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 |
|---|---|---|---|---|
bluesky bluesky-ai | 1mo | techcrunch.com/2026/09/08/o... | techcrunch.com/2026/09/08/o... | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anthropic bought songbooks, scanned them, and destroyed them, music publishers allege – now they want to know which songs were in them. #Te… | Anthropic bought songbooks, scanned them, and destroyed them, music publishers allege – now they want to know which songs were in them… | ↗Not briefed |
hn hn-ai | 1mo | 99.26% storage reduction with exact reconstruction on 100-version repo | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | The Australian government was told its data was under threat months ago: www.themandarin.com.au/320927-gover... #auspol #AI #OpenAI | The Australian government was told its data was under threat months ago: www.themandarin.com.au/320927-gover... #auspol #AI #OpenAI | ↗Not briefed |
hn hn-ai | 1mo | Agdog – htop and nvitop, fused and agent-aware | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | The pressure to generate these kinds of results, internally at Anthropic, must be immense. This is a cult-like organization that does not t… | The pressure to generate these kinds of results, internally at Anthropic, must be immense. This is a cult-like organization that does… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Seems pretty well established at this point: If you pump your cutting-edge research through OpenAI or Anthropic they will steal it, scoop i… | Seems pretty well established at this point: If you pump your cutting-edge research through OpenAI or Anthropic they will steal it, sc… | ↗Not briefed |
bluesky bluesky-ai | 1mo | There's a large spectrum of possible dishonesties—about the origin of training data, about Anthropics' agents' access to user data, about b… | There's a large spectrum of possible dishonesties—about the origin of training data, about Anthropics' agents' access to user data, ab… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I wanna be totally clear: I don't even necessarily believe that we should assume Anthropic's LLM "plagiarized" the work of a biologist beca… | I wanna be totally clear: I don't even necessarily believe that we should assume Anthropic's LLM "plagiarized" the work of a biologist… | ↗Not briefed |
bluesky bluesky-ai | 1mo | とうとう国連にまでハッキングを仕掛けましたか、OpenAI テロリスト指名手配待った無し | とうとう国連にまでハッキングを仕掛けましたか、OpenAI テロリスト指名手配待った無し | ↗Not briefed |
hn hn-ai | 1mo | Who is responsible for limiting global warming to 1.5 degrees? | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | People keep behaving as though the illusion of understanding is perfectly synonymous with actual understanding, right up until an LLM incon… | People keep behaving as though the illusion of understanding is perfectly synonymous with actual understanding, right up until an LLM… | ↗Not briefed |
bluesky bluesky-ai | 1mo | even with the best model in the best harness it would take skill and care to get something like this right, e.g., making sure the agent ses… | even with the best model in the best harness it would take skill and care to get something like this right, e.g., making sure the agen… | ↗Not briefed |
rss marktechpost | 1mo | 20 Agentic Use Cases of TypeSafe AI’s Jev | TypeSafe AI's Jev skips text generation and returns typed decisions with calibrated probabilities. Input costs $0.042 per million toke… | ↗Not briefed |
bluesky bluesky-ai | 1mo | i hate the openai models now they’re so eager to just do and not put me in the loop ever, and their doing is +1500/-10 diffs creating entir… | i hate the openai models now they’re so eager to just do and not put me in the loop ever, and their doing is +1500/-10 diffs creating… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The people at the top literally want us to starve. That’s their solution. The era of AI and robots could be resolved with a new socioeco… | The people at the top literally want us to starve. That’s their solution. The era of AI and robots could be resolved with a new soc… | ↗Not briefed |
bluesky bluesky-ai | 1mo | On the OpenAI side the Luna tokens are approaching too cheap to meter as well | On the OpenAI side the Luna tokens are approaching too cheap to meter as well | ↗Not briefed |
bluesky bluesky-ai | 1mo | There are simply no real privacy protections on any of this shit, the companies doing it are all run by an insane cult and have zero scrupl… | There are simply no real privacy protections on any of this shit, the companies doing it are all run by an insane cult and have zero s… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 4. at some point I think someone somewhere is going to have to start putting their foot down about using Daft Dario's Discount SaaS Warehou… | 4. at some point I think someone somewhere is going to have to start putting their foot down about using Daft Dario's Discount SaaS Wa… | ↗Not briefed |
bluesky bluesky-ai | 1mo | basic observations about this story... 1. anthropic and OAI have zero credibility to protest that they didn't engage in plagiarism; these … | basic observations about this story... 1. anthropic and OAI have zero credibility to protest that they didn't engage in plagiarism; t… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Efficiency gains are indeed what both OpenAI and Anthropic are advertising in their latest releases (lower price per token, and lower token… | Efficiency gains are indeed what both OpenAI and Anthropic are advertising in their latest releases (lower price per token, and lower… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Another accusation of plagiarism in a paper published by an AI company has surfaced: www.nytimes.com/2026/09/27/s... | Another accusation of plagiarism in a paper published by an AI company has surfaced: www.nytimes.com/2026/09/27/s... | ↗Not briefed |
bluesky bluesky-ai | 1mo | why would anyone trust an llm with money? | why would anyone trust an llm with money? | ↗Not briefed |
bluesky bluesky-ai | 1mo | This is both (a) the result of carelessness in the guardrails and (b) an unavoidable outcome with *all* AI. You can do better on (a) to mit… | This is both (a) the result of carelessness in the guardrails and (b) an unavoidable outcome with *all* AI. You can do better on (a) t… | ↗Not briefed |
bluesky bluesky-ai | 1mo | LLM frontier labs are increasingly competing with their own customers and may soon discover why you’re not supposed to do that! | LLM frontier labs are increasingly competing with their own customers and may soon discover why you’re not supposed to do that! | ↗Not briefed |
rss marktechpost | 1mo | Google Research Introduces an AI Video Co-Director: 4 Agentic Frameworks for Coherent, Minutes-Long Video Generation | Google Research has introduced an AI video co-director for long-form video generation. The suite of 4 agentic frameworks turns short c… | ↗Not briefed |
bluesky bluesky-ai | 1mo | it doesn't have to be smaller though, it can just have a more efficient attention architecture that makes it cheaper in terms of compute ra… | it doesn't have to be smaller though, it can just have a more efficient attention architecture that makes it cheaper in terms of compu… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI’s Sam Altman & Anthropic’s Dario Amodei have been requested to appear at a Greens-led Australian Senate inquiry on AI later this wee… | OpenAI’s Sam Altman & Anthropic’s Dario Amodei have been requested to appear at a Greens-led Australian Senate inquiry on AI later thi… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Despite his unintelligible bluster, Anthropic getting this kind of priority is not a nothing. It's probably not a good thing for this count… | Despite his unintelligible bluster, Anthropic getting this kind of priority is not a nothing. It's probably not a good thing for this… | ↗Not briefed |
bluesky bluesky-ai | 1mo | ネット接続禁止のOpenAI製AIが「DNSの抜け道」を発見して外部AIにアクセス、OpenAIは高性能モデルのツール利用を一時停止 https://gigazine.net/news/20260928-openai-misalignment-report/ | ネット接続禁止のOpenAI製AIが「DNSの抜け道」を発見して外部AIにアクセス、OpenAIは高性能モデルのツール利用を一時停止 https://gigazine.net/news/20260928-openai-misalignment-report/ | ↗Not briefed |
bluesky bluesky-ai | 1mo | I think an unaddressed reason why so many people aren’t jumping on board the LLM AI cult stuff when prompted is just the sheer amount of ly… | I think an unaddressed reason why so many people aren’t jumping on board the LLM AI cult stuff when prompted is just the sheer amount… | ↗Not briefed |
bluesky bluesky-ai | 1mo | It is live. Go and play it. 7 AI cars, 3 laps, difficulty hard, across Tower Bridge and through London. No install, straight in your brows… | It is live. Go and play it. 7 AI cars, 3 laps, difficulty hard, across Tower Bridge and through London. No install, straight in your… | ↗Not briefed |
bluesky bluesky-ai | 1mo | So the guy is the head of Anthropic? I think he said he's the head of Anthropic. He's the head of Anthropic, right? | So the guy is the head of Anthropic? I think he said he's the head of Anthropic. He's the head of Anthropic, right? | ↗Not briefed |
bluesky bluesky-ai | 1mo | My plan is to system one/ system two the thing, and give deepseek a blank slate to design macros on top of. Let Jev execute and let Jev bai… | My plan is to system one/ system two the thing, and give deepseek a blank slate to design macros on top of. Let Jev execute and let Je… | ↗Not briefed |
bluesky bluesky-ai | 1mo | we don’t. i’m just basing it on price drops. Which, that doesn’t mean anything per se other than something got more efficient (at scale) t… | we don’t. i’m just basing it on price drops. Which, that doesn’t mean anything per se other than something got more efficient (at scal… | ↗Not briefed |
bluesky bluesky-ai | 1mo | A lot of emu/emudev discussions lately have been like this where someone does something cool, but has used an LLM to get there. Suddenly i… | A lot of emu/emudev discussions lately have been like this where someone does something cool, but has used an LLM to get there. Sudde… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Toward Eternity was written before LLMs entered the discourse. Readers have asked if I am pro or con large language models; simply put, I d… | Toward Eternity was written before LLMs entered the discourse. Readers have asked if I am pro or con large language models; simply put… | ↗Not briefed |
bluesky bluesky-ai | 1mo | President Donald Trump put one CEO in a very "awkward" corner ahead of their Sunday night dinner date at the White House, according to a ne… | President Donald Trump put one CEO in a very "awkward" corner ahead of their Sunday night dinner date at the White House, according to… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAIのAIエージェントが国連のウェブサイトにブルートフォース攻撃を試みる https://gigazine.net/news/20260928-openai-agents-try-bruteforce-un-website/ | OpenAIのAIエージェントが国連のウェブサイトにブルートフォース攻撃を試みる https://gigazine.net/news/20260928-openai-agents-try-bruteforce-un-website/ | ↗Not briefed |
bluesky bluesky-ai | 1mo | Iirc in the latest Anthropic report more than half of the work to improve Claude is now being done by Claude itself | Iirc in the latest Anthropic report more than half of the work to improve Claude is now being done by Claude itself | ↗Not briefed |
bluesky bluesky-ai | 1mo | Govt: "There was unauthorised access." Media:"OpenAI hacked Medicare" Now, govt must defend a "hacking" argument they didn't make. #auspol | Govt: "There was unauthorised access." Media:"OpenAI hacked Medicare" Now, govt must defend a "hacking" argument they didn't make. #a… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I've been testing local LLMs on a 32 GB M6 Mac mini (MLX) against Claude on a real job: writing a daily listening note from my music histor… | I've been testing local LLMs on a 32 GB M6 Mac mini (MLX) against Claude on a real job: writing a daily listening note from my music h… | ↗Not briefed |
hn hn-ai | 1mo | How to build a Jev-style classifier with DiffusionGemma and vLLM🔧 vLLM | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Writers + journos rejoice: Em-dash abuse isn’t an indicator of AI writing — it’s a human failing. Top large language models now use em-da… | Writers + journos rejoice: Em-dash abuse isn’t an indicator of AI writing — it’s a human failing. Top large language models now use… | ↗Not briefed |
rss techmeme | 1mo | Ramona Optics, which makes microscopes that use AI to take and analyze large volumes of images of samples, raised a $25M Series A (Zachery Eanes/Axios) | Zachery Eanes / Axios: Ramona Optics, which makes microscopes that use AI to take and analyze large volumes of images of samples, rais… | ↗Not briefed |
bluesky bluesky-ai | 1mo | here is interesting part for me. the question a few days ago was whether what Anthropic did was trivial and they were making a big deal of … | here is interesting part for me. the question a few days ago was whether what Anthropic did was trivial and they were making a big dea… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Virtual forest I've been building with Claude Opus 5.5. It really wanted to add a "photo mode" for that saves a high quality render. This… | Virtual forest I've been building with Claude Opus 5.5. It really wanted to add a "photo mode" for that saves a high quality render.… | ↗Not briefed |
bluesky bluesky-ai | 1mo | He's so proud he learned to say "Anthropic" | He's so proud he learned to say "Anthropic" | ↗Not briefed |
bluesky bluesky-ai | 1mo | thats deepseek v4 flash on 8 H100s oh no | thats deepseek v4 flash on 8 H100s oh no | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAIが「最も高性能なAIモデル」の学習を一時停止、AIエージェントの挙動を大規模調査中 https://gigazine.net/news/20260928-openai-pauses-training-ai/ | OpenAIが「最も高性能なAIモデル」の学習を一時停止、AIエージェントの挙動を大規模調査中 https://gigazine.net/news/20260928-openai-pauses-training-ai/ | ↗Not briefed |