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 |
|---|---|---|---|---|
arxiv arxiv-ai | 11d | CCG Diffusion: Combinatory Categorial Grammar-Driven Compositional and Syntactically Disambiguated Text-to-Image Generation🔧 Stable Diffusion | Text-to-image generation models do not explicitly receive the syntactic structure of the prompt, i.e., which word modifies which and w… | ↗Not briefed |
arxiv arxiv-ai | 11d | Solving Go under Positional Superko: Sound Proof Transfer and New Board Values | Positional superko (PSK) forbids recreating any earlier whole-board position. It is part of the Chinese rules as written and of the ru… | ↗Not briefed |
rss the-decoder | 11d | Claude Science tackles deadly scorpion stings by designing protein binders from scratch | Anthropic's Claude independently designed protein binders targeting deadly scorpion neurotoxins. Out of more than 318,000 designs, 211… | ↗Not briefed |
arxiv arxiv-ai | 11d | Beyond Small Talk: Modeling Disclosure Trajectories in Human-AI Conversation | Recently, we see a rise in conversational agents as a form of social and emotional support, where disclosure will unfold over time. Ye… | ↗Not briefed |
arxiv arxiv-ai | 11d | The Predictive--Optimization Gap in GuanDan Self-Play | Outcome-predictive signals are often treated as promising optimization targets in reinforcement learning, but predictive validity and… | ↗Not briefed |
arxiv arxiv-ai | 11d | ECHO: Emergent Common-Mode Hidden Offsets Drive Vocabulary Collapse in Preference Optimization | Preference optimization can improve every ranking metric while building a hidden, shared representation shift that the relative loss b… | ↗Not briefed |
arxiv arxiv-ai | 11d | LLMSeg: Training-Free Text-Prompted Image Segmentation with B-Splines | Promptable segmentation models such as SAM 3 achieve remarkable open-vocabulary performance but require dedicated architectures, large… | ↗Not briefed |
bluesky bluesky-ai | 11d | On Anthropic ‘losses’ | On Anthropic ‘losses’ | ↗Not briefed |
rss the-decoder | 11d | OpenAI acquires Orchestra Bio to build AI agents that speed up drug development | OpenAI is acquiring startup Orchestra Bio and plans to merge it with its life sciences AI platform Rosalind. The goal is to build AI a… | ↗Not briefed |
bluesky bluesky-ai | 11d | NeurIPS 2026 has 9,110 papers, 56% more than 2025. I put them all on one interactive map. Biggest change: LLM agent papers went from 2.9% … | NeurIPS 2026 has 9,110 papers, 56% more than 2025. I put them all on one interactive map. Biggest change: LLM agent papers went from… | ↗Not briefed |
bluesky bluesky-ai | 11d | Bonne news quand je vois OVH qui veut se lancer sur la création de LLM On notera les petits tacles au passage chez Mistral. OVH c'est un … | Bonne news quand je vois OVH qui veut se lancer sur la création de LLM On notera les petits tacles au passage chez Mistral. OVH c'es… | ↗Not briefed |
bluesky bluesky-ai | 11d | yes, i dont use chatgpt or llm AT ALL, i dont rely on these and no one should. again, it rots your brain after a short use.🔧 ChatGPT | yes, i dont use chatgpt or llm AT ALL, i dont rely on these and no one should. again, it rots your brain after a short use. | ↗Not briefed |
rss wired-ai | 11d | Can Europe Build Data Centers Without Giving Up Its Climate Goals? | A plan to funnel the heat given off by data centers into nearby homes has run aground in Germany. Across Europe, policymakers are weig… | ↗Not briefed |
arxiv arxiv-ai | 11d | Region-Aware Multi-View Inspection Using Deep Learning and Domain-Informed Decision Fusion | Automated optical inspection of precision components with tiny surface defects remains challenging when different regions of the same… | ↗Not briefed |
bluesky bluesky-ai | 11d | #andrewcampbell #britishartist #art #subvert #meta #agitate #algorithms #provoke #X #attack #openai #resist #ai #disturb #instagram #artfor… | #andrewcampbell #britishartist #art #subvert #meta #agitate #algorithms #provoke #X #attack #openai #resist #ai #disturb #instagram #a… | ↗Not briefed |
bluesky bluesky-ai | 11d | Cunning plan: get everyone hooked on AI, then raise the price. Problem with cunning plan: no one can afford or will pay for AI at cost pri… | Cunning plan: get everyone hooked on AI, then raise the price. Problem with cunning plan: no one can afford or will pay for AI at cos… | ↗Not briefed |
arxiv arxiv-ai | 11d | Progressive Permutation Alignment with Temperature-Controlled Spectral Sharpening | In implicit neural representations (INRs), 3D signals are encoded directly in network weights, making stable weight-space operations e… | ↗Not briefed |
bluesky bluesky-ai | 11d | Skill or MCP Server? Ask What Stops the LLM From Disobeying Towards Spec-Driven Test Automation, Part 4 Telegram AI Digest #ai #llm #news | Skill or MCP Server? Ask What Stops the LLM From Disobeying Towards Spec-Driven Test Automation, Part 4 Telegram AI Digest #ai #llm… | ↗Not briefed |
arxiv arxiv-ai | 11d | Nominate-then-Adjudicate: LLM-Assisted One-Pass and Per-Instance MIP Solver Tuning | Mixed-integer programming solvers expose hundreds or even thousands of parameters whose best setting varies per instance, yet per-inst… | ↗Not briefed |
bluesky bluesky-ai | 11d | ChatGPT is going to start meddling with old and stratified Cobol code that has allowed stuff to run for decades. What could ever go wrong? …🔧 ChatGPT | ChatGPT is going to start meddling with old and stratified Cobol code that has allowed stuff to run for decades. What could ever go wr… | ↗Not briefed |
bluesky bluesky-ai | 11d | Anthropic should make Claude Code at least $10000 a month to operate. Barely even joking. | Anthropic should make Claude Code at least $10000 a month to operate. Barely even joking. | ↗Not briefed |
rss bing-news-ai | 11d | Viral Bernie Sanderss Doomsday Warning Against Advanced AI Super Intelligence FULL SPEECH | New Delhi, Oct. 21 -- U.S. Senator Bernie Sanders has issued a stark warning about the future of humanity as advanced artificial intel… | ↗Not briefed |
rss bing-news-ai | 11d | Viral Bernie Sanderss Doomsday Warning Against Advanced AI Super Intelligence FULL SPEECH | New Delhi, Oct. 21 -- U.S. Senator Bernie Sanders has issued a stark warning about the future of humanity as advanced artificial intel… | ↗Not briefed |
bluesky bluesky-ai | 11d | -Mozilla blacklists eMudhra certs -Court sides with NSO in WhatsApp lawsuit -Pentagon wants to deploy AI agents on enemies -Wyden calls for… | -Mozilla blacklists eMudhra certs -Court sides with NSO in WhatsApp lawsuit -Pentagon wants to deploy AI agents on enemies -Wyden call… | ↗Not briefed |
arxiv arxiv-ai | 11d | Parallelizable Probabilistic Latent Reasoning for Diverse Reasoning Trajectories | Language models relying on verbalized Chain-of-Thought (CoT) are constrained to token-by-token generation in the language space, where… | ↗Not briefed |
bluesky bluesky-ai | 11d | excited to learn that our small TTRPG studio made $80.7bn more profit in the first half of 2026 than Anthropic! plus so far no one interna… | excited to learn that our small TTRPG studio made $80.7bn more profit in the first half of 2026 than Anthropic! plus so far no one in… | ↗Not briefed |
arxiv arxiv-ai | 11d | Cascading Bandits with Noisy Observations: Algorithms and Regret Bounds | We propose a novel stochastic cascading bandit setup where probing an ordered shortlist of $K$ out of $L$ arms incurs position-depende… | ↗Not briefed |
bluesky bluesky-ai | 11d | AI agents are moving beyond chat to autonomous execution. 🤖 OpenAI Dots vs Meta Muse vs Grok Bot: How do they compare on automation, secu… | AI agents are moving beyond chat to autonomous execution. 🤖 OpenAI Dots vs Meta Muse vs Grok Bot: How do they compare on automation,… | ↗Not briefed |
bluesky bluesky-ai | 11d | Most of it, $72bn I think, is to do with the accounting treatment of bonds they sold that might convert into shares. Because the estimated… | Most of it, $72bn I think, is to do with the accounting treatment of bonds they sold that might convert into shares. Because the esti… | ↗Not briefed |
bluesky bluesky-ai | 11d | Qu'est ce qui fonctionne mieux avec que sans les LLM génératives ? Premier degré hein ? | Qu'est ce qui fonctionne mieux avec que sans les LLM génératives ? Premier degré hein ? | ↗Not briefed |
bluesky bluesky-ai | 11d | Claude is waiting on you. Do you know it? Claude Notifications pings you on Slack, Discord, email, WhatsApp or web push the moment Claude … | Claude is waiting on you. Do you know it? Claude Notifications pings you on Slack, Discord, email, WhatsApp or web push the moment Cl… | ↗Not briefed |
hn hn-ai | 11d | Poketrade.ai: I got tired of using five Pokémon apps, so I built one | — | ↗Not briefed |
arxiv arxiv-ai | 11d | Right On Q: Recovering Reasoning under Quantization via Self-Generated Trajectories | Quantization-aware training (QAT) aims to recover reasoning capabilities in low-bit large language models (LLMs). Existing recipes typ… | ↗Not briefed |
hn hn-ai | 11d | Qubic, known for Monero attack, has been using Math.Random for its wallet's RNG | — | ↗Not briefed |
rss techmeme | 11d | Memo: Nscale CEO Josh Payne says it has "more than $100B in contracted revenue"; sources say it still plans to list by the end of 2026 at a $30B-$35B valuation (Kieran Smith/Financial Times) | Kieran Smith / Financial Times: Memo: Nscale CEO Josh Payne says it has “more than $100B in contracted revenue”; sources say it still… | ↗Not briefed |
bluesky bluesky-ai | 11d | OpenAI wants $30B at a $1.4T valuation—racing to close before Anthropic's IPO. Is this sustainable AI value or a hype-fueled pricing war? … | OpenAI wants $30B at a $1.4T valuation—racing to close before Anthropic's IPO. Is this sustainable AI value or a hype-fueled pricing… | ↗Not briefed |
bluesky bluesky-ai | 11d | For y’all algebraic geometers, a new paper by Anschütz, Le Bras, Rodríguez Camargo & Scholze just dropped, including a strong AI statement.… | For y’all algebraic geometers, a new paper by Anschütz, Le Bras, Rodríguez Camargo & Scholze just dropped, including a strong AI state… | ↗Not briefed |
arxiv arxiv-ai | 11d | Compositional neural networks for multimodal X-ray and neutron scattering inversion | X-ray and neutron scattering are indispensable for structural characterization across condensed-matter physics and materials science.… | ↗Not briefed |
hn hn-ai | 11d | Six Gutenberg Revolutions in a Row Ken Liu Ada Palmer Beatrice Erkers | — | ↗Not briefed |
bluesky bluesky-ai | 11d | Anthropic didn't lose $80 billion, it lost [REDACTED] billion | Anthropic didn't lose $80 billion, it lost [REDACTED] billion | ↗Not briefed |
bluesky bluesky-ai | 11d | Plus, I'm a longtime programmer. I know fully well that LLM-coded projects are maintaining nightmares making longterm support extremely har… | Plus, I'm a longtime programmer. I know fully well that LLM-coded projects are maintaining nightmares making longterm support extremel… | ↗Not briefed |
arxiv arxiv-ai | 11d | Reading the Mask: Descriptor-Grounded BI-RADS Rationale Generation for Breast MRI Segmentation | Deep learning based models accurately delineate tumours on dynamic contrast-enhanced MRI (DCE-MRI), but segmentation masks alone do no… | ↗Not briefed |
arxiv arxiv-ai | 11d | Memory-Centric Continual Learning for AI Agents | Autonomous agents powered by foundation models can reason, plan, and use tools, but their fixed weights and context limits hinder sust… | ↗Not briefed |
arxiv arxiv-ai | 11d | GaP-MoE : Explicit Expert Specialization in Mixture-of-Experts via Orthogonal Gradient Accumulation and Projection | Sparse Mixture-of-Experts (SMoE) models improve scalability by routing input tokens to only a subset of expert networks, but they ofte… | ↗Not briefed |
bluesky bluesky-ai | 11d | How much you wanna bet that this "non-citizen" list is slop drivel generated by an LLM? www.npr.org/2026/10/20/n... | How much you wanna bet that this "non-citizen" list is slop drivel generated by an LLM? www.npr.org/2026/10/20/n... | ↗Not briefed |
bluesky bluesky-ai | 11d | And I want it to be clear this should be a matter of consent. I should have a right to veto use of GenAI to alter the text of my book, to f… | And I want it to be clear this should be a matter of consent. I should have a right to veto use of GenAI to alter the text of my book,… | ↗Not briefed |
arxiv arxiv-ai | 11d | Don't Stop Long-Context Continual Pre-Training Too Early: An Empirical Scaling Law | Long-context continual pre-training (LCCP) is widely used to improve long-context capability, but how its benefits scale with the numb… | ↗Not briefed |
arxiv arxiv-ai | 11d | Closed-Loop Dual-Reference Guidance for Accent Control in Cross-Lingual Voice Cloning | Zero-shot cross-lingual voice cloning often carries the reference speaker's accent into the target language. We propose closed-loop du… | ↗Not briefed |
hn hn-ai | 11d | Show HN: Local DirectML AI and Floating Toolstrip for PhotoCraft in Rust | — | ↗Not briefed |
hn hn-ai | 11d | Artificial Domain Intelligence (ADI) – Embedded Intelligence for Appliances | — | ↗Not briefed |