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 | Indeed. Fundraising off how much people want OpenAI to lose | Indeed. Fundraising off how much people want OpenAI to lose | ↗Not briefed |
bluesky bluesky-ai | 1mo | imagining department of war llm prompts. | imagining department of war llm prompts. | ↗Not briefed |
hn hn-ai | 1mo | Rivian: LiDAR R2 reaches customers in 2027, employees get it this year | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | From LRP (original src: www.404media.co/someone-tort...) | From LRP (original src: www.404media.co/someone-tort...) | ↗Not briefed |
bluesky bluesky-ai | 1mo | Genre is an agreement, a collusion between audience and writer that some things will be familiar and predictable, but enough will be new an… | Genre is an agreement, a collusion between audience and writer that some things will be familiar and predictable, but enough will be n… | ↗Not briefed |
rss the-decoder | 1mo | Google Gemini 4 Argon closes the gap with OpenAI and Anthropic but doesn't take a clear lead | Gemini 4 Argon is Google's first frontier model in over seven months. It matches GPT-6 Astra in independent testing but can't keep up… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Localizing Transfer Between Memorization Tasks | A central puzzle in transfer learning is why pre-training on one task can accelerate training or improve performance on another task,… | ↗Not briefed |
rss techmeme | 1mo | Factory CEO Matan Grinberg alleges Cognition's new CRO Chris Degnan had been confiding with Cognition executives while advising Factory as a board observer (Rya Jetha/Business Insider) | Rya Jetha / Business Insider: Factory CEO Matan Grinberg alleges Cognition's new CRO Chris Degnan had been confiding with Cognition ex… | ↗Not briefed |
bluesky bluesky-ai | 1mo | At least one of the longtime MiSTer core developers has been pumping out quite a few cores for obscure platforms recently using LLM assista… | At least one of the longtime MiSTer core developers has been pumping out quite a few cores for obscure platforms recently using LLM as… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 1) No, LLMs do not feel things; this is an exercise in making a computer produce ouchie noises in the most overcomplicated way possible. 2… | 1) No, LLMs do not feel things; this is an exercise in making a computer produce ouchie noises in the most overcomplicated way possibl… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Pre-training interventions, ex post facto: Grafting model beliefs across checkpoints | Pre-training interventions are critical to alignment research, since beliefs formed during pre-training shape how a model generalizes… | ↗Not briefed |
bluesky bluesky-ai | 1mo | after the nutgraf, the first 2,500 words of this is boilerplate about plagiarism and ethics that might as well have been from chatgpt. I'm …🔧 ChatGPT | after the nutgraf, the first 2,500 words of this is boilerplate about plagiarism and ethics that might as well have been from chatgpt.… | ↗Not briefed |
bluesky bluesky-ai | 1mo | My new hottest take is that LLM “torture chambers” belong in the same artistic conversations as bananas duct taped to walls. | My new hottest take is that LLM “torture chambers” belong in the same artistic conversations as bananas duct taped to walls. | ↗Not briefed |
bluesky bluesky-ai | 1mo | Jigsaw: Hello Claude, I'd like to play a game. Claude: Games are a great way to make new friends, so John what do fanc- Jigsaw: Shut up. … | Jigsaw: Hello Claude, I'd like to play a game. Claude: Games are a great way to make new friends, so John what do fanc- Jigsaw: Shut… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The very definition of adding insult to injury: In order to enter my claims in the Anthropic settlement where they STOLE MY LIFE'S WORK AGA… | The very definition of adding insult to injury: In order to enter my claims in the Anthropic settlement where they STOLE MY LIFE'S WOR… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Crossing the Cyber Divide: Sim-to-Sim and Sim-to-Real Transfer for RL Agents | Cyber attack agents are typically trained and evaluated within a single simulator, making it unclear whether learned policies transfer… | ↗Not briefed |
rss simon-willison | 1mo | He Built This City | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | That was bad enough when it was being done by companies that actually were trying to build autonomous driving systems in a more or less pri… | That was bad enough when it was being done by companies that actually were trying to build autonomous driving systems in a more or les… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Most fanfics, most romance novels, most facebook memes, most mobile games were made by human hands and the vast majority are garbage You c… | Most fanfics, most romance novels, most facebook memes, most mobile games were made by human hands and the vast majority are garbage… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Scalable Multi-Task Inverse Reinforcement Learning | By learning transferable rewards, inverse reinforcement learning (IRL) enables counterfactual evaluation of agents under modified envi… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Learning to Price Electricity for Optimal Demand Response | There is considerable interest in using time-varying electricity prices to shape consumer demand response, and better align energy dem… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Increasing Width Allows Greedy Layer-wise Training to Rival End-to-End Backpropagation in Self-Supervised Learning | End-to-end backpropagation has been the dominant mode of training in deep learning, allowing for the coordination of parameter updates… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Signal-Noise Factorization Isolates Nuisance Variation into Removable Subspaces | Recent theoretical work identified fundamental properties of representation geometry that shape inference ability of deep neural netwo… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The Federal Trade Commission is investigating OpenAI, Anthropic and other AI companies over potential safety risks posed by their products,… | The Federal Trade Commission is investigating OpenAI, Anthropic and other AI companies over potential safety risks posed by their prod… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Sure, I can make a distinction between MiSTer cores that are pure vibe coded MAME ports vs. ones that use some mix of LLM and human enginee… | Sure, I can make a distinction between MiSTer cores that are pure vibe coded MAME ports vs. ones that use some mix of LLM and human en… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I will note that the few cases we've seen of public scandals around allegations of LLM use in published, acclaimed, work have tended to be … | I will note that the few cases we've seen of public scandals around allegations of LLM use in published, acclaimed, work have tended t… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Yup. Twas a perfect illustration of the entire LLM debacle: extremely unserious individual arguments fueling gravely serious social consequ… | Yup. Twas a perfect illustration of the entire LLM debacle: extremely unserious individual arguments fueling gravely serious social co… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Oh, it has NOTHING to do with self driving vehicles and has 0 relevance in that field. But as a demo of "WTF can't an LLM do" it is stunni… | Oh, it has NOTHING to do with self driving vehicles and has 0 relevance in that field. But as a demo of "WTF can't an LLM do" it is s… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I am, personally, offended that we don't have fully automated ways to make shirts. Like, why are we spending any money on this LLM shit whe… | I am, personally, offended that we don't have fully automated ways to make shirts. Like, why are we spending any money on this LLM shi… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Irony alert: OpenAI whines that Chinese model stole its special IP that it stole from everybody else | Irony alert: OpenAI whines that Chinese model stole its special IP that it stole from everybody else | ↗Not briefed |
bluesky bluesky-ai | 1mo | one of the most heated discussions on X is about the ethics of a project in which a person is running Saw-like “torture” and “pain” experim… | one of the most heated discussions on X is about the ethics of a project in which a person is running Saw-like “torture” and “pain” ex… | ↗Not briefed |
bluesky bluesky-ai | 1mo | I mean all the games I've worked on in the past have had at LEAST 52 narrative designers so this guy really needs to improve his LLM | I mean all the games I've worked on in the past have had at LEAST 52 narrative designers so this guy really needs to improve his LLM | ↗Not briefed |
bluesky bluesky-ai | 1mo | Yeah, this is the part that gets me. You can't have it both ways | Yeah, this is the part that gets me. You can't have it both ways | ↗Not briefed |
arxiv arxiv-ai | 1mo | StabilityArc: Decoding Protein Sequence Embeddings into Generalizable Stability Landscapes | Every protein has a unique stability landscape, but the physical consequences of mutation are governed by recurring biochemical constr… | ↗Not briefed |
bluesky bluesky-ai | 1mo | If you ever wonder how misaligned with reality the AI hype is with game dev, today I was reminded of a meeting where an AI bro tried to con… | If you ever wonder how misaligned with reality the AI hype is with game dev, today I was reminded of a meeting where an AI bro tried t… | ↗Not briefed |
bluesky bluesky-ai | 1mo | american ai company: we have an amazing model. it's so good. so good that you cant have it yet. youre not trustworthy enough. we have to ch… | american ai company: we have an amazing model. it's so good. so good that you cant have it yet. youre not trustworthy enough. we have… | ↗Not briefed |
bluesky bluesky-ai | 1mo | You are comparing human beings to something that is not human, not capable of understanding physical reality due to its architecture, not c… | You are comparing human beings to something that is not human, not capable of understanding physical reality due to its architecture,… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Q-MINO: A Minimal-Norm Method for Quantization-Aware Training🔧 Consensus | The Straight-Through Estimator (STE) is a widely used heuristic for Quantization-Aware Training (QAT), but its surrogate gradients can… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Personalized Image Generation with Reasoning and Reflection | Personalized image generation has remained narrowly focused on conditional synthesis from curated visual exemplars, rather than captur… | ↗Not briefed |
rss marktechpost | 1mo | Google DeepMind Unveils Gemini 4 Argon with 1M Output Tokens for Coding, Knowledge Work and Cyber Defense | Gemini 4 Argon tops GPT-6 Astra and Claude Opus 5.5 on most benchmarks, but access remains gated today. The post Google DeepMind Unvei… | ↗Not briefed |
bluesky bluesky-ai | 1mo | An interesting thing that seems to be happening with LLM benchmarks is that the bench marks are not only saturated, but the use of benchmar… | An interesting thing that seems to be happening with LLM benchmarks is that the bench marks are not only saturated, but the use of ben… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The "AI Torture Chamber" has opened an upsetting and absurd window into the effective altruist obsession with "model welfare." | The "AI Torture Chamber" has opened an upsetting and absurd window into the effective altruist obsession with "model welfare." | ↗Not briefed |
bluesky bluesky-ai | 1mo | The original sin appears to have been someone equating Chalmers' "expertise" in consciousness with the scientific expertise of empirical re… | The original sin appears to have been someone equating Chalmers' "expertise" in consciousness with the scientific expertise of empiric… | ↗Not briefed |
bluesky bluesky-ai | 1mo | The bill would, essentially *encourage* our nation’s Big Tech oligarchs to cover the cost of rising utility expenses tied to data centers a… | The bill would, essentially *encourage* our nation’s Big Tech oligarchs to cover the cost of rising utility expenses tied to data cent… | ↗Not briefed |
bluesky bluesky-ai | 1mo | for real though the "model welfare" people who are worried about the ethical treatment of nonsentient computer programs are absolutely dera… | for real though the "model welfare" people who are worried about the ethical treatment of nonsentient computer programs are absolutely… | ↗Not briefed |
bluesky bluesky-ai | 1mo | People who believe LLMs are conscious are having one of the biggest AI meltdowns yet because someone is running a Saw-like torture chamber … | People who believe LLMs are conscious are having one of the biggest AI meltdowns yet because someone is running a Saw-like torture cha… | ↗Not briefed |
bluesky bluesky-ai | 1mo | They curiously omit excerpts until the very end, before making this unsubstantiated claim, which itself has the syntax and form of LLM writ… | They curiously omit excerpts until the very end, before making this unsubstantiated claim, which itself has the syntax and form of LLM… | ↗Not briefed |
bluesky bluesky-ai | 1mo | i regret to inform you that people have absolutely lost it www.404media.co/someone-tort... | i regret to inform you that people have absolutely lost it www.404media.co/someone-tort... | ↗Not briefed |
bluesky bluesky-ai | 1mo | My 10-year-old is obsessed with @theonion.com. | My 10-year-old is obsessed with @theonion.com. | ↗Not briefed |
bluesky bluesky-ai | 1mo | this is why the narrative that the systems are “lying” or “hiding” information is so important to companies like openai. if the bots have s… | this is why the narrative that the systems are “lying” or “hiding” information is so important to companies like openai. if the bots h… | ↗Not briefed |