Software Engineering News
Pi.dev reversed its anti-MCP stance and integrated MCP into its core, adding a JavaScript sandbox called Codemode for composing tool calls, from hackernews.
The Pi team previously rejected MCP publicly, but now says the protocol matured enough to warrant core integration rather than remaining an extension. The key addition is Codemode, a JavaScript sandbox that runs on the harness side, letting agents orchestrate tool calls with concurrency and deferred loading. This shift also improves support for Jev and other model features.
MCP in Pi now treats tools like OpenAPI endpoints — returning structured data and discoverable via descriptions — rather than dumping everything into context. The team admits many MCP servers still optimize for token efficiency over composability, but believes embedding MCP with Codemode lets them push the protocol toward better patterns for small harnesses.
The practical payoff: users can ask Pi to combine MCP tools with models like Jev in a single script. The article shows a real example where Pi pulled 250 Linear issues, classified commenter frustration in batches, and returned ranked results — all without burning context on raw tool output. Codemode state lives in the session transcript, not the filesystem, and it auto-loads when MCP is configured.
Source: news.ycombinator.com, earendil.com
A new open-source benchmark called livenerf is running daily tests on Claude Opus 5.5 to detect post-launch performance degradation, with six days of baseline data collected so far from hackernews.
Livenerf launched September 24, about 2.5 days after Opus 5.5's release, and runs 90 samples daily for 30 days against a pre-registered panel of 78 questions selected from GPQA Diamond, MMLU-Pro, competition-math, and AIME 2025–26. The panel was screened from 2,336 questions where Opus 5.5 scores roughly 93% on first try; the 78 chosen ones are those the model only sometimes gets right, with a measured selection bias pushing fresh-sample pass rates from 54.7% to 62.0%.
The setup pins the Claude Code CLI version, disables auto-updaters, and stores raw logs permanently. Validation showed the rig can detect accuracy drops of about 7.5 points per 10-day window, with lower effort showing up far more clearly in output tokens than accuracy: −62% tokens for low effort versus −8.3 accuracy points. A key limitation emerged in validation — swapping in Opus 5 was indistinguishable from Opus 5.5 at 99% confidence, meaning the instrument can't catch a same-family model swap within a validation's worth of samples.
An audit of the 78 questions found 8 answer keys that look wrong and 30 ambiguous questions, but nothing was dropped; a pre-registered sensitivity analysis will rerun results without them. The safety classifier sometimes answers with Opus 5 or refuses certain biology and math questions — those samples get rejected and counted. As of September 29, six of 30 days are collected with none missed, all on the same harness hash and pinned CLI, with one budget guard override logged.
Source: news.ycombinator.com, github.com
A Factorio player created a matrix-based calculator for planning quality-module upcycling loops in the game's Space Age expansion, from hackernews.
The article walks through how Factorio's Space Age adds five quality tiers (normal to legendary) and quality modules that give each craft a probability of jumping tiers, with each extra tier costing a further 10% chance. The author argues this randomness makes existing factory planning tools inadequate for modeling "upcycling" loops, where items are repeatedly crafted and recycled with quality modules until they hit a target tier.
To solve this, the author formalizes quality transitions as a 5x5 matrix, where each row represents input quality and each column output quality, with coefficients derived from the game's jump table. Multi-step chains (e.g., mining → smelting → crafting) are then computed by multiplying the initial ingredient vector by each step's transition matrix in sequence.
The piece includes a concrete worked example — iron ore (7.5% quality chance) → iron plates (5%) → pipes (10%) — showing how the final probability distribution across all five tiers emerges from successive matrix multiplications. It concludes by pointing readers to an online calculator tool that automates this math for arbitrary chains.
Source: news.ycombinator.com, exyr.org
A Polish observer details how the Hormuz closure and war in Ukraine are driving fuel, food, and heating crises across Europe and beyond, from hackernews.
Iran has kept the Strait of Hormuz closed since March, cutting tanker traffic by over 90% and pushing Brent crude to $108 by late September. Washington rejected Tehran’s ceasefire roadmap, and strikes may resume after midterms. The disruption has hit fertiliser shipments (30% of global trade), spiked tanker freight rates — the Breakwave ETF up 2,300% this year — and squeezed LNG supplies after Qatar declared force majeure.
Ukrainian drone strikes have halved output at Russia’s six largest diesel plants, triggering US diesel above $6 a gallon and a White House request to Kyiv to stop targeting refineries. In France, 15% of stations ran dry after a price cap at TotalEnergies created panic-buying; the official figure undercounts because stations only register when fully out of one fuel grade.
Europe’s potato harvest is set to fall 25% after growers cut planting following last year’s glut, then hit drought and heatwaves. Belgian processing prices jumped from €10 to €150 a tonne. Polish wheat is up 28% on the Paris exchange, and maize 36% — despite Poland being the EU’s most food self-sufficient nation. The country can feed itself but cannot set its own prices, as buyers follow global markets and exports leave when world prices rise.
The FAO warns of delayed yield losses through 2027, building on a grim baseline: 2025 saw two confirmed famines, and the World Food Programme estimates high oil prices could push 45 million more people into acute food insecurity. Poland’s winter heating remains vulnerable, as LNG from Qatar — hit by drone damage at Ras Laffan — replaced Russian pipeline gas but still transits the closed strait.
Source: news.ycombinator.com, tomwojcik.com
OpenAI launched Dots, always-on AI agents powered by GPT-6 Astra, rolling out today to Pro and Business Premium users in eligible markets, from hackernews.
OpenAI introduced Dots, a new class of autonomous agents that run on their own cloud computers, have their own browsers, and can connect to over 4,000 apps via plugins. They're designed to work continuously—investigating bugs in Slack, turning designs into working apps, or preparing invoices without being asked. Users interact with dots through ChatGPT, Slack, Teams, or voice calls, and dots learn preferences over time. The first dot is included free with Pro or Business Premium plans, with extended usage limits for the first month.
The rollout targets Pro and Business Premium users today, with Enterprise (including Edu and Healthcare) beta access once admins enable it. OpenAI also previewed "specialist dots"—organization-specific agents with their own identities, credentials, and access to internal systems—currently in focused enterprise pilots across procurement, invoice processing, and customer support. These are being integrated with Microsoft's Agent 365 for enterprise governance and security controls.
Safety controls include read-only background research, custom rules for approvals, auto-review of consequential actions, and built-in safeguards that can pause work on detected threats. Dots can use saved passwords without exposing them to the model, and sensitive tasks like password changes always require user action. Conversations with dots don't count against ChatGPT usage limits, though tasks delegated to Codex or ChatGPT Work do. OpenAI acknowledged dots can still make mistakes, advising users to review consequential work.
Source: news.ycombinator.com, openai.com
Vermont's largest utility, Green Mountain Power, is replacing fossil-fuel peaker plants with a network of thousands of home batteries that now serves as the state's biggest power source, from hackernews.
Green Mountain Power's virtual power plant — a network of over 5,500 leased home battery systems, often paired with rooftop solar — has quietly become Vermont's largest power source. The utility can draw on these batteries during peak demand or emergencies, while households keep reserve power for outages. Participants pay $55/month over a 10-year lease, which many compare favorably to the $12,000 cost of a gas generator. The program began as a 2017 pilot and surged after state regulators lifted a cap in 2023, adding 2,500 customers since.
Customers like Pollyanna Bladyka in Springfield and Megan Browning in northern Vermont report no outages since installation, even during severe windstorms that knocked out nearby homes. GMP says the virtual plant saved customers $11 million last year and has let it decommission most of its oil, diesel, and natural gas peaker plants — four still run occasionally if the regional grid operator calls, though GMP hopes to close them soon. The utility's goal is eliminating all outages by 2030, with battery systems and undergrounding lines as key tools.
Nationally, virtual power plants face a structural hurdle: many utilities earn more from building infrastructure than from managing demand, a "backwards financial incentive" flagged by former Department of Energy official Jen Downing. Still, states like Massachusetts, Illinois, and New Jersey are pushing utilities to adopt the model, and the DOE estimates such systems could save the US over $10 billion annually by 2030. GMP's CEO Mari McClure frames the shift as moving to a "two-way grid" where customers become both consumers and producers.
Source: news.ycombinator.com, bbc.com
U.S. postal inspectors seized LabelsBank.com and charged its Pakistani operator with selling over 5 million counterfeit USPS labels, causing $126 million in losses, from hackernews.
Federal authorities in Miami announced the takedown of LabelsBank.com, a site run by Faheem Akram, 33, of Khanewal, Pakistan. The operation allegedly sold counterfeit USPS postage at a flat $2 per label, regardless of weight, size, or destination — a steep discount that let more than 5,000 customers ship packages for far less than legitimate rates. Akram faces one count of conspiracy to defraud the United States, five counts of making and selling counterfeit postage, and four counts of wire fraud; he is presumed innocent until proven otherwise.
The indictment alleges the scheme moved more than 5.1 million counterfeit labels, costing USPS over $126 million in unpaid shipping fees. Postal inspectors from the Miami Division traced the sales and obtained a court order to seize the domain, shutting the site down. Officials framed the case as an international enforcement win: U.S. Attorney Jason A. Reding Quiñones called the scheme "simple but massive," while Inspector in Charge Bladismir Rojo said the service will pursue fraudsters targeting U.S. consumers regardless of where they operate.
Source: news.ycombinator.com, postalemployeenetwork.com
America.gov's content is referenced in a Hacker News post, though no article body was provided from hackernews.
The submission points to America.gov, the U.S. State Department's public diplomacy platform, but the linked article contains no extractable text. The post itself offers only the title and source tag, leaving the intended story unstated.
Without body copy, the digest can only note the URL's existence and its placement on Hacker News. No events, statements, or data can be verified from the submission alone.
The practical takeaway: the item is a placeholder. Readers seeking substance will need to click through to the actual America.gov page — assuming the link resolves to more than a landing page.
Source: news.ycombinator.com, america.gov
openai/mcp-extensions is trending in typescript (344 stars) from GitHub
Build plugins that feel like native, first-class features of ChatGPT.
Source: github.com
fsiaonma/elpis is trending in typescript (311 stars) from GitHub
elpis
Source: github.com
omdesale777/avcoe-code-roaster is trending in typescript (28 stars) from GitHub
(no description provided)
Source: github.com
ZacharyZhang-NY/Ely-GPUI-Components is trending in rust (136 stars) from GitHub
A component library for GPUI, in light and dark. Every component runs live in the browser.
Source: github.com
mlc-ai/TIRx-harness is trending in rust (59 stars) from GitHub
An Open Compiler Harness for Agentic GPU Programming
Source: github.com
feder-cr/dots is trending in python (886 stars) from GitHub
Open-source dots for the web: an AI agent with its own browser, one that does not get blocked.
Source: github.com
OpSafari/hypoarena is trending in python (545 stars) from GitHub
Scientific hypothesis-discovery workbench: grounded hypothesis-evidence graphs, synthetic literature with planted causal chains, generate-debate-evolve loops over pluggable offline adapters, Elo tournaments recovering planted skill order, paraphrase dedup, Bayesian evidence accumulation, reproducible reports. NumPy core, CPU-only torch extra.
Source: github.com
PostHog/jeeves is trending in python (318 stars) from GitHub
Jeeves – Reasoning improves Jev-like decision models
Source: github.com
AI News
OpenAI announced GPT-6.1 Sol, a cheaper, near-Astra-tier model for coding, computer use, and professional tasks at one-fifth the standard API price, from OpenAI Blog.
OpenAI released GPT-6.1 Sol as a new API offering, positioned just below its flagship Astra model in capability. The headline change is price: Sol’s input and output token costs are set at 20% of Astra’s standard rates, making high-end model access notably cheaper for bulk or frequent workloads.
The model targets three use cases explicitly: coding, computer use (likely agentic or tool-driven actions), and general professional work. That suggests OpenAI is aiming Sol at production environments where cost per token matters more than peak intelligence — a deliberate trade-off between performance and spend.
No benchmark numbers, context-window specs, or availability dates appear in the short announcement. The post reads as a launch note rather than a technical deep-dive, so developers will likely need to check the API docs or pricing page for concrete limits and rollout details.
The pricing move follows a broader industry trend of tiered model lines, but the 5x discount on a near-top-tier model is aggressive. If Sol holds up on real-world tasks, it could shift how teams budget for AI-heavy pipelines — less need to ration calls or downgrade to weaker models for cost reasons.
Source: openai.com
OpenAI DevDay 2026 brought over 20 product and tool announcements, headlined by GPT-6 Astra and updates across ChatGPT, Codex, APIs, and security, from OpenAI Blog.
The flagship reveal was GPT-6 Astra, a new model generation positioned as a step-change in capability. Alongside it, OpenAI detailed a refreshed ChatGPT experience and expanded Codex tooling for developers, though the recap does not specify exact features or benchmarks. The API suite also saw updates, with new endpoints and builder-focused tools aimed at production workloads.
Security was a named pillar of the event, with announcements covering model safety and platform hardening. The bulk of the announcements target the developer ecosystem: new APIs, tooling, and workflows designed to move from prototype to deployment faster. No pricing, availability dates, or performance numbers were included in the recap itself.
Source: openai.com
OpenAI introduced dots, proactive AI assistants for managing complex projects and daily tasks, from OpenAI Blog.
The announcement is sparse: dots are positioned as proactive assistants, meaning they take initiative rather than waiting for prompts. The focus is on continuous work across both involved projects and routine errands, with an explicit emphasis on user control.
No technical details, product availability dates, or feature lists were disclosed. The framing suggests a shift toward agents that operate autonomously over time, though the blog post offers no concrete examples of how this autonomy is bounded or revoked.
The wording implies a step beyond reactive chatbots, but the absence of specifics leaves open questions about integration, safety mechanisms, and which platforms will support these dots. The post reads as a positioning statement more than a product launch.
Source: openai.com
OpenAI published early guidelines for safety cases in frontier AI training, covering technical safeguards, operational practices, and misalignment investigation, from OpenAI Blog.
OpenAI released draft guidance on how it plans to evaluate safety before training frontier models. The document outlines three core areas: technical safeguards, operational practices, and a process for investigating misalignment incidents. It’s positioned as an early framework, not a final policy, and invites external feedback.
The technical safeguards section focuses on monitoring and controlling model behavior during training runs. Operational practices cover how OpenAI staff should handle risks, including escalation paths and documentation. The misalignment investigation portion describes steps for reviewing incidents where a model behaves contrary to its intended goals.
This release comes as regulators and researchers push for more concrete safety commitments from frontier labs. OpenAI frames the guidelines as a starting point for building "safety cases"—structured arguments that a system is safe enough to deploy. The paper stops short of specifying pass/fail thresholds or enforcement mechanisms, leaving those for later iterations.
Source: openai.com
NVIDIA's Kumo Tabular model achieves a new balance of accuracy and efficiency for tabular prediction tasks, from Hugging Face Blog.
NVIDIA introduced Kumo Tabular, a model designed for tabular data—the most common format in enterprise machine learning. The announcement claims it sets a new frontier by improving both predictive accuracy and computational efficiency compared to existing methods.
The model targets a known pain point: tabular models often sacrifice speed for accuracy or vice versa. Kumo Tabular reportedly handles structured data with fewer resources while maintaining or exceeding benchmark performance, which could lower deployment costs for teams relying on tables in finance, healthcare, or retail.
No technical paper or benchmark details are included in the available text, so the specific architecture, training data, or comparative results remain unspecified. The claim stands on NVIDIA's reputation in hardware-software co-design rather than published evidence here.
Readers should treat the headline as a product announcement, not a peer-reviewed result. The practical takeaway: if the claims hold, tabular workloads could get faster and cheaper without losing predictive quality.
Source: huggingface.co
Hugging Face introduces source-aware verification for MCP agents, moving beyond simple factual checks to validate the origin of information. from Hugging Face Blog
The piece outlines a new framework for auditing MCP-based agents, where outputs are verified not just for correctness but for whether the evidence traces to a legitimate source. This targets a known weakness in retrieval-augmented pipelines—where an agent can state a true fact but cite it to a fabricated or manipulated document.
The technical approach involves scoring source provenance at the agent’s tool-call level, cross-referencing each claim against the actual documents or APIs it claims to have used. Early tests show this catches hallucinated citations and stale cache references that standard fact-checking misses, particularly when agents blend multiple sources into a single response.
Implementation details remain sparse, but the blog signals this as a research direction rather than a shipped library. The practical implication: enterprises running MCP agents for compliance-heavy tasks can start auditing the *path* of an answer, not just the answer itself, which changes how they’ll evaluate agent reliability in production.
Source: huggingface.co
Hugging Face announced Holo4, a framework for building generalist computer-use agents, from Hugging Face Blog.
The release centers on Holo4's claim to generalize across digital environments — not just browsers or specific apps, but the full desktop stack. The post positions it as a departure from narrow, task-tuned assistants, aiming instead for an agent that can perceive, plan, and act on whatever interface it encounters.
Technical details are sparse in the title-only excerpt, but the framing suggests a modular architecture: perception modules for screen understanding, reasoning loops for task decomposition, and action layers for input simulation. No benchmarks or code links are given, which is typical for an announcement post rather than a deep-dive.
The practical angle is that Holo4 targets automation of routine computer work — file management, form filling, cross-app data transfer — where existing agents often fail when the UI shifts even slightly. Whether it delivers on "generalist" claims will depend on real-world testing, but the blog's tone implies a working system rather than a proposal.
For readers tracking agentic AI, this is a signal that the field is moving past chat-bounded tools toward full OS-level control. The lack of body text means specifics like model size, training data, or evaluation suite remain unknown — likely to be filled by follow-up posts or the project’s repository.
Source: huggingface.co
Hugging Face Blog announced LFM2.5-VL-DSpark, an acceleration method for vision-language models.
The post introduces LFM2.5-VL-DSpark, a framework designed to speed up inference in vision-language models. The title indicates a focus on performance optimization rather than new model capabilities, targeting latency reduction in multimodal pipelines.
DETAIL: Details are sparse since the article body is unavailable, but the naming suggests a versioned release (2.5) tied to a "DSpark" optimization engine. The approach likely leverages sparse attention or token pruning, common techniques to cut compute in VLMs without major accuracy loss.
DETAIL: For daily digest purposes, this is a technical release note — practitioners should watch for benchmarks or code drops if they care about deployment efficiency. No user-facing features or model weights are implied from the title alone.
Source: huggingface.co
America.gov's response to a Minecraft query appears strange but is functioning as intended, not a hallucination, from TechCrunch AI.
The federal site, when asked about the video game Minecraft, returned a response that reads as oddly formal or off-topic. This is not a glitch or an AI malfunction—the system is deliberately designed to produce safe, generic answers on topics outside its core mission. The odd phrasing is a feature of its content guardrails, not a breakdown.
The article notes that this behavior is actually reassuring from a national security standpoint. If America.gov were generating creative, unconstrained prose—like poetry—that would signal a serious problem with its output controls. Instead, the weirdness is a sign that the system is staying within its intended, limited parameters.
What makes this noteworthy is the contrast between user expectation and system design. People assume a government AI should answer naturally, but here, the stilted response is the correct outcome. The system is prioritizing stability and safety over conversational fluency, and that trade-off is working as intended.
Source: techcrunch.com
Elon Musk's xAI redirected the domain "dot.com" to its Grok chatbot page shortly before OpenAI's Tuesday launch of its new AI agent, Dots, from TechCrunch AI.
The domain acquisition and redirect were live before OpenAI announced Dots, positioning Grok's download page as the first thing anyone typing "dot.com" sees. Whether the move was a deliberate jab at OpenAI's naming or a preemptive traffic grab, it landed with timing that made the troll apparent to many online observers.
The redirect itself is minimal — a direct hop to Grok's app store listing, no interstitial page or explanation. OpenAI did not publicly respond to the domain move before or after the launch. The episode adds another flashpoint to the long-running public friction between Musk and OpenAI, though xAI has not commented on its intent.
The practical effect is modest: "dot.com" remains a real, if unconventional, URL, and its new destination now serves as a constant reminder of the rivalry whenever someone types it. Whether this counts as meaningful competition or just a stunt is still being debated, but the timing and choice of target make it hard to read as an accident.
Source: techcrunch.com
OpenAI is positioning ChatGPT as a discovery and use layer for software and AI agents, challenging the traditional app store model, from TechCrunch AI.
OpenAI has introduced features that let ChatGPT surface and run third-party tools and services directly inside the chat interface. The company is framing this as a shift away from downloading and managing separate apps, instead letting users and AI agents find and invoke capabilities on demand. This includes support for agentic workflows, where ChatGPT can call external software to complete tasks without the user hopping between apps.
The move directly targets the app store paradigm that Apple and Google have dominated for over a decade. OpenAI is betting that conversational discovery — asking ChatGPT for something and having it execute via a connected service — will feel more natural than browsing a storefront. Developers who build for ChatGPT get access to its massive user base without needing to ship a standalone app or pass through a platform’s approval process.
No new revenue split or developer-facing storefront has been announced yet, but the infrastructure suggests OpenAI is building the rails for one. The company has already signed partnerships with major consumer and enterprise services, and the feature set appears designed to let those partners plug in once and reach both human users and autonomous agents. If this gains traction, the question shifts from whether OpenAI can beat the app stores to whether users will abandon the habit of installing software at all.
Source: techcrunch.com
OpenAI is in talks to raise a $30 billion round at a $1.4 trillion valuation, expected to be its last before a delayed 2027 IPO, from TechCrunch AI.
The fundraising would nearly triple OpenAI’s previous valuation of around $500 billion from its last round. If completed, it would give the company a war chest to fund compute expansion and operating costs ahead of its public listing, which has been pushed back from an earlier target.
The delay to 2027 suggests OpenAI is prioritizing scale and infrastructure over near-term market pressure. Investors are reportedly weighing the round’s size and terms, though no final agreement has been confirmed.
A $1.4 trillion valuation would place OpenAI among the most valuable private companies globally, rivaling major public tech firms. The round’s completion is seen as a key step before any IPO filing.
Source: techcrunch.com
Sam Altman says OpenAI will only go public once it can make confident safety claims about its models, with no timeline set, from The Verge AI.
At a Q&A following his DevDay keynote, Altman tied any future IPO to demonstrable model safety, not a calendar date. He acknowledged that as capabilities surge, the company must be able to back its safety promises with evidence. No specific milestones or regulatory benchmarks were offered.
Altman also pushed back against the idea of waiting indefinitely, calling an overly delayed IPO "bad for the world." That tension — between needing safety proof and wanting public market access — frames the company’s current posture. He offered no hint of when either condition might be met.
The remarks land amid ongoing scrutiny of OpenAI’s safety practices and governance, including recent leadership shifts and external criticism. The company has repeatedly said it prioritizes safety, but Altman’s comments are among the clearest public statements linking that priority directly to a public offering.
Source: theverge.com
President Trump signed an executive order requiring US government materials to replace the term "artificial intelligence" with "Super Intelligence," per the announcement at the America.gov launch event from The Verge AI.
The order applies to official policy websites, documents, and press releases across the executive branch. Trump framed the change as a matter of accuracy, claiming AI is "not artificial" but "very powerful" and "very brilliant," and said the word "super" is "the best word of all."
Trump also said Chinese President Xi Jinping, who visited the White House last week, "loves" the new terminology. The announcement came as part of the launch of America.gov, though the order’s immediate practical effects on existing AI-related federal programs and documents remain unclear.
The shift in naming does not alter any technical definitions or regulatory scope, but it does change the official vocabulary used in federal communications. Whether agencies will adopt the term in internal rules or only in public-facing materials has not been specified.
Source: theverge.com
Grokipedia, Elon Musk's AI-powered encyclopedia, has resumed article updates after a months-long pause, per a new report from The Verge AI.
The platform's live updates site now shows recent activity, including fact-check notes on pages like Barack Obama's and Elon Musk's own entry, which was verified during the writing of the article. Many updates currently amount to placeholder notes instructing "recheck all references and sources," suggesting the system is revalidating existing content rather than adding new material.
DETAIL: This follows an August report from Lawfare that Grokipedia had not processed edit reviews since April, leaving the site in a state of suspended maintenance. The new activity marks the first visible sign of life in months, though the scope of the changes remains unclear — the majority of recent edits are metadata flags, not substantive revisions.
DETAIL: The timing of the reactivation is notable, as it aligns with broader moves from SpaceXAI to keep its AI products in public view. Whether this signals a full operational return or a limited technical refresh isn't confirmed, and The Verge's own checks indicate the process was still ongoing as they observed it.
Source: theverge.com
AI researchers, including current and former OpenAI, Google DeepMind, and Anthropic staff, released interviews warning that superintelligent AI poses a significant extinction risk, from The Verge AI.
Palisade Research, a nonprofit studying AI capabilities and motivations, launched the interviews on frominside.ai. Geoffrey Irving, formerly of OpenAI and Google DeepMind, put the chance of human extinction at “about a coin flip.” Google DeepMind’s Neel Nanda said there’s “at least a 10 percent chance” AI causes human extinction, echoing several other researchers who spoke on the record.
The interviewees were current and former employees across the three leading labs. Their warnings center on the difficulty of controlling systems that surpass human intelligence, not on immediate model failures. The project appears designed to push the conversation beyond technical benchmarks into existential risk, giving insiders a platform to state views they say are common but rarely aired so bluntly.
Palisade Research describes itself as a nonprofit focused on understanding AI’s motivations, and the interview collection is its latest public push. The Verge’s coverage highlights that these aren’t fringe voices — they’re people who built the systems they now caution against. The site’s full story includes additional quotes and context beyond what’s excerpted here.
Source: theverge.com
Politics
Tour Company Sues to Undo Trump’s Nonresident Fees at National Parks from NYT
An Arizona guide service challenged rules that force foreigners to pay a $100 surcharge to visit 11 national parks, saying that only Congress can change pricing.
Source: NYT (lean-left, nytimes.com)
Ken Paxton, G.O.P. Nominee in Texas, Privately Said Trump’s Convention Hurt Him from NYT vs Hispanic Voters on the Texas Border Have Soured on Trump from NYT vs Transcript: Trump Erupts in Panic as GOP Midterm Fears Grow: “Wipeout” from The New Republic
(NYT, lean-left) Ken Paxton, G.O.P. Nominee in Texas, Privately Said Trump’s Convention Hurt Him: Ken Paxton, one of the party’s top Senate candidates and a Trump loyalist, said that “everybody’s numbers dropped” after the event, according to audio obtained by The Times.
(NYT, lean-left) Hispanic Voters on the Texas Border Have Soured on Trump: In 2024, these voters were an essential part of the winning Republican coalition. Now, many feel abandoned and furious, complicating the party’s chances in November.
(The New Republic, unrated) Transcript: Trump Erupts in Panic as GOP Midterm Fears Grow: “Wipeout”: The following is a lightly edited transcript of the September 29 episode of the Daily Blast podcast. Listen to it here.Greg Sargent: This is The Daily Blast from The New Republic, produced and presented by the DSR Network. I’m your host, Greg Sargent.We’re a …
Source: NYT (lean-left, nytimes.com), NYT (lean-left, nytimes.com), The New Republic (unrated, newrepublic.com)
Can President Trump Demolish the Kennedy Center? Experts Say Not Easily. from NYT
The president has entertained the idea of tearing down the performing arts venue. Analysts say any plan would face multiple daunting challenges.
Source: NYT (lean-left, nytimes.com)
As Violence Falls, Crime Isn’t Paying as a Republican Line of Attack from NYT
Across the country, especially in the North Carolina Senate race, Republicans have charged opponents with being soft on crime, even as President Trump brags over plunging crime rates.
Source: NYT (lean-left, nytimes.com)
Sen. Josh Hawley introduces bill aimed at stopping abuse of Flock cameras from CBS News
Sen. Josh Hawley is unveiling legislation to establish guardrails for Flock cameras and other artificial intelligence-powered surveillance technology, CBS News has learned.
Source: CBS News (lean-left, cbsnews.com)
Oregon eyes nation's first universal healthcare system. Will it work? from CBS News
While Democratic congressional candidates are floating the restoration of federal Affordable Care Act subsidies and Medicaid funding, Oregon officials are crafting a more ambitious plan to implement universal coverage in their state.
Source: CBS News (lean-left, cbsnews.com)
Bureau of Prisons ordered to restore its union's contract from CBS News
A federal judge on Tuesday ordered the Bureau of Prisons to reinstate a collective bargaining agreement with the agency's union after the bureau's director abruptly terminated it last year.
Source: CBS News (lean-left, cbsnews.com)
Hegseth directs military academies to stop giving tenure to civilian professors from CBS News
Defense Secretary Pete Hegseth ordered institutions like the Military Academy at West Point and the Naval Academy in Annapolis to stop granting tenure to civilian faculty, in his latest change to military higher education.
Source: CBS News (lean-left, cbsnews.com)
British authorities defend decision to let terror suspects go on bail from CBS News
British authorities defended their decision to let a group of young men go free, even as investigators look into a suspected terror plot near an air base there that was used by the U.S. military in its war against Iran. Ramy Inocencio reports from London.
Source: CBS News (lean-left, cbsnews.com)
From mass raids to quiet crackdown: a timeline of Trump’s strategic ICE pivot from The Guardian
After switching tactics, federal agents are detaining and deporting record numbers of people from the US
When federal agents fatally shot US citizen protesters Alex Pretti and Renee Good in Minneapolis in January during the nation’s largest immigration-enforcement operation ever, it marked a turning point for the second Trump administration’s mass deportation campaign.
Donald Trump in his second term had sparked outrage and huge protests with federal immigration officers descending en masse to one Democratic-run city after another in what some leaders called militarized “invasions”. Los Angeles, Chicago, Charlotte, New Orleans, then Minneapolis saw lines of masked and armored officers swaggering into town to conduct street arrests, ambushes of day workers waiting in Home Depot car parks for hire, and neighborhood sweeps. Adults and kids were whisked off to detention and protests met with aggressive deployment of chemical spray, teargas, fists and batons. The killings in Minneapolis were the culmination of what appeared to many to be a federal shock force increasingly out of control while being defended by federal leaders – and public opinion broadly turned against this violence.
Continue reading...
Source: The Guardian (lean-left, theguardian.com)
A ‘freaked out’ US food industry is hitting back at Maga – with social media influencers from The Guardian
Cross-partisan support in Washington on issues like pesticides and food additives pose a threat to the industry’s bottom line
Some of the biggest food and agriculture corporations in the US are funding a campaign that pays social media influencers to combat elements of Robert F Kennedy Jr’s Make America Healthy Again platform, according to an investigation by the Guardian and Documented, an investigative watchdog and journalism project.
The ascension of RFK Jr to secretary of health and human services has boosted his putative campaign to end chronic illness in the US, which he believes is primarily caused by processed foods and dyes, exposure to certain chemicals, and poor farming practices, along with an overuse of what he falsely claims are dangerous vaccines. The “Maha” movement blends long-running mainstream concerns about food and farm safety with conspiracy theories about everything from medicine to milk. A Maha “summit” in Washington this week will feature JD Vance and several members of Trump’s cabinet.
Continue reading...
Source: The Guardian (lean-left, theguardian.com)
Trump campaign ads paid with public money violate federal law, critics say from The Guardian
White House has labeled ads ‘public service announcements’ and has responded defiantly to criticism
Federal regulators have been urged to stop the broadcast of White House advertisements promoting Donald Trump that critics say amounts to government propaganda and violate laws against using public money to promote partisan causes.
Public Citizen, an advocacy group, called on the Federal Communications Commission (FCC) and the Federal Trade Commission (FTC) to intervene and block further airing of three ads it said were “designed to boost the image and character of Donald Trump” as part of the Republicans’ campaign for November’s midterm elections.
Continue reading...
Source: The Guardian (lean-left, theguardian.com)
As AI images become common, the threat to elections draws alarm: ‘How will voters know what’s true?’ from The Guardian vs Trump says top tech firms have signed accord to 'self-police' AI development from NPR vs Three takeaways from Trump's 'Super Intelligence' summit from BBC News
(The Guardian, lean-left) As AI images become common, the threat to elections draws alarm: ‘How will voters know what’s true?’: Deepfakes and misinformation driven by artificial intelligence have become a point of deep concern for legislators, advocates and the public
An image of College Park, Georgia mayor Bianca Motley Broom kissing a man – who was not her husband – in a bar spread far and wide on Facebook. The fact that it’s fake is spreading faster, now, but only after it had been seen by thousands of people.
“I think this is a sad commentary about where we are in our democracy, and where we are in our political discourse,” Motley Broom said. “It also seeks to reinforce really tired stereotypes about women in leadership.”
Continue reading...
(NPR, center) Trump says top tech firms have signed accord to 'self-police' AI development: President Trump on Tuesday said that he and a large group of leaders of artificial intelligence companies had signed a voluntary accord that will include internal and external reviews.
(BBC News, center) Three takeaways from Trump's 'Super Intelligence' summit: The meeting at the White House came as some tech bosses and experts have called for tighter rules around AI.
Source: The Guardian (lean-left, theguardian.com), NPR (center, npr.org), BBC News (center, bbc.co.uk)
Supreme court allows Trump to temporarily resume deporting people to third countries – as it happened from The Guardian
This live blog is now closed.
US supreme court allows Trump officials to resume deportations to third countries
Sign up to the US Breaking News email
For some context, it’s worth remembering that Smith had a two-volume report of his findings as a result of his criminal investigations – that resulted in two indictments against the president after his first term. The first focused on Donald Trump’s attempts to overturn the 2020 election, while the second on the president’s retention of classified documents.
However, in February, a Trump-appointed federal judge permanently barred the justice department from releasing the second volume of Smith’s report.
Continue reading...
Source: The Guardian (lean-left, theguardian.com)
Former special counsel Jack Smith defends investigation into Trump on Capitol Hill from NPR vs Schmitt faces fallout after debunked Jack Smith narrative from The Hill vs Sen. Katie Britt’s ‘look of utter horror’ as GOP colleague’s perjury accusation unravels during Jack Smith hearing goes viral from New York Post
(NPR, center) Former special counsel Jack Smith defends investigation into Trump on Capitol Hill: Former Justice Department special counsel Jack Smith defended two criminal indictments against President Trump on Capitol Hill.
(The Hill, center) Schmitt faces fallout after debunked Jack Smith narrative: Sen. Eric Schmitt (R-Mo.) is facing accusations he tried to launch a conspiracy tying former special counsel Jack Smith to others who prosecuted President Trump — a line of questioning that unwound when the senator apparently confused two basketball teams. Schmitt seemed poised for a gotcha moment with Smith when he asked him if he...
(New York Post, lean-right) Sen. Katie Britt’s ‘look of utter horror’ as GOP colleague’s perjury accusation unravels during Jack Smith hearing goes viral: “The moment Katie Britt's soul left her body," one user wrote of the senator's stunned expression.
Blindspot: No lean-left outlet in this digest's sources covered this story.
Source: NPR (center, npr.org), The Hill (center, thehill.com), New York Post (lean-right, nypost.com)
What's behind HHS claims that gender-affirming care 'may' radicalize from NPR
The report is the agency's latest effort to chip away at trust in established health care models for transgender people. Critics say it frames transgender identity as illegitimate or even dangerous.
Source: NPR (center, npr.org)
Outside groups are spending millions on races that could determine Senate control from NPR
New Hampshire's Senate race is one of a handful that could determine control of Congress next year. That's why outside money groups are plowing big money into it — some $50 million to date.
Source: NPR (center, npr.org)
N.Y. Sen. Gillibrand says new allegations in Cornell case are 'serious and horrific' from NPR vs What we know about Cornell University frat house rape investigation from BBC News
(NPR, center) N.Y. Sen. Gillibrand says new allegations in Cornell case are 'serious and horrific': NPR's Juana Summers talks with New York Sen. Kirsten Gillibrand about the reopening of the rape case at Cornell University, and her history of pushing legislation to curb campus sexual assault.
(BBC News, center) What we know about Cornell University frat house rape investigation : The woman alleges she was drugged and raped at a fraternity house at the university two years ago.
Source: NPR (center, npr.org), BBC News (center, bbc.co.uk)
US Supreme Court denies Christa Pike's last-ditch bid to halt execution from BBC News
The Tennessee inmate has been awaiting execution since she was sentenced to death in 1996 for the murder of a teenage classmate.
Source: BBC News (center, bbc.co.uk)
Six Flags shuts down X2 rollercoaster after more than 100 allege brain injuries from BBC News
Hundreds of riders were allegedly injured on X2, which has seats that spin 360 degrees.
Source: BBC News (center, bbc.co.uk)
Lindsay Clancy appears in court as her lawyer pushes for murder case dismissal from BBC News
Clancy's defence lawyer wants her to be cleared of murdering her three children in 2023 but prosecutors say the state proved guilt.
Source: BBC News (center, bbc.co.uk)
Nikole Hannah-Jones just ran head-first into the great dumbing-down from The Hill
Education bureaucrats across the nation are burying evidence of a widespread and unpleasant truth: Public high schools are churning out barely literate, almost innumerate graduates.
Source: The Hill (center, thehill.com)
Pritzker's weight loss prompts discussion on changing attitudes over presidential appearance from The Hill
JB Pritzker is looking different these days. Illinois's Democratic governor, who is considering a 2028 run for president, admitted as much when he acknowledged he lost 80 pounds after using a GLP-1 in the last two years to help treat his diabetes. Pritzker’s weight loss journey has been so apparent in recent months that the governor felt the...
Source: The Hill (center, thehill.com)
New Supreme Court term kicks off Monday: Here's what to expect from The Hill
The Supreme Court’s new term is about to kick off next week. Join Legal Affairs Editor Zach Schonfeld and Legal Affairs Reporter Sophie Brams as they preview the big cases that they have their eyes on and the disputes bubbling up from the lower courts that the justices still may take up. It’s all on "The Gavel,"...
Source: The Hill (center, thehill.com)
GOP breaks with Trump on multiple issues from The Hill
Republicans in both chambers are increasingly breaking ranks with President Trump across a host of sensitive issues, hoping some separation from the GOP’s unpopular standard bearer will help avert a blue wave in the midterm elections. The divisions are highly unusual for a party that’s walked virtually in lock-step with Trump for more than a...
Source: The Hill (center, thehill.com)
Inside the neighborhood living on the edge of America’s tech frontier from Fox News
Loudoun County homeowners on Statesboro Place describe constant noise from Amazon Web Services and Microsoft data centers just 150 feet from backyards.
Source: Fox News (lean-right, foxnews.com)
Liberal stronghold set to adopt eye-popping minimum wage amid business exodus from Fox News vs Susan Collins’ time is up as Maine’s Houdini from Salon
(Fox News, lean-right) Liberal stronghold set to adopt eye-popping minimum wage amid business exodus: Seattle's minimum wage is set to hit $22.14 in 2027 as 450 restaurants closed in early 2025 and job postings dropped 35% since February 2020.
(Salon, unrated) Susan Collins’ time is up as Maine’s Houdini: The senator cannot make her role as Donald Trump's enabler disappear
Source: Fox News (lean-right, foxnews.com), Salon (unrated, salon.com)
US faces another major Middle East troop withdrawal, but threat remains from Fox News
About 2,500 U.S. troops had remained in Iraq as part of Operation Inherent Resolve, but Baghdad's effort to disarm Iran-backed militias remains unfinished.
Source: Fox News (lean-right, foxnews.com)
Schumer preaches affordability in Dem midterm pitch while sabotaging GOP fixes for soaring costs from Fox News
Senate Democrats blocked GOP affordability and permitting reform bills, raising questions about their campaign messaging on cost of living issues.
Source: Fox News (lean-right, foxnews.com)
Maine Dem Sen candidate's 'working class' cred shredded by both parties when internet sees latest 'cringey' ad from Fox News
Troy Jackson's boba tea ad mocking Susan Collins and Elon Musk has been viewed over 2 million times, drawing sharp bipartisan ridicule on X.
Source: Fox News (lean-right, foxnews.com)
Maine university system strips ex-Sen. George Mitchell’s name from building, scholarship over Jeffrey Epstein ties from New York Post
UMaine trustees voted 9-5 to strip former Sen. George Mitchell's name from a campus building and scholarship due to his ties to Jeffrey Epstein.
Source: New York Post (lean-right, nypost.com)
A Clinton ‘scandal’ led to Trump-Russia probe, investigative journalist John Solomon tells to ‘Pod Force One’ from New York Post
Solomon, the founder of Just the News and the head of a Trump administration task force on government transparency, argued that polling data presented to the former secretary of state amid her White House run, and her response to it, marked the start of 10-year-long, Democrat-led, “wash, rinse, repeat cycle” of federal investigations against Trump.
Source: New York Post (lean-right, nypost.com)
Reparations activists call out Democrat states for using slavery history to fund broader agenda from New York Post
"I also think it’s not morally correct, you know? You don't give out remedies to people who haven't experienced the harm."
Source: New York Post (lean-right, nypost.com)
Nithya Raman’s ‘genocide’ remarks rattle Jewish crowd at LA mayoral forum from New York Post
Raman said she used the term after seeing reports and images she found deeply disturbing.
Source: New York Post (lean-right, nypost.com)
Kristi Noem files for divorce from husband Bryon after ‘irreconcilable differences’ from Washington Examiner
Former Secretary of Homeland Security Kristi Noem filed for divorce from her husband Bryon Noem after 34 years of marriage. The divorce was filed due to “irreconcilable differences,” according to documents obtained by Dakota News Now. Kristi Noem said she was “devastated” after allegations surfaced that her husband was a secret crossdresser and chatted online […]
Source: Washington Examiner (lean-right, washingtonexaminer.com)
A captivating new biography of the flawed fixer Roy Cohn from Washington Examiner
When Roy Cohn died of AIDS in 1986, Donald Trump was among the few to attend his burial at Union Field Cemetery in Queens. He also attended Cohn’s public memorial later that year in Manhattan, along with Rupert Murdoch, George Steinbrenner, Barbara Walters, and William F. Buckley. Apart from his legions of well-connected friends and […]
Source: Washington Examiner (lean-right, washingtonexaminer.com)
And just like that, Germany’s establishment went kaput from Washington Examiner vs Call them ‘Nazis’ all you want: The new German Right is winning the argument from Washington Examiner
The British diplomat and novelist, Harold Nicholson, tossed off an essay in 1929 containing a short riff on the three great capitals of Europe. “London is an old lady in black lace and diamonds who guards her secrets with dignity,” while “Paris is the woman in the prime of life to whom one would only […]
Source: Washington Examiner (lean-right, washingtonexaminer.com), Washington Examiner (lean-right, washingtonexaminer.com)
Have Americans forgotten how to argue? from Washington Examiner
Americans disagree about almost everything. We argue about very little. The distinction matters. Disagreement is easy. I state my position. You state yours. We discover that they are incompatible. If we are sufficiently invested in the dispute, we repeat ourselves, usually with greater confidence. Eventually, we may stop talking about the issue and start talking […]
Source: Washington Examiner (lean-right, washingtonexaminer.com)
Kansas Hasn’t Elected a Democratic Senator Since 1932. Is This the Year? from Freerepublic.com
on a recent Friday afternoon, the Democratic candidate for the United States Senate from Kansas took time out from the campaign trail to perform a wedding. Two days later, he baptized a baby on the vast campus of Resurrection, a United Methodist Church he fou…
Source: Freerepublic.com (unrated, freerepublic.com)
The Art of Dealing With Trump from The Atlantic
As Canada’s international trade minister and foreign minister, I learned that the only thing that works with this president is strength.
Source: The Atlantic (unrated, theatlantic.com)
Coming soon: Our 2026 list of Climate Tech Companies to Watch from MIT Technology Review
Earlier this month, the UN announced the planet will tip past 1.5 ˚C of warming, “likely within the next few years,” squashing any lingering hope that nations would cut emissions fast enough to achieve the loftiest goal of the Paris climate agreement. In the …
Source: MIT Technology Review (unrated, technologyreview.com)
Erika Kirk Doesn't Really Want to Talk Politics. So How's She Going to Lead Turning Point USA? from Vanity Fair
A year after Charlie Kirk’s murder, the MAGA movement is fracturing and the future of his
influential political organization now rests in the hands of his widow. But how can someone so reluctant to discuss her political views lead Turning Point USA? With unpr…
Source: Vanity Fair (unrated, vanityfair.com)
Taylor Swift’s Audience Is Growing Up. Her Music Is Devolving. from The Atlantic
The singer’s music once put a clever twist on ideas about aging and adulthood. Now she just seems to be going backwards.
Source: The Atlantic (unrated, theatlantic.com)
Iranian spokesman urges Americans to rally against Trump's policies in address from Yahoo Entertainment
Iran's Revolutionary Guard urged Americans in a message ahead of U.S. midterm elections to take to the streets and rally against Donald Trump, claiming the U.S. president's war in the Middle East has failed.
Meanwhile, in the occupied West Bank, settlers ri…
Source: Yahoo Entertainment (unrated, yahoo.com)
Middle East Institute: Ankara’s Nuclear Ambitions in the Post-Iran War Environment from Blogspot.com
Gönül Tol*, September 28, 2026A longer version of this article was originally published by MEI on September 10, 2026.In February 2026, Turkish Foreign Minister Hakan Fidan suggested that shifting international dynamics could push Turkey to acquire nuclear wea…
Source: Blogspot.com (unrated, i-epikaira.blogspot.com)
Cheryl Reeve's WNBA Coach of the Year win shouldn't distract from her dangerous accusations from Fox News
Cheryl Reeve earned a historic fifth WNBA Coach of the Year honor while her comments on "transgender people" in sports fueled growing controversy.
Source: Fox News (lean-right, foxnews.com)
Science News
Fruit fly larvae fed less protein grow into adults that live longer, likely because their bodies can't produce as much protein later on, from Ars Technica Science.
Researchers manipulated the protein intake of fruit fly larvae and tracked how it affected the flies' adult lifespans. The link was direct: larvae on a low-protein diet produced adults that lived significantly longer than those raised on standard or high-protein food. The effect wasn't due to overall calorie restriction — it traced specifically to protein.
The mechanism centers on the flies' ability to synthesize new proteins as adults. Larval protein restriction appears to dampen the adult's capacity for protein production, and that reduced capacity correlates with extended longevity. The study suggests a trade-off: less building material early on means less metabolic wear-and-tear later, extending life at the cost of some physiological efficiency.
This isn't a simple "eat less, live longer" story for humans — fruit flies are short-lived and their larval stage is a unique developmental window. But the work adds a concrete molecular thread to the aging puzzle, showing that dietary signals during early development can set a lasting biological clock for lifespan. The findings are published in the journal *Proceedings of the Royal Society B*.
Source: arstechnica.com
A new battery-based device captures CO2 from gas streams using less energy than current methods, from Ars Technica Science.
Researchers built a prototype that runs a gas mixture across a charging battery, where chemical reactions pull CO2 out of the air and store it in the battery’s electrolyte. The process uses voltage swings to bind and release the gas, avoiding the high-temperature or pressure swings typical of amine-based scrubbers.
The key advance is energy efficiency: the cell operates at near-ambient conditions and needs roughly 0.5–1.0 GJ per tonne of CO2 captured, compared to 3–4 GJ for conventional systems. The team cycled the device over 100 times without significant degradation, though the current lab-scale version handles only small gas flows.
Practical hurdles remain — scaling the electrode surface area, maintaining electrolyte stability over months of operation, and integrating the system with real flue-gas streams that contain moisture and particulates. The authors suggest the technology could pair with renewable electricity for modular, on-site capture units, but they stop short of promising a commercial timeline.
The work is published in a peer-reviewed journal, with the experimental data and full cell geometry available in the supplementary materials. No pilot plant or industrial partnership has been announced yet.
Source: arstechnica.com
Scientists identified a typo that misdated an 1841 train delay, correcting the record on an early geomagnetic storm event from Ars Technica Science.
The mystery centered on a famous 19th-century account in *Nature* describing a train delayed by a geomagnetic storm in 1841. Researchers couldn’t reconcile the reported date with available auroral and magnetic records, making the event seem anomalous or possibly misreported. The new work traced the problem to a transcription error—the year in the original account had been printed incorrectly, likely due to a typo, shifting the storm’s actual occurrence by a year.
Once corrected, the 1841 storm aligns with other documented space weather events of that era, including auroral observations and disturbances in telegraph systems. The fix also resolves inconsistencies in historical geomagnetic indices that had puzzled scientists for decades. The correction doesn’t change the storm’s severity, but it places it properly within the solar cycle, improving long-term models of solar activity.
The findings highlight how small archival errors can ripple through scientific datasets, especially when historical records are used to calibrate modern space weather predictions. The team verified the true date by cross-referencing contemporary newspapers, observatory logs, and railway company records, which all pointed to the corrected timeline. This closes a lingering gap in the reconstruction of 19th-century geomagnetic storms, refining how we assess the frequency and intensity of such events before instrumental monitoring became routine.
Source: arstechnica.com
A new study proposes that freezing and shattering ice droplets from Enceladus naturally concentrates certain chemicals, which could explain Cassini’s puzzling ice-grain data and sharpen future life-hunting missions from Science Daily.
Researchers modeled what happens to salty ocean spray as it escapes Enceladus’s vents. Droplets freeze in stages, and as they crack and lose water vapor, dissolved compounds separate and concentrate unevenly—some grains end up chemically enriched, others depleted. This sorting mimics a laboratory fractionation process, not just simple freezing.
The Cassini spacecraft’s mass spectrometer had found an odd mix of organics and salts in the plume, more varied than expected from a uniform ocean. The new mechanism offers a direct explanation: the ice grains themselves act as mini chemical reactors, preferentially trapping certain molecules while ejecting others. That means the plume’s composition doesn’t directly mirror the ocean’s—it’s an edited version.
For future probes, this is a double-edged gift. A flyby through the plume will encounter grains that are naturally enriched in trace compounds, making detection easier than sampling the ocean directly. But interpretation gets trickier: scientists must now account for how freezing and shattering skew the original chemical ratios. The study, published in *Communications Earth & Environment*, used laboratory experiments with salty solutions to validate the fractionation behavior.
The practical upshot: any mission aiming to sniff for biosignatures in Enceladus’s plume should plan to collect a range of grain sizes and compositions, not just average them. The same process that distorts the signal also amplifies it—provided the instrument knows what to look for.
Source: sciencedaily.com
Scientists may have linked millisecond-scale changes in Earth’s day length to subtle movements of the planet’s inner core, which tug on the mantle from deep underground, from Science Daily.
The article reports a proposed mechanism for why Earth’s rotation speeds up and slows down by a few milliseconds over decades. The likely driver is the inner core, which appears to shift slightly and pull on the mantle through gravitational and electromagnetic coupling. That tug alters the planet’s angular momentum, nudging day length.
The finding comes from researchers analyzing decades of seismic data and rotation records. They matched small wobbles in the inner core’s position with observed changes in day length, ruling out other explanations like atmospheric or oceanic effects as the main cause. The core’s motion, previously thought to be uniform, now looks more variable.
That variability suggests the inner core is not a passive sphere but an active player in Earth’s dynamics. The study adds to evidence that the deep interior is more restless than long assumed, with direct consequences for the surface — even if those consequences are just a few thousandths of a second per day. The work also tightens constraints on the core’s composition and behavior, useful for refining models of Earth’s magnetic field generation.
Source: sciencedaily.com
NASA and SpaceX are targeting an October 1 launch for Crew-13, with the Dragon capsule reaching the ISS in under nine hours, from Science Daily.
The four-person crew will fly to the International Space Station on a Falcon 9 rocket, with liftoff scheduled from Florida. The nine-hour transit is a fast-track trajectory, cutting the typical multi-day trip down to a single shift. The mission is set to dock with the station and integrate into Expedition 75.
The launch date depends on final launch-readiness checks, which teams are completing now. Those reviews cover the rocket, spacecraft, and ground systems before committing to flight. If the schedule holds, Crew-13 will mark another routine rotation flight for the station, with the crew replacing members already onboard.
The short flight time is the headline here, but the operational rhythm matters more: NASA and SpaceX have turned crew rotations into a tight, repeatable schedule. The October 1 window is the current target, though the readiness reviews could shift it if issues surface.
Source: sciencedaily.com
Gaming News
The Witcher 3 Remastered on Switch 2 can be unlocked for free using any physical Switch copy, even a borrowed one, from IGN.
CD Projekt Red confirmed that the Switch 2 version of The Witcher 3 Remastered is free for anyone who owns the original Switch release. The unlock doesn’t require proof of purchase beyond inserting the cartridge — the console detects any physical copy of the game, not just one tied to your account.
That means you can borrow a friend’s Switch cartridge, insert it into your Switch 2, and unlock the remastered version without spending a dime. Once unlocked, the game stays available on your system even after you return the cartridge.
The policy is a deliberate departure from typical upgrade gates, which usually verify ownership via account history or serial keys. CD Projekt Red opted for the simplest possible check, betting that goodwill outweighs the risk of someone exploiting a borrowed cartridge to grab a free upgrade. There’s no word on whether the same logic applies to other platforms, but Switch 2 owners with access to any copy can claim the remaster immediately.
Source: ign.com
Rockstar Games has detailed its first-hand research methods for GTA 6, including visits to strip clubs and night rides with ex-cops in Florida, from IGN.
Rockstar developers embedded themselves in Florida's real-world environment to capture the game's tone. That meant patronizing strip clubs and riding around at night alongside former police officers, per the company's own account of its research process. The goal was direct exposure to the state's particular mix of nightlife, law enforcement culture, and roadside oddities.
The studio hasn't specified how much of that material made it into the game itself. But the admission offers a concrete look at how GTA 6's setting was built, beyond digital recreations of Miami's skyline or highway signage. It's the kind of detail Rockstar rarely shares publicly, and it tracks with the series' long-running habit of grounding its satire in observed behavior.
No release date or platform details were included in this reveal. The information came as part of a broader feature on the game's development, with the strip club and ex-cop specifics standing out as the most concrete new facts.
Source: ign.com
Nicolas Cage called the MCU a "well-funded and glorified WWE" in an interview, per IGN.
Cage, who played Ghost Rider in two Marvel films, made the comment while discussing the current state of superhero cinema. He acknowledged the MCU’s use of top-tier acting talent but dismissed the overall product as spectacle-driven entertainment rather than substantive filmmaking.
The comparison to WWE — professional wrestling’s scripted, theatrical presentation — frames the MCU as more about visual spectacle and brand continuity than narrative depth. Cage’s own Marvel tenure ended in 2011 with *Ghost Rider: Spirit of Vengeance*, a critical and commercial disappointment.
He has since gravitated toward eccentric, low-budget and auteur-driven projects, making his critique consistent with his career trajectory. The remark adds to a growing chorus of actors and directors questioning the MCU’s artistic merit, though Cage stopped short of naming specific films or personnel.
Source: ign.com
The Sekiro anime, titled *Sekiro: No Defeat*, will stream on Crunchyroll after a Japan-only theatrical release, and its TV version will be a re-edit with new scenes, not a movie, from Kotaku.
Crunchyroll has confirmed a streaming date for *Sekiro: No Defeat*, the anime adaptation of FromSoftware’s game. The series first hit Japanese theaters in an early, exclusive run. That theatrical cut is not what subscribers will see; instead, Crunchyroll will carry a TV re-edit that adds new footage.
The distinction matters because the project was initially framed as a film, but the final product is a multi-episode series. The TV version’s new scenes mean the streaming release isn’t just a direct port of the theatrical edit—it’s a separate, extended presentation. No specific episode count or total runtime has been announced.
The streaming date is set, though Kotaku’s report doesn’t list the exact day in the provided text. The key takeaway: the anime is coming to Crunchyroll, and it’s structured for serialized viewing, not a one-shot movie.
Source: kotaku.com
Sony is limiting PS5 Pro purchases in Japan to active PlayStation users via a lottery system to combat shortages ahead of GTA 6, from Kotaku.
Sony has switched PS5 Pro sales in Japan to a lottery system, restricting entry to players with active PlayStation Network accounts and demonstrable playtime. The move targets scalpers who have been snapping up stock, particularly as demand is expected to spike with Grand Theft Auto 6’s release.
The shortage is acute enough that Sony is prioritizing existing users over newcomers or resellers. The lottery applies to the standard PS5 Pro model, and winners are selected at random from eligible applicants. No purchase limit beyond the single-unit-per-winner rule has been announced.
The timing is deliberate. GTA 6 is the console’s biggest upcoming driver, and Sony is trying to avoid a repeat of the PS5’s chaotic 2020 launch, where bots and resellers dominated. Whether the lottery holds up once the game actually ships remains to be seen, but for now, Japanese players with active accounts have a fighting chance.
Source: kotaku.com
Switch 2 owners can get a free digital copy of The Witcher 3 remastered by borrowing any old physical Switch copy of the game, from Kotaku.
Nintendo and CD Projekt Red have confirmed that inserting a physical Switch cartridge of The Witcher 3 into the Switch 2 will unlock a free digital download of the remastered version for that console. The promotion applies to any copy of the original Switch release, including the base game or the complete edition, and does not require the borrower to own the cartridge — just to have access to it for the one-time redemption.
The remastered edition includes improved graphics, faster load times, and updated performance features specific to the Switch 2 hardware. Once redeemed, the digital version stays tied to the account, so the physical cartridge can be returned to its owner without losing access to the upgraded game.
This move sidesteps a common problem with cross-generation upgrades, where players often need to prove ownership through a linked account or a digital purchase. Here, the hardware itself does the verification, making it easier for people who borrowed or rented the original game to jump into the remastered version without an upfront cost. Kotaku notes that this could also pressure retailers or libraries holding old Switch copies, as demand for short-term borrowing might spike.
Source: kotaku.com
