgage crisisand subsequent economic recession. And this isn’t just a problem for homeowners – costly insurancealso affects rentersandaffordable housing providers, as landlords pass on costs to tenants, defer maintenance, postpone new building or face financial distress. Some individuals, neighborhoods and cities can afford to pay their way out of this spiral. They can build costly infrastructure to mitigate storm damage, absorb losses and rebuild after a disaster. Case studies from Miamishow how climate risk already is beginning to sort neighborhoods by wealth, resilience and insurability. High insurance costs could further push affordable homeownership out of reach in places like Miami Gardens and similar communities, where housing costs are on the rise. Meanwhile, investors are buying elevated land that is less likely to flood in communities like Little Haiti, displacing communities and limiting their access to affordable housing. Homeowners rely on insurance topool risksso that no individual absorbs the full cost of a shock. But as for-profit insurers look to protect themselves from growing losses, they necessarily become much more selective about who gets protection and at what cost. But this piecemeal, property market-driven form of adaptation defers a larger and more expensive collective reckoning: What happens if larger numbers of residents can no longer afford to stay, or otherwise decide the risks are too high, andmove elsewhere? Where do they go, and what becomes of the places they leave behind? In other words, risk sharing becomes risk sorting. And without strong mechanisms to counter this, a split occurs in places like Shore Acres. Florida’s coastal communities already are showing signs of this “splintering protectionism” – a patchwork of individuals and neighborhoods that are financially protected or excluded under growing climate risk. These patterns oftenrecall and reinforcehistoric forms ofracial and economic injustice in Florida housing marketsandmore broadlyin the U.S. One immediate response would be to create public policies that make adaptation and insurance work better together. In other words, homeowners need help both with storm-proofing their homes to reduce damage up front and with paying to repair and rebuild when necessary after a storm. The state of Florida has gradually built acomplex systemof semipublic insurance institutions, but it hasn’t meaningfully tackled resilience at the home and neighborhood scale. Efforts to stimulate private financial market solutions for homeowners haveproven challenging, in part because individuals and private markets cannot coordinate comprehensive community adaptation strategies. ManyU.S. and international reform proposalsfocus on linking insurance backstops, such as expanded public insurance options, to concrete measures that stimulate home- and community-level resilience, including stronger building codes, home retrofitting, new infrastructure and better spatial planning. These proposals recognize that leaving decisions about adapting houses located in vulnerable areas up to individual homeowners is ineffective. Those who can afford it may make expensive updates to their homes if they think the risks are high enough, but those who can’t afford it are simply out of luck. Public institutions like thehousing resilience agenciesproposed by theClimate and Community Institutecould help connect insurance and adaptation in new ways. Extensive international case studies also provide a richbasis for reimaginingour insurance and resilience institutions. Local, state or even federal versions of these agencies could offer consumer insurance for individuals alongside adaptation investment programs as a one-stop shop. These agencies could pool risks through a single-payer insurance system and reduce those risks through investment in resilience measures. As government agencies, they would be focused on long-term safety and affordability rather than making a profit. Such agencies could also incorporate transparent and democratic decision-making, giving more power to communities over decisions that are typically “black-boxed” by private market actors. Regardless of how Florida chooses to move forward, insurance reform debates should not lose sight of thesefundamental questions: What, and who, are we trying to protect, on what time horizon, and at what costs? The current system is already answering these questions, deciding the fate of Shore Acres and any number of similar communities. The risk-sorting dynamic that’s driving adaptation is also opening new and deepening existing financial fault lines in neighborhoods like this. On one side are those who can afford high insurance rates and the costs of protective measures, such as storm shutters. On the other side are those who can’t afford insurance or to rebuild their ruined homes after a storm. I believe Florida’s challenge is not simply to stabilize insurance markets, but to create new forms of collective protection that connect finance, risk reduction and decisions about how communities live with climate risk.

MARTINA MISER En la cubierta de un barco bateeiro, con el mar de A Illa a sus espaldas, los convocantes de la manifestación que el próximo domingo se celebrará en etsa localidad arousana explicaron este martes las claves de esa protesta. Será una movilización por tierra y por mar. Por tierra, confían en que sean miles de personas las que a mediodía se reúnan en la entrada de la localidad, en O Bao, para caminar juntas hasta el muelle de O Xufre. Por mar, aspiran a reunir medio millar de embarcaciones. No será un reto fácil de cumplir: los mariscadores que se han acogido al cese de actividad han tenido que entregar los roles de sus lanchas y no podrán sumarse, por tanto, a la protesta. Quienes sí estarán serán numerosas embarcaciones tradicionales de la ría. Así lo aseguró Salvador Allo, que habló en nombre «das embarcacións e dos oficios tradicionais, que están desaparecendo». Allo contó la vieja historia de dos dornas cuyos patrones, en un día de mal tiempo, tomaron dos decisiones distintas: uno actuar para salvarse de la tormenta, mientras el otro se enrocaba en un «deixa ir». «Deixar ir nunca chegou a terra», dijo Salvador Allo. Y es que inhibirse de los problemas, fingir que no están ahí, no evita que se sufran las consecuencias. Así que hizo un llamamiento a la ciudadanía de toda la ría para salir a movilizarse. «As empresas, se queren estar aquí e usar o noso», teñen que respectalo, sentenció. Antes de que Salvador tomase la palabra, los representantes de las tres plafaformas convocantes tomaron la palabra. Por la PDRA, Xaquín Rubido, Xocas, fue contundente. La ría de Arousa ha tocado fondo «ante a pasividade absoluta da Xunta de Galicia». A lo largo de los últimos años, todos los pasos que se han dado desde la Administración autonómica, dijo, parecen responder a un único objetivo: cansar a la gente del mar y provocar su salida de este, creando el escenario perfecto para «cambiar o modelo produtivo» y abrir las puertas de la ría «a empresas multinacionais que explotarán os nosos recursos, xerarán máis contaminación, xerarán moitos menos postos de traballo e levarán os beneficios de aquí». Xocas utilizó los datos de producción tanto del sector marisquero como del bateeiro para ilustrar ese decilve productivo en el que el mar arousano lleva inmerso desde hace años. Solo en los últimos tres, la caída de la producción ha sido de un 90 % en «especies senlleiras da nosa ría», como la almeja babosa, la fina o el berberecho. La prioridad debería ser ahora revertir esa situación, pero en lugar de ello, dicen desde la PDRA, la Xunta «nos quere impor proxectos contaminantes» y antagónicos: Altri, la mina de Touro-O Pino, el almacén de betún del puerto de Vilagarcía o la eleción del punto de vertido para el dragado del Lérez. Por la plataforma vecinal Mina de Touro-O Pino Non, habló Che Cancelo, quien hizo hincapié en que «se a estas alturas non en en marcha Altri e non se reabriu a mina non é porque haxa argumentos na súa contra, que os hai, e todos, se non pola forza de todo un pobo e pola presión que exerceu. Foi esa presión a que obrigou á Xunta a retrasar eses proxectos». La elección de las palabras que empleó Cancelo fue cuidadosa. «Falo de retrasar porque nin as empresas depredadoras renunciaron, nin a Xunta fixo outra cousa que dar largas». Por esa razón considera necesario que la movilización social continúe. «Non nos van cansar, porque nos estamos a xogar o noso modo de vida e o noso futuro», advirtió. De futuro también habló Juan Pedro Sánchez, presidente de la plataforma Altri Non. Señala que los vecinos y vecinas a quienes representa «nunca nos deixamos enganar cos anuncios da Xunta» que dan por cerrado el tema de la pastera. Convencidos de que la intención de la Administración es resucitar el proyecto en cuanto baje la vigilancia social, prometen mantenerla viva. «Manteremos a loita fronte ao engano», sentenció.

France-based geophysical services company Viridien, formerly known as CGG, has joined forces with Norway’s oil and gas player Aker BP to bolster hydrocarbon search by enhancing multi-client ocean bottom node (OBN) seismic data acquisition, imaging, and technology development on the Norwegian Continental Shelf (NCS). Viridien and Aker BP have entered into a strategic partnership agreement to strengthen collaboration across multi-client OBN seismic and geoscience workflows for exploration on the Norwegian Continental Shelf, with the aim of accelerating the discovery and development of additional reserves. This agreement establishes a five-year co-operation framework designed to align strategy and bring consistent effort to innovation and technology development, intending to optimize exploration efforts off the coast of Norway. Petter Sørhaug, Exploration & Reservoir Development at Aker BP, commented:“Our strategic partnership with Viridien supports Aker BP’s ambition to deliver profitable and sustainable growth on the Norwegian Continental Shelf. We have built a close and constructive collaboration with Viridien over time. This agreement is a natural extension of our cooperation, and it also follows Aker BP’s strategy to transform exploration and production through selected strategic partnerships. “Aker BP sees ocean bottom node seismic as a key enabler for improved imaging and higher-quality subsurface data, both in exploration and field development. Together with Viridien we will further strengthen our capabilities in seismic acquisition, imaging, and processing to unlock greater value from our portfolio.” The partnership, which combines the Norwegian player’s operational experience with the French firm’s imaging technology and expertise, also expands the latter’s multi-client OBN coverage to support large-scale exploration. Dechun Lin, Head of Earth Data, Viridien, underlined:“This multi-year agreement reflects Viridien’s strategy for long-term collaboration and investment in the NCS. “Our expertise in seismic imaging and local knowledge has played a significant role de-risking near-field exploration and enabling recent discoveries. We look forward to continuing working with Aker BP to develop technologies that will enhance the efficiency, cost-effectiveness, and value of OBN.” The collaboration with Aker BP comes shortly after Viridienembarked on a multi-client ocean bottom node surveyin the Central North Sea, covering both the UK and Norwegian sectors. Take the spotlight and anchor your brand in the heart of the offshore world! Join us for a bigger impact and amplify your presence at the core hub of the offshore energy community!

The International Maritime Organization’s (IMO) targets, the full integration of shipping into the EU Emissions Trading System (EU ETS) by 2026, and the FuelEU Maritime initiative are turning decarbonization into a daily challenge for shipping companies. Against this backdrop, Schneider Electric and Msystems are positioning themselves as strategic partners in energy and digital transformation. Leveraging extensive experience in maritime applications and expertise in IT infrastructure and digital solutions, the two companies focus on energy management, data transparency, and vessel lifecycle optimization. Through their participation at Posidonia 2026, Schneider Electric and Msystems are demonstrating how the convergence of energy and IT can create tangible business value, supporting the development and integration of digital infrastructure onboard vessels. The discussion around sustainability has matured. The question is no longer whether to act, but how. Energy efficiency, electrification, and the use of real-time data have become critical drivers for reducing carbon footprints while safeguarding profitability. Through maritime-focused technologies such as EcoStruxure™, Schneider Electric provides an integrated architecture that connects energy management, automation, and digital services. The result is enhanced visibility, better-informed decision-making, and meaningful reductions in both operating costs and emissions. Msystems brings years of specialized expertise in both information technology and maritime operations, delivering comprehensive onboard IT infrastructure solutions. With a focus on technologies such as micro data centers and remote management systems, the company helps ensure uninterrupted operations and operational reliability, meeting the growing demands of an increasingly digitalized maritime environment. “The rapid expansion of onboard IT systems is transforming vessels into floating data centers. From navigation to equipment condition monitoring, the need for reliable power and uninterrupted operation has become critical,” said Tasos Sarris, Channel Sales Business Developer for Southeast Europe at Schneider Electric. Schneider Electric addresses these requirements through integrated micro data center solutions, marine-grade UPS systems, and modular architectures that ensure business continuity even under the most demanding operating conditions. Reliability is no longer simply a technical specification—it is a business imperative. Modern vessels are highly digitalized and energy-intensive, making uninterrupted power supply essential. At the same time, shore power, also known as Onshore Power Supply (OPS), is evolving from an option into a requirement, as ports and vessels adapt to new regulatory standards. In this context, certified low- and medium-voltage solutions for maritime applications are playing an active role in supporting the development of the global shore-power ecosystem. “Our success depends not only on technology but also on having the right partners,” said Kyriaki Roussianou, Strategic Account Manager at Msystems. “The long-standing partnership between Schneider Electric and Msystems is a prime example, combining technical expertise, certified solutions, and a deep understanding of the maritime industry’s needs.” From micro data centers to DNV-certified UPS systems designed for demanding maritime applications, supported by comprehensive after-sales services and worldwide technical coverage, the approach takes into account the entire lifecycle of solutions, prioritizing total cost of ownership rather than isolated capital investments. As shipbuilding lead times continue to increase and regulatory uncertainty persists, strategic focus is shifting toward the modernization of existing fleets and modular upgrade solutions. Shipping companies are seeking ways to adapt their vessels to evolving requirements without relying solely on newbuild programs. In this environment, Schneider Electric supports continuous optimization through targeted upgrades in energy systems, automation, and digital infrastructure, enabling controlled investment costs and minimal operational disruption. At the same time, attention is increasingly shifting from initial capital expenditure to total lifecycle cost. Through solutions that enable predictive maintenance, improved energy management, and enhanced reliability, operators can extend the operational lifespan of vessels beyond 30 years. The result is fleets that are more resilient, more efficient, and better prepared to meet the challenges of the future. Για να εμφανίζονται περισσότερα άρθρα τηςΝαυτεμπορικήςστις αναζητήσεις σας εύκολα και γρήγορα, πρέπει να προσθέσετε το site στις προτιμώμενες πηγές σας. Μπορείτε να το κάνετε πηγαίνονταςεδώ.

COP31 hosts Turkey urged countries Tuesday to join a voluntary push to make electricity account for 35 percent of global energy demand by 2035 as it outlined its priorities for the UN climate talks. The November summit in Antalya is taking shape as the Middle East conflict roils global energy markets, exposing fossil fuel importers to price spikes and supply shortages. The electrification target unveiled in Bonn was “a flagship initiative” of COP31 that could respond to this crisis and help insulate economies from fossil fuel price shocks, the Turkish conference organisers said in a statement. Thousands of climate negotiators are in Bonn this week and next to draft agreements and lay the groundwork for the final decisions taken by political leaders at the summit due to start November 9. Turkey said raising the global share of energy demand met by electricity from roughly 20 percent to 35 percent by 2035 would speed up the shift from fossil fuels to renewable power. “By electrifying daily life, from transport to buildings and industry, we can protect families and businesses from volatile energy markets,” incoming COP31 president Murat Kurum said in a statement. The goal will not require formal agreement by the nearly 200 nations taking part in the annual talks because it is part of the voluntary program that runs alongside the binding negotiations. This so-called “action agenda” encourages countries to join non-binding pledges and other initiatives to turn commitments made at the UN-sponsored climate talks into action on the ground. – Clean switch – In simple terms, electrification means replacing technologies that burn fossil fuels directly — such as gas heating systems and diesel vehicles — with electric alternatives. But for electrification to drive down heat-trapping emissions and tackle climate change, the extra electricity must come primarily from renewable sources — rather than fossil fuels. “If you electrify and you increase coal, then what are you doing?” veteran COP observer and E3G analyst Alden Meyer told AFP in Bonn. “You do need to both expand electrification and squeeze fossil fuels out of the electricity system at the same time.” The electrification target unveiled by Turkey did not explicitly state how that extra power should be produced. In 2025, renewables reached 34 percent of global electricity generation, overtaking coal’s 33 percent share for the first time in 100 years, according to energy think tank Ember. Australia, which is steering the formal negotiations in a COP31 co-hosting arrangement with Turkey, said electrification could cut emissions and shore up energy security. “I see them as different sides of the same coin. Electrification reduces the need for fossil fuels,” COP31 negotiations chief Chris Bowen, who is also Australia’s climate and energy minister, told AFP in an interview in Bonn on Monday.

Houston-based energy player Vaalco Energy has brought back online a field off the coast of Côte d’Ivoire, following an overhaul to extend the lifespan of a floating production, storage, and offloading (FPSO) vessel deployed at the African asset. Vaalco has confirmed the restart of production from theBaobab fieldon the CI-40 block, offshore Côte d’Ivoire, following theFPSO Baobab Ivoirien’s revamp after the vesselceased hydrocarbon operationsin January 2025. George Maxwell, Vaalco’s Chief Executive Officer, commented:“We are excited that the Baobab field on the CI-40 block offshore Côte d’Ivoire has restarted production in line with our projected timeline. We have the CI-40 block license extended through 2038 and believe that there is significant development drilling upside at Baobab. “In early 2024, we had no assets in Côte d’Ivoire and now we have developed a strong position with development and exploration potential. We are at a critical junction, with successes in the Gabon drilling campaign and the Baobab field returning to production, and we believe that the remainder of 2026 will be very impactful.” This content is available after accepting the cookies. FPSO on its way back to Côte d’Ivoire as US firm continues its drilling ops in Gabon Once the nine-month refurbishment wascompleted in Dubai, the FPSO returned to Côte d’Ivoire in early Q2 2026 to be moored in position and reconnected to the field infrastructure. As a result, production has resumed from four producing wells, with the remaining three producers expected to come online shortly. The company claims that the field is performing in line with its expectations. TheFPSO refurbishmentwas undertaken to extend the vessel’s life and ensure its long-term operational capacity, as a significant development drilling program at Baobab is planned to begin in the second half of 2026. This Phase 5 drilling program is expected to include four producers, two to three injectors, and two workovers, providing potential meaningful additions to production from the main Baobab field. Maxwell emphasized:“We remain focused on execution and driving meaningful growth through our organic capital programs that we believe will translate into value for our shareholders in 2026 and beyond.” Take the spotlight and anchor your brand in the heart of the offshore world! Join us for a bigger impact and amplify your presence at the core hub of the offshore energy community!