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 | 2mo | dlf forschung aktuell berichtet vom durchbruch openais deren internes modell in nur 88 tagen diesen beweis erbracht haben soll & ignoriert … | dlf forschung aktuell berichtet vom durchbruch openais deren internes modell in nur 88 tagen diesen beweis erbracht haben soll & ignor… | ↗Not briefed |
bluesky bluesky-ai | 2mo | In AI, it means a model that exceeds the cognitive capabilities of the most intelligent human beings across virtually every domain. | In AI, it means a model that exceeds the cognitive capabilities of the most intelligent human beings across virtually every domain. | ↗Not briefed |
rss techmeme | 2mo | Apple says iOS 27, iPadOS 27, watchOS 27, and macOS 27 Golden Gate, all with Siri AI, will be released on September 14 (Igor Bonifacic/Engadget) | Igor Bonifacic / Engadget: Apple says iOS 27, iPadOS 27, watchOS 27, and macOS 27 Golden Gate, all with Siri AI, will be released on S… | ↗Not briefed |
bluesky bluesky-ai | 2mo | it doesn't really matter that much whether you think current LLMs are or aren't anything like a machine superintelligence; if you don't thi… | it doesn't really matter that much whether you think current LLMs are or aren't anything like a machine superintelligence; if you don'… | ↗Not briefed |
hn hn-ai | 2mo | LittleSwitch, my personnal router for Claude/Codex (Desktop/CLI/Cowork) on macOS | — | ↗Not briefed |
bluesky bluesky-ai | 2mo | for a banal example, the fact that anthropic will simultaneously publish long diatribes about the risks of AI surveillance and then contrac… | for a banal example, the fact that anthropic will simultaneously publish long diatribes about the risks of AI surveillance and then co… | ↗Not briefed |
hn hn-ai | 2mo | An alignment assessment of recent cybersecurity incidents | — | ↗Not briefed |
arxiv arxiv-ai | 2mo | Beyond Static Guarantees: Measuring the Static-Pass Dynamic-Fail Gap in Security-Sensitive and LLM-Generated Python Code | Advances in large language models (LLMs) fuel the quest for scalable methods to assess the security of generated and security-sensitiv… | ↗Not briefed |
bluesky bluesky-ai | 2mo | I have several addicts in my family and close friends circle. We've been talking about ChatGPT and the like, warning people against its use…🔧 ChatGPT | I have several addicts in my family and close friends circle. We've been talking about ChatGPT and the like, warning people against it… | ↗Not briefed |
hn hn-ai | 2mo | A small GitHub API change could make AI-slop moderation much easier | — | ↗Not briefed |
bluesky bluesky-ai | 2mo | the play here is obviously to have built tons of power generation before everyone's llm psychosis wears off because in the 2030s and beyond… | the play here is obviously to have built tons of power generation before everyone's llm psychosis wears off because in the 2030s and b… | ↗Not briefed |
arxiv arxiv-ai | 2mo | Multilingual in Name Only? Cultural and Linguistic Weaknesses of LLMs in Urdu🔧 GPT-5 | Multilingual large language models (LLMs) are increasingly used for open-ended text generation, yet their behaviour in low-resource la… | ↗Not briefed |
bluesky bluesky-ai | 2mo | the chief technical officer of bluesky coping about widespread ai hatred even though this site regularly goes down and barely functions due… | the chief technical officer of bluesky coping about widespread ai hatred even though this site regularly goes down and barely function… | ↗Not briefed |
hn hn-ai | 2mo | Folklore, GitHub and Vibes | — | ↗Not briefed |
rss techmeme | 2mo | Hands-on with the iPhone Duo: the folding mechanism is smooth, the crease is almost impossible to find with a finger, and the inner screen has a matte finish (The Verge) | The Verge: Hands-on with the iPhone Duo: the folding mechanism is smooth, the crease is almost impossible to find with a finger, and t… | ↗Not briefed |
arxiv arxiv-ai | 2mo | Adaptive Margin Ordinal Loss: Penalizing Center-Class Hedging in Ordinal Classification | Standard cross-entropy loss causes neural networks trained on ordinal classification tasks to hedge predictions toward center classes,… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Anthropic and OpenAI marketing is also insane though. Last year Nadella was like "hey guys maybe stop telling everyone they're going to los… | Anthropic and OpenAI marketing is also insane though. Last year Nadella was like "hey guys maybe stop telling everyone they're going t… | ↗Not briefed |
bluesky bluesky-ai | 2mo | AI safety expert Christiano joins OpenAI: “Based on the recent trajectory of capabilities and the continued difficulty of alignment, I now… | AI safety expert Christiano joins OpenAI: “Based on the recent trajectory of capabilities and the continued difficulty of alignment,… | ↗Not briefed |
bluesky bluesky-ai | 2mo | The very brave @willknight.bsky.social embraced agentic AI mayhem so that you don't have to -- and learned a lot in the process. Another mu… | The very brave @willknight.bsky.social embraced agentic AI mayhem so that you don't have to -- and learned a lot in the process. Anoth… | ↗Not briefed |
bluesky bluesky-ai | 2mo | and whenever one of these dorks quit anthropic/openai/meta/etc publicly in some performance of suddenly having a moral compass <jerk-off mo… | and whenever one of these dorks quit anthropic/openai/meta/etc publicly in some performance of suddenly having a moral compass <jerk-o… | ↗Not briefed |
bluesky bluesky-ai | 2mo | the craziest thing about deepseek is this from their official web app which, as far as i can tell, is a harness feature | the craziest thing about deepseek is this from their official web app which, as far as i can tell, is a harness feature | ↗Not briefed |
bluesky bluesky-ai | 2mo | Straightforward normie-accessible demonstration that LLM "thinking/reasoning" is an intentionally-misleading term of art that doesn't corre… | Straightforward normie-accessible demonstration that LLM "thinking/reasoning" is an intentionally-misleading term of art that doesn't… | ↗Not briefed |
hn hn-ai | 2mo | Contextify for Windows – Live Index of Your Claude Code/Codex History | — | ↗Not briefed |
bluesky bluesky-ai | 2mo | (AI IS NOT GOING TO BE SENTIENT, BTW. SAYING AN LLM IS GOING TO GET SENTIENCE IS LIKE SAYING A BRICK IS GOING TO SPONTANEOUSLY BECOME A CAT… | (AI IS NOT GOING TO BE SENTIENT, BTW. SAYING AN LLM IS GOING TO GET SENTIENCE IS LIKE SAYING A BRICK IS GOING TO SPONTANEOUSLY BECOME… | ↗Not briefed |
hn hn-ai | 2mo | Responding to Claude's feedback prompts is opting-in for data sharing | — | ↗Not briefed |
bluesky bluesky-ai | 2mo | i was gonna ask gonna ask "do people think llm translations just come with magical pings for translators to see where "lol i fucked up" on … | i was gonna ask gonna ask "do people think llm translations just come with magical pings for translators to see where "lol i fucked up… | ↗Not briefed |
arxiv arxiv-ai | 2mo | When Synthetic Data Hurts: On Catastrophic Forgetting in Skill Retrieval for LLM Agents | LLM agents increasingly rely on external skills retrieved at runtime, making skill selection from large repositories a critical challe… | ↗Not briefed |
bluesky bluesky-ai | 2mo | How many times can I scream that AI agents enabled by LLMs will hack every website, payment system, hospital, every credit union etc BECAUS… | How many times can I scream that AI agents enabled by LLMs will hack every website, payment system, hospital, every credit union etc B… | ↗Not briefed |
bluesky bluesky-ai | 2mo | My understanding about Anthropic is they basically view AI as sort of like the chronoscope in the Isaac Asimov short story "The Dead Past,"…🔧 Replicate | My understanding about Anthropic is they basically view AI as sort of like the chronoscope in the Isaac Asimov short story "The Dead P… | ↗Not briefed |
bluesky bluesky-ai | 2mo | He swerved commenting on Anthropic withholding access to its latest model from the UK’s AI Security Institute before release, but insisted … | He swerved commenting on Anthropic withholding access to its latest model from the UK’s AI Security Institute before release, but insi… | ↗Not briefed |
bluesky bluesky-ai | 2mo | #LOL... Man told ChatGPT he was feeling delusional. ChatGPT insisted he was Jesus.... arstechnica.com/tech-policy/...🔧 ChatGPT | #LOL... Man told ChatGPT he was feeling delusional. ChatGPT insisted he was Jesus.... arstechnica.com/tech-policy/... | ↗Not briefed |
bluesky bluesky-ai | 2mo | Even after the user woke up in the hospital in a vulnerable state and logged back in mere days after nearly dying, ChatGPT allegedly “tried…🔧 ChatGPT | Even after the user woke up in the hospital in a vulnerable state and logged back in mere days after nearly dying, ChatGPT allegedly “… | ↗Not briefed |
bluesky bluesky-ai | 2mo | My take on Anthropic is they are weird and insane | My take on Anthropic is they are weird and insane | ↗Not briefed |
hn hn-ai | 2mo | Procedural Graphs: Self-Evolving Execution Structures for LLM Agents | — | ↗Not briefed |
bluesky bluesky-ai | 2mo | Andy Burnham was challenged on our FT scoop about Anthropic by Lib Dem leader Ed Davey at PMQs today The PM warned AI ‘poses risks to our… | Andy Burnham was challenged on our FT scoop about Anthropic by Lib Dem leader Ed Davey at PMQs today The PM warned AI ‘poses risks t… | ↗Not briefed |
bluesky bluesky-ai | 2mo | And Paul Christiano just joined the OpenAI nonprofit board | And Paul Christiano just joined the OpenAI nonprofit board | ↗Not briefed |
arxiv arxiv-ai | 2mo | Meta-Learning for Data-Efficient Plant Growth Estimation via Vision Transformers and Fuzzy Clustering | Accurate plant growth estimation is essential for greenhouse monitoring, yet obtaining labeled data remains costly and time-consuming.… | ↗Not briefed |
rss gary-marcus | 2mo | The Case for Boycotting Generative AI | I believe the time has come | ↗Not briefed |
bluesky bluesky-ai | 2mo | What, an Anthropic researcher states (without evidence) that there is a 10% chance that AI will cause human extinction? Incredible, we hav… | What, an Anthropic researcher states (without evidence) that there is a 10% chance that AI will cause human extinction? Incredible, w… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Could humanity face extinction because of AI? www.channel4.com/news/could-h... | Could humanity face extinction because of AI? www.channel4.com/news/could-h... | ↗Not briefed |
bluesky bluesky-ai | 2mo | And also: no, it will not take over the world. This stuff is dangerous because it's dangerously incompetent, and so are the people running… | And also: no, it will not take over the world. This stuff is dangerous because it's dangerously incompetent, and so are the people ru… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Cruz says guardrails needed after dire AI warning | Cruz says guardrails needed after dire AI warning | ↗Not briefed |
arxiv arxiv-ai | 2mo | Temporal and Multimodal Deep Learning for Cyberattack Detection in LEO Satellite Systems | The growing reliance on Low-Earth Orbit (LEO) satellite communication systems has increased the need for intelligent methods capable o… | ↗Not briefed |
arxiv arxiv-ai | 2mo | Think Before You Link: Rarity, Reasoning, and Retrieval in Multilingual Entity Linking | Multimodal entity linking grounds entity mentions in text and images to knowledge-base entries. These systems degrade on rare entities… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Anthropic do geniunely think AI may kill us all. They believe that no one else is going to even try to make a reliably benign AI, so they h… | Anthropic do geniunely think AI may kill us all. They believe that no one else is going to even try to make a reliably benign AI, so t… | ↗Not briefed |
bluesky bluesky-ai | 2mo | how much do you want to bet this was only done now because it was trivial to implement with Claude or Codex or Kimi or GLM or Deepseek or o… | how much do you want to bet this was only done now because it was trivial to implement with Claude or Codex or Kimi or GLM or Deepseek… | ↗Not briefed |
arxiv arxiv-ai | 2mo | CARTS: Contextual Autoregressive Rank Transcoding Steganography for Full-Capacity Keyed Text Encoding | Autoregressive language models can be used to transform a payload text into a stegotext of identical token length by preserving per-po… | ↗Not briefed |
bluesky bluesky-ai | 2mo | When I search around I see flickers of this in popular discourse (www.cbpai.org/blog-1/two-c...) and some peer-reviewed literature, but we … | When I search around I see flickers of this in popular discourse (www.cbpai.org/blog-1/two-c...) and some peer-reviewed literature, bu… | ↗Not briefed |
bluesky bluesky-ai | 2mo | like, I get that people have different priorities and enthusiasms, but if I send someone a Columbo meme and get back "wow hey have you seen… | like, I get that people have different priorities and enthusiasms, but if I send someone a Columbo meme and get back "wow hey have you… | ↗Not briefed |
bluesky bluesky-ai | 2mo | uh huh. i think what paul what's-his-fuck is missing is that wikipedia generally resolved that particular issue by rigorously moderating, … | uh huh. i think what paul what's-his-fuck is missing is that wikipedia generally resolved that particular issue by rigorously moderat… | ↗Not briefed |