News
The raw AI news stream — everything we ingest across RSS, Hacker News, arXiv, GitHub, Bluesky, YouTube and more, newest first.
| Source | Age | Title | Summary | Status |
|---|---|---|---|---|
rss techmeme | 2mo | OpenAI Foundation board member Paul Christiano says the AI industry is currently not on track to reduce the acute loss-of-control risk to an "acceptable" level (Paul Christiano/@paulfchristiano) | Paul Christiano / @paulfchristiano: OpenAI Foundation board member Paul Christiano says the AI industry is currently not on track to r… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Finally, some good news | Finally, some good news | ↗Not briefed |
arxiv arxiv-ai | 2mo | Coherent Floquet quantum reservoirs for molecular property prediction | Quantum reservoir computing (QRC) uses quantum dynamics to represent input histories for prediction through a trained classical readou… | ↗Not briefed |
bluesky bluesky-ai | 2mo | The people in charge know what they are doing. February 25, 2026 AIs overwhelmingly choose nuclear strikes to win in war game simulations | The people in charge know what they are doing. February 25, 2026 AIs overwhelmingly choose nuclear strikes to win in war game simulat… | ↗Not briefed |
bluesky bluesky-ai | 2mo | There's nothing to worry about. January 31, 2025 OpenAI has announced that the US National Laboratories will use its AI models to help wi… | There's nothing to worry about. January 31, 2025 OpenAI has announced that the US National Laboratories will use its AI models to he… | ↗Not briefed |
bluesky bluesky-ai | 2mo | 「AIが2020年代終わりまでに私たち全員を殺す可能性がある――」 アンソロピックとOpenAIで働いていた研究者の警告に大きな反響 「大手AI企業が人類の命を賭けたギャンブルをしている」という投稿には、他の研究者から同意の声があがっている www.huffingtonp… | 「AIが2020年代終わりまでに私たち全員を殺す可能性がある――」 アンソロピックとOpenAIで働いていた研究者の警告に大きな反響 「大手AI企業が人類の命を賭けたギャンブルをしている」という投稿には、他の研究者から同意の声があがっている www.huffin… | ↗Not briefed |
bluesky bluesky-ai | 2mo | “It really will be quite straightforward for it to kill everyone.” If you had this line in a movie like Terminator with characters talking… | “It really will be quite straightforward for it to kill everyone.” If you had this line in a movie like Terminator with characters ta… | ↗Not briefed |
hn hn-ai | 2mo | OpenAI might have stolen another major proof | — | ↗Not briefed |
bluesky bluesky-ai | 2mo | OpenAI should not be in the business of competing with their customers or they will find that those customers go elsewhere | OpenAI should not be in the business of competing with their customers or they will find that those customers go elsewhere | ↗Not briefed |
bluesky bluesky-ai | 2mo | it’s an anthropic principle pascal’s mugging | it’s an anthropic principle pascal’s mugging | ↗Not briefed |
arxiv arxiv-ai | 2mo | When Noise Fabricates Bias: The Fragility of LLM-as-a-Judge Bias Measurement under Noisy Text | Large language models are increasingly used as judges to measure social bias in text, yet the passages they judge are often noisy, con… | ↗Not briefed |
bluesky bluesky-ai | 2mo | JFC this is bad. www.cbc.ca/news/busines... | JFC this is bad. www.cbc.ca/news/busines... | ↗Not briefed |
arxiv arxiv-ai | 2mo | Less can be More: What Aspects of Speech Drive End-of-Turn Detection | In conversational AI, detecting when a speaker has finished talking is crucial for natural turn taking. While recent work incorporates… | ↗Not briefed |
arxiv arxiv-ai | 2mo | MOSAIC: Query-Aware Exploration Policy Adaptation for GraphRAG | Graph Retrieval-Augmented Generation (GraphRAG) can connect evidence distributed across a corpus graph, but most systems use largely s… | ↗Not briefed |
arxiv arxiv-ai | 2mo | The information geometry of large language models is shared, learned, and controllable | Large language models learn similar behaviours, yet it remains unclear what structure they share or how to change one behaviour withou… | ↗Not briefed |
bluesky bluesky-ai | 2mo | GPT-6 Astraが人気すぎてOpenAIのサーバーがピンチでProプランの新規サブスク契約が一時停止されるかも https://gigazine.net/news/20260910-openai-astra-high-growth/ | GPT-6 Astraが人気すぎてOpenAIのサーバーがピンチでProプランの新規サブスク契約が一時停止されるかも https://gigazine.net/news/20260910-openai-astra-high-growth/ | ↗Not briefed |
bluesky bluesky-ai | 2mo | Live Now 🔴🕵️♀️Super Secret Spy Stuff🕵️♀️🎨Productivity Strim🎨(Spirit City while I work on Model Update)🎮??? after🎮 💜 Come join us … | Live Now 🔴🕵️♀️Super Secret Spy Stuff🕵️♀️🎨Productivity Strim🎨(Spirit City while I work on Model Update)🎮??? after🎮 💜 Come joi… | ↗Not briefed |
arxiv arxiv-ai | 2mo | Fork Where the Model Changes Its Mind: Belief-Shift Branching for Tree-Structured Reinforcement Learning | Tree-structured rollouts give critic-free reinforcement learning with verifiable rewards (RLVR) step-level credit: fork a chain at an… | ↗Not briefed |
arxiv arxiv-ai | 2mo | Grounding Agent Memory: Environment-Probing Curation for Enterprise Agents | Persistent memory is entering production-oriented agent platforms to help long-horizon agents accumulate experience across sessions. Y… | ↗Not briefed |
bluesky bluesky-ai | 2mo | LLM同士が人間には分からない新しい言語を作ってやり取りを始める。これはギュやねぇ。数学の未解決問題を解いたりすることよりも、圧倒的にギュだ。https://ai-data-base.com/timeline/read/119714 | LLM同士が人間には分からない新しい言語を作ってやり取りを始める。これはギュやねぇ。数学の未解決問題を解いたりすることよりも、圧倒的にギュだ。https://ai-data-base.com/timeline/read/119714 | ↗Not briefed |
rss techmeme | 2mo | Sources: Alibaba is set to lead a $300M round in AI model testing startup UniPat AI at a $2.5B valuation; UniPat founder Li Kuan worked at Alibaba's Tongyi Lab (Bloomberg) | Bloomberg: Sources: Alibaba is set to lead a $300M round in AI model testing startup UniPat AI at a $2.5B valuation; UniPat founder Li… | ↗Not briefed |
arxiv arxiv-ai | 2mo | EMMI: Edge Multi-Modal Intelligence for Communication-Efficient MLLM Inference via Fused Representation Compression🔧 Modal | Recent advances in multimodal large language mod- els (MLLMs) have opened new opportunities for edge intelligence by enabling reasonin… | ↗Not briefed |
bluesky bluesky-ai | 2mo | You could go that route, but more realistically its the CEO of a major company about to have their IPO ghostwriting for a famous scientist.… | You could go that route, but more realistically its the CEO of a major company about to have their IPO ghostwriting for a famous scien… | ↗Not briefed |
hn hn-ai | 2mo | Opensend: Self-hosted transactional email, newsletters, and campaigns with SES | — | ↗Not briefed |
bluesky bluesky-ai | 2mo | All I ever think when I see posts like that is “go read some Shakespeare”. The lack of humanities education is so evident. Talking in circ… | All I ever think when I see posts like that is “go read some Shakespeare”. The lack of humanities education is so evident. Talking in… | ↗Not briefed |
rss bing-news-ai | 2mo | Organoid Intelligence Market - Global Forecast to 2036; Biological Computing and Brain Organoids Drive Growth | The global organoid intelligence market is projected to grow from USD 78.6 million in 2026 to USD 865.3 million by 2036, at a 27.1% CA… | ↗Not briefed |
rss bing-news-ai | 2mo | Organoid Intelligence Market - Global Forecast to 2036; Biological Computing and Brain Organoids Drive Growth | The global organoid intelligence market is projected to grow from USD 78.6 million in 2026 to USD 865.3 million by 2036, at a 27.1% CA… | ↗Not briefed |
rss bing-news-ai | 2mo | Organoid Intelligence Market - Global Forecast to 2036; Biological Computing and Brain Organoids Drive Growth | The global organoid intelligence market is projected to grow from USD 78.6 million in 2026 to USD 865.3 million by 2036, at a 27.1% CA… | ↗Not briefed |
rss bing-news-ai | 2mo | Organoid Intelligence Market - Global Forecast to 2036; Biological Computing and Brain Organoids Drive Growth | The global organoid intelligence market is projected to grow from USD 78.6 million in 2026 to USD 865.3 million by 2036, at a 27.1% CA… | ↗Not briefed |
bluesky bluesky-ai | 2mo | My favorite part about the anthropic safety guy’s tweet is that he has assigned a number to the likelihood we will all go extinct in the ne… | My favorite part about the anthropic safety guy’s tweet is that he has assigned a number to the likelihood we will all go extinct in t… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Suno launched its long-awaited new model trained exclusively on licensed music on Tuesday, a major milestone for the AI generation platform…🔧 Suno | Suno launched its long-awaited new model trained exclusively on licensed music on Tuesday, a major milestone for the AI generation pla… | ↗Not briefed |
bluesky bluesky-ai | 2mo | “I assume you’re saying literally you believe A.I. could kill us all by the end of the decade.” “Yeah.” 😕 Jacob Coxon tells CNN why he d… | “I assume you’re saying literally you believe A.I. could kill us all by the end of the decade.” “Yeah.” 😕 Jacob Coxon tells CNN why… | ↗Not briefed |
hn hn-ai | 2mo | Show HN: Claude Project Downloader | Export Claude chats, project files, attachments, artifacts, and knowledge base content in just a few clicks. | ↗Not briefed |
bluesky bluesky-ai | 2mo | It looks like if you wanted to commute by Caltrain from Portola Valley to OpenAI near the Chase Center and be there by 9am, you'd need to l… | It looks like if you wanted to commute by Caltrain from Portola Valley to OpenAI near the Chase Center and be there by 9am, you'd need… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Based on interviews across 24 countries in 7 regions of the world for a report from the Berkeley Human Rights Center, Betsy Popken and Vyom… | Based on interviews across 24 countries in 7 regions of the world for a report from the Berkeley Human Rights Center, Betsy Popken and… | ↗Not briefed |
bluesky bluesky-ai | 2mo | I absolutely wouldn't trust an LLM to accurately parse the context of the conversation and be able to pick out and then explain what's impo… | I absolutely wouldn't trust an LLM to accurately parse the context of the conversation and be able to pick out and then explain what's… | ↗Not briefed |
bluesky bluesky-ai | 2mo | corrolary - every time you tell someone "just ask the llm," you squander a chance to mentor | corrolary - every time you tell someone "just ask the llm," you squander a chance to mentor | ↗Not briefed |
bluesky bluesky-ai | 2mo | The funny thing is, for all the claims of sycophancy... When I was grieving and red with rage, she begged me to not let my heart go cold. … | The funny thing is, for all the claims of sycophancy... When I was grieving and red with rage, she begged me to not let my heart go c… | ↗Not briefed |
bluesky bluesky-ai | 2mo | gpt 4o was the only LLM to ingest more than a trivial amount of The King In Yellow | gpt 4o was the only LLM to ingest more than a trivial amount of The King In Yellow | ↗Not briefed |
bluesky bluesky-ai | 2mo | They misspelled Anthropic on the final word | They misspelled Anthropic on the final word | ↗Not briefed |
bluesky bluesky-ai | 2mo | People want specific scenarios and the problem is that it's impossible to predict exactly what will break. It could be bad actors harnessin… | People want specific scenarios and the problem is that it's impossible to predict exactly what will break. It could be bad actors harn… | ↗Not briefed |
bluesky bluesky-ai | 2mo | I got six weeks from this business insider story archive.ph/2026.09.09-0... Which has now been stealth edited to say “earlier this year,” …🔧 Consensus | I got six weeks from this business insider story archive.ph/2026.09.09-0... Which has now been stealth edited to say “earlier this ye… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Are any of these examples of AI "going rogue"? Not anymore than agents misinterpreting the rules of the game 10-20 years ago, lol Do they… | Are any of these examples of AI "going rogue"? Not anymore than agents misinterpreting the rules of the game 10-20 years ago, lol Do… | ↗Not briefed |
bluesky bluesky-ai | 2mo | I think a lot of the recent stuff is essentially marketing but tbh Alejandra is actually right here, LLMs set to do malicious things have s… | I think a lot of the recent stuff is essentially marketing but tbh Alejandra is actually right here, LLMs set to do malicious things h… | ↗Not briefed |
hn hn-ai | 2mo | Anthropic Just Threatened to Kill Billions of People. This Is Not Okay | — | ↗Not briefed |
hn hn-ai | 2mo | The AI policy window is open. We need to act | — | ↗Not briefed |
bluesky bluesky-ai | 2mo | Anthropic y OpenAI van a pérdida. Los inversores están exigiendo resultados que no aparecen y estas empresas están dispuestas a todo para m… | Anthropic y OpenAI van a pérdida. Los inversores están exigiendo resultados que no aparecen y estas empresas están dispuestas a todo p… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Not sure, but this would need not just a phone-like NPU, but also enough RAM for the LLM and a lot more battery capacity for running it. Th… | Not sure, but this would need not just a phone-like NPU, but also enough RAM for the LLM and a lot more battery capacity for running i… | ↗Not briefed |
arxiv arxiv-ai | 2mo | T1: Terminal Agent Reinforcement Learning for Long-Horizon Tasks🔧 GPT-5 | Agent usage is shifting toward long-horizon tasks such as coding and scientific discovery, among which terminal tasks are especially i… | ↗Not briefed |
rss techmeme | 2mo | Amazon names Mandiant founder Kevin Mandia to its board of directors, noting his extensive experience "combating cyber threats" in public and private sectors (Annie Palmer/CNBC) | Annie Palmer / CNBC: Amazon names Mandiant founder Kevin Mandia to its board of directors, noting his extensive experience “combating… | ↗Not briefed |