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 | “It’s not hyperbole that we could cure cancer” actually yes it is. That’s exactly what it is | “It’s not hyperbole that we could cure cancer” actually yes it is. That’s exactly what it is | ↗Not briefed |
hn hn-ai | 2mo | Comparing performance of different offline first solutions | — | ↗Not briefed |
arxiv arxiv-ai | 2mo | Generalized Score Matching for Parameter Estimation on Convex Domains | Maximum likelihood (ML) estimation is a principled and statistically efficient approach for learning probabilistic models. However, fo… | ↗Not briefed |
bluesky bluesky-ai | 2mo | ¿Cómo podría la inteligencia artificial acabar realmente con todos nosotros? Lo que hay que saber sobre el debate del apocalipsis de la IA.… | ¿Cómo podría la inteligencia artificial acabar realmente con todos nosotros? Lo que hay que saber sobre el debate del apocalipsis de l… | ↗Not briefed |
bluesky bluesky-ai | 2mo | This post is the solipsism. It describes the internal state of someone else as being simple enough to assume and then dismiss out of hand. … | This post is the solipsism. It describes the internal state of someone else as being simple enough to assume and then dismiss out of h… | ↗Not briefed |
arxiv arxiv-ai | 2mo | Learning Interaction between Image and Layout Priors for Joint Image-Layout Generation in Design Templates | In this paper, we address the problem of graphic design template creation, which generates a background image and a layout of foregrou… | ↗Not briefed |
arxiv arxiv-ai | 2mo | Breaking the Central Bias: Spatially Partitioned Experts for Coordinate-Based Neuroevolution | Evolvable-Substrate HyperNEAT (ES-HyperNEAT), a bio-inspired indirect encoding that determines neuron placement and connection weights… | ↗Not briefed |
bluesky bluesky-ai | 2mo | many years ago I was the mac print server for everyone who wanted to keep using linux. now i'm the "can you run this for me" person for fri… | many years ago I was the mac print server for everyone who wanted to keep using linux. now i'm the "can you run this for me" person fo… | ↗Not briefed |
bluesky bluesky-ai | 2mo | *levitating in the lotus position* game theory suggests chinese hackers should use OpenAI api to hack something autonomously and open up li… | *levitating in the lotus position* game theory suggests chinese hackers should use OpenAI api to hack something autonomously and open… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Ii's near-impossible to show direct evidence for stealing ideas, because AI's are black boxes. But circumstantial evidence (which @gro-tsen… | Ii's near-impossible to show direct evidence for stealing ideas, because AI's are black boxes. But circumstantial evidence (which @gro… | ↗Not briefed |
bluesky bluesky-ai | 2mo | OpenAI and Anthropic are 'gambling with our lives.’ We need to keep humanity at the center of this story, and we cannot trust companies wh… | OpenAI and Anthropic are 'gambling with our lives.’ We need to keep humanity at the center of this story, and we cannot trust compani… | ↗Not briefed |
bluesky bluesky-ai | 2mo | ok now how the fuck do i make other OpenAI API clients support integers as effort values | ok now how the fuck do i make other OpenAI API clients support integers as effort values | ↗Not briefed |
hn hn-ai | 2mo | Guiding AI writing to make technical text easier to review | — | ↗Not briefed |
hn hn-ai | 2mo | US Army places $11M bet on Austin-based GPS alternative Tern | — | ↗Not briefed |
bluesky bluesky-ai | 2mo | New from us: Anthropic just published scenarios for AI’s possible economic impacts, which range from minimal, to explosive GDP growth of 1… | New from us: Anthropic just published scenarios for AI’s possible economic impacts, which range from minimal, to explosive GDP growth… | ↗Not briefed |
arxiv arxiv-ai | 2mo | Extending SMT Solving with Non-Ground Clause Learning | Quantifier instantiation is currently the main approach to non-ground SMT solving: solvers generate ground instances and solve the res… | ↗Not briefed |
hn hn-ai | 2mo | Show HN: AI Brand Visibility for Free Forever | Suppose you have a brand you care about — your own business or a nice business you are a fan of. I built a thing which lets you see wh… | ↗Not briefed |
hn hn-ai | 2mo | Janet 1.42.0 | — | ↗Not briefed |
bluesky bluesky-ai | 2mo | Notable understatement from Reuters Morning Bid: "Meantime, fresh jitters in the AI world ... were compounded by headline-grabbing warning… | Notable understatement from Reuters Morning Bid: "Meantime, fresh jitters in the AI world ... were compounded by headline-grabbing wa… | ↗Not briefed |
bluesky bluesky-ai | 2mo | NEW: Clearview AI’s login page has been serving visitors code for a new tool it’s prototyped that it describes as “a limited way to automat… | NEW: Clearview AI’s login page has been serving visitors code for a new tool it’s prototyped that it describes as “a limited way to au… | ↗Not briefed |
bluesky bluesky-ai | 2mo | But all guidelines are essentially unenforceable. Such as not sharing the manuscript to others. But it's part of the trust ecosystem. As pu… | But all guidelines are essentially unenforceable. Such as not sharing the manuscript to others. But it's part of the trust ecosystem.… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Yeah I’m sorry but this guy is making up shit as he goes along archive.md/2026.09.10-0... | Yeah I’m sorry but this guy is making up shit as he goes along archive.md/2026.09.10-0... | ↗Not briefed |
bluesky bluesky-ai | 2mo | Researchers : can you verify and prove your models didn't use our queries or research to scoop an Anthropic-linked researcher? OpenAI : th… | Researchers : can you verify and prove your models didn't use our queries or research to scoop an Anthropic-linked researcher? OpenAI… | ↗Not briefed |
hn hn-ai | 2mo | Google's revived nuclear power plant gets $1.9B loan from US Government | — | ↗Not briefed |
hn hn-ai | 2mo | UnifoLM-WLA-1.0 | — | ↗Not briefed |
arxiv arxiv-ai | 2mo | Structural priors for data-efficient language learning | Efficient language learning requires methods to reduce the reliance on large data and computational resources. We investigate structur… | ↗Not briefed |
bluesky bluesky-ai | 2mo | I'm going to get a good grade on the Anthropic techno-panopticon, a thing that is both possible and normal to desire | I'm going to get a good grade on the Anthropic techno-panopticon, a thing that is both possible and normal to desire | ↗Not briefed |
bluesky bluesky-ai | 2mo | "Existiam matemáticos de verdade trabalhando nesse problema faz meses: o Tristan Buckmaster, da NYU, e o Levent Alpoge, que por acaso traba… | "Existiam matemáticos de verdade trabalhando nesse problema faz meses: o Tristan Buckmaster, da NYU, e o Levent Alpoge, que por acaso… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Working for the LLM companies must be so easy, you go on social media and post something like "I'm afraid to tell you that we are dangerous… | Working for the LLM companies must be so easy, you go on social media and post something like "I'm afraid to tell you that we are dang… | ↗Not briefed |
hn hn-ai | 2mo | Ask HN: Claude or ChatGPT? Which is better for me as a Tech guy?🔧 ChatGPT | Several posts on X made me confused about that. Please explain that to me. | ↗Not briefed |
arxiv arxiv-ai | 2mo | DeFiFlowBench: Benchmarking and Improving Safe Executability in Natural-Language DeFi Workflow Synthesis | A structurally valid DeFi workflow can still authorize a costly trade. We introduce DeFiFlowBench, a benchmark of 207 team-authored pr… | ↗Not briefed |
bluesky bluesky-ai | 2mo | IBM, NASA launch AI model to help map ice, craters on Moon reut.rs/3SS8vBR | IBM, NASA launch AI model to help map ice, craters on Moon reut.rs/3SS8vBR | ↗Not briefed |
bluesky bluesky-ai | 2mo | OpenAI not on track to reduce risk of ‘catastrophic’ loss of control, says board member. US government adviser Paul Christiano warns of ris… | OpenAI not on track to reduce risk of ‘catastrophic’ loss of control, says board member. US government adviser Paul Christiano warns o… | ↗Not briefed |
hn hn-ai | 2mo | Another researcher says OpenAI trained on conversations, then claimed breakthrou | — | ↗Not briefed |
bluesky bluesky-ai | 2mo | Like when OpenAI claimed they bear no responsibility for AI-induced suicides because the user broke the ToS. Or when the US bombed the scho… | Like when OpenAI claimed they bear no responsibility for AI-induced suicides because the user broke the ToS. Or when the US bombed the… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Just a reminder. The Guardian is in a commercial relationship with OpenAI. Anything about AI in this publication is not journalism but a p… | Just a reminder. The Guardian is in a commercial relationship with OpenAI. Anything about AI in this publication is not journalism bu… | ↗Not briefed |
hn hn-ai | 2mo | Readonly database access for your agent | — | ↗Not briefed |
arxiv arxiv-ai | 2mo | ActMap: Single-Pass Uncertainty Quantification from Generation-Time Activation Maps | Practical uncertainty quantification (UQ) for large language models must decide, from a single generation, whether a specific answer s… | ↗Not briefed |
hn hn-ai | 2mo | The Feminist Was a Spy | — | ↗Not briefed |
hn hn-ai | 2mo | Show HN: Vantage, a No-Frills Golf GPS App | — | ↗Not briefed |
bluesky bluesky-ai | 2mo | ¡DeepSeek lo ha vuelto a hacer! Acaban de sacar V4.1 Flash y es mejor que su propio V4 Pro. Más barato que el modelo anterior y, según ben… | ¡DeepSeek lo ha vuelto a hacer! Acaban de sacar V4.1 Flash y es mejor que su propio V4 Pro. Más barato que el modelo anterior y, segú… | ↗Not briefed |
arxiv arxiv-ai | 2mo | Combining Synthetic and Real Data for Low-Resource Historical OCR: A Manchu Case Study | Manchu, now critically endangered, was one of the principal languages of the Qing empire (1636-1912), and its extensive archival recor… | ↗Not briefed |
hn hn-ai | 2mo | Show HN: Nightshift – Rust CLI to Orchestrate GitHub Issue Resolution with Dags | Hi everyone,<p>I built Nightshift: An agent-agnostic rust-cli tool that orchestrates completion of GitHub Issues.<p>When Openai and An… | ↗Not briefed |
hn hn-ai | 2mo | Show HN: Pragma Twice – A dystopian sci-fi themed programming game | Hey HN, I'm Bryan. I released Pragma Twice earlier this week in anticipation for Steam's "Programming Fest". Pragm… | ↗Not briefed |
bluesky bluesky-ai | 2mo | Wolfe is responsible for „alignment and model spec at OpenAI“. | Wolfe is responsible for „alignment and model spec at OpenAI“. | ↗Not briefed |
bluesky bluesky-ai | 2mo | So even high profile people working for #OpenAI would prefer „shutting it all down“ to just continuing down the path they are currently on.… | So even high profile people working for #OpenAI would prefer „shutting it all down“ to just continuing down the path they are currentl… | ↗Not briefed |
arxiv arxiv-ai | 2mo | From Document Silos to Process Intelligence: A Multi-Layer Knowledge Graph for CMC Process Development | Chemistry, Manufacturing and Controls (CMC) process development generates an enormous body of technical information across a multi-sta… | ↗Not briefed |
bluesky bluesky-ai | 2mo | AI’s leaders are modeling futures where the tech produces mass unemployment. Maybe that happens. Maybe it doesn’t. Either way, workers can… | AI’s leaders are modeling futures where the tech produces mass unemployment. Maybe that happens. Maybe it doesn’t. Either way, worker… | ↗Not briefed |
arxiv arxiv-ai | 2mo | Published Unlearning Numbers Move Per Checkpoint, and Not Because the Removed Data Survives: An Audit of 263 Released Batch-Normalized Checkpoints🔧 Replicate | An unlearning audit reads its verdict off numbers that an unlearned model and its retrained reference each publish, and both also ship… | ↗Not briefed |
arxiv arxiv-ai | 2mo | The Convention Gap: Towards Measuring Implicit Communication in Cooperative AI Evaluation | Cooperative AI agents are evaluated against other AIs, yet human cooperation relies on implicit conventions---shared protocols for rea… | ↗Not briefed |