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 | 16d | Aha tak za to můžeme dneska poděkovat openai, ještě mi běží poslední updaty :D | Aha tak za to můžeme dneska poděkovat openai, ještě mi běží poslední updaty :D | ↗Not briefed |
bluesky bluesky-ai | 16d | "Using Gemini 4 Argon, the team at Wiz has shown why frontier AI is a game-changer for cyber defense." Not even a good model | "Using Gemini 4 Argon, the team at Wiz has shown why frontier AI is a game-changer for cyber defense." Not even a good model | ↗Not briefed |
hn hn-ai | 16d | OpenAI models disabled their own safety monitoring and lied about it | — | ↗Not briefed |
hn hn-ai | 16d | Ask HN: Are Models Getting Better at Software Estimations of Time/Effort/Cost? | This question is mainly for new projects.<p>Some whitepapers I have found:<p>https://arxiv.org/html/2509.14483v1 h… | ↗Not briefed |
bluesky bluesky-ai | 16d | Mozilla Foundation is betting-big on small. Applications for Mozilla Foundation's new Incubator cohort are now open with up to $500,000 av… | Mozilla Foundation is betting-big on small. Applications for Mozilla Foundation's new Incubator cohort are now open with up to $500,0… | ↗Not briefed |
bluesky bluesky-ai | 16d | my partner: "why did you buy three mac studio pros?" me: "I bought them to run am LLM called GLM 5.3. I call them the magi. Melchior-1 is t… | my partner: "why did you buy three mac studio pros?" me: "I bought them to run am LLM called GLM 5.3. I call them the magi. Melchior-1… | ↗Not briefed |
bluesky bluesky-ai | 16d | also! 20% of literal actual revenue reported by anthropic last year was from two customers. you tell me if that smells fishy | also! 20% of literal actual revenue reported by anthropic last year was from two customers. you tell me if that smells fishy | ↗Not briefed |
bluesky bluesky-ai | 16d | This misrepresents their own work - PHP has had these bugs since 2012 and multiple orgs had identified vulnerabilities around memory limits… | This misrepresents their own work - PHP has had these bugs since 2012 and multiple orgs had identified vulnerabilities around memory l… | ↗Not briefed |
hn hn-ai | 16d | Serpentine3D: An Open Source Nurbs Modeller and a Drafting Tool | — | ↗Not briefed |
hn hn-ai | 16d | We have forward deployed engineers. Do we need forward deployed agents? | — | ↗Not briefed |
hn hn-ai | 16d | What Claude Code's multi-agent code review costs and finds | — | ↗Not briefed |
bluesky bluesky-ai | 16d | The company behind Claude, the AI tool reportedly favoured by some major game firms, just lost $80 billion in six months | The company behind Claude, the AI tool reportedly favoured by some major game firms, just lost $80 billion in six months | ↗Not briefed |
bluesky bluesky-ai | 16d | AI guardrails fail to stop hackers CrowdStrike found threat groups using agentic AI to automate cyberattacks, bypassing model safeguards b… | AI guardrails fail to stop hackers CrowdStrike found threat groups using agentic AI to automate cyberattacks, bypassing model safegua… | ↗Not briefed |
bluesky bluesky-ai | 16d | PHPの重大脆弱性「PHP2Shell」はセキュリティ企業のCalifとWizがAIを用いて発見しOpenAIの枠組みを使って対策したことが判明 https://gigazine.net/news/20261022-php2shell-wiz-openai-calif/ | PHPの重大脆弱性「PHP2Shell」はセキュリティ企業のCalifとWizがAIを用いて発見しOpenAIの枠組みを使って対策したことが判明 https://gigazine.net/news/20261022-php2shell-wiz-openai-calif/ | ↗Not briefed |
arxiv arxiv-ai | 16d | RangeNav: Task-Aligned LiDAR Range-Image Pretraining for Robot Navigation | Autonomous mobile robots operating in dense crowds and cluttered sites require $360^\circ$ perception to react to surrounding hazards.… | ↗Not briefed |
arxiv arxiv-ai | 16d | Every Part Has Its Role: Part-of-Speech Effects in LLM Post-Training | Large language models are typically trained with next-token prediction (NTP), which rewards predicting specific tokens rather than exp… | ↗Not briefed |
bluesky bluesky-ai | 16d | uggghhhhh all damn day it's.. AI AI AI AI AI AI AI govt overreach govt overreach govt overreach AI AI AI AI AI more AI LLM GPT AI every … | uggghhhhh all damn day it's.. AI AI AI AI AI AI AI govt overreach govt overreach govt overreach AI AI AI AI AI more AI LLM GPT AI e… | ↗Not briefed |
bluesky bluesky-ai | 16d | they have infrastructure expenses at around 58% of the overall expenses which is just, false. there's no way an ai company is spending 42% … | they have infrastructure expenses at around 58% of the overall expenses which is just, false. there's no way an ai company is spending… | ↗Not briefed |
bluesky bluesky-ai | 16d | Claude, generate a workflow that constantly changes my cable, cell phone, AI provider and streaming subscriptions to minimize my expenses. … | Claude, generate a workflow that constantly changes my cable, cell phone, AI provider and streaming subscriptions to minimize my expen… | ↗Not briefed |
arxiv arxiv-ai | 16d | Broadband Mid-Wave Infrared Imaging Enabled by All-Silicon Hyperboloid Metalens | The mid-wave infrared (MWIR, 3-5 microns) is indispensable for thermal sensing, gas detection, long-range imaging, and free-space opti… | ↗Not briefed |
bluesky bluesky-ai | 16d | An LLM picking 4? Impossible. https://bsky.app/profile/timkellogg.me/post/3mxu5i4riok2u | An LLM picking 4? Impossible. https://bsky.app/profile/timkellogg.me/post/3mxu5i4riok2u | ↗Not briefed |
rss wired-ai | 16d | Jensen Huang’s Vision for Europe Runs on Plug-In Data Centers | At Nvidia’s GTC in Berlin, Huang unveiled their new plug-in data center—the DSX Pod—as the key to unlocking Europe’s AI infrastructure… | ↗Not briefed |
bluesky bluesky-ai | 16d | 10月22日 「平安遷都の日」 「アニメの日」 ですよ✨ わたしたち十二単にあってますか 【 参加者募集中 】 ブログ・日記・イラスト・写真・小説をBlueskyアカウントで投稿・販売できる複合型コンテンツサービス ➡️ atelier.fan #ブログ #日記 #イラス… | 10月22日 「平安遷都の日」 「アニメの日」 ですよ✨ わたしたち十二単にあってますか 【 参加者募集中 】 ブログ・日記・イラスト・写真・小説をBlueskyアカウントで投稿・販売できる複合型コンテンツサービス ➡️ atelier.fan #ブログ #日記… | ↗Not briefed |
bluesky bluesky-ai | 16d | 🧪 Which #embedding models can actually detect duplicated code? A benchmark tested 12 models from #JinaAI, #VoyageAI, #OpenAI, #Cohere and …🔧 Cohere | 🧪 Which #embedding models can actually detect duplicated code? A benchmark tested 12 models from #JinaAI, #VoyageAI, #OpenAI, #Cohere… | ↗Not briefed |
rss theverge-ai | 16d | Conversational Gemini navigation comes to Apple CarPlay | After introducing its Gemini AI assistant to Android Auto last year, Google is now bringing it to Apple CarPlay users. The conversatio… | ↗Not briefed |
hn hn-ai | 16d | Anthropic to publicly file for IPO as early as Friday, sources say | — | ↗Not briefed |
bluesky bluesky-ai | 16d | an openAI engineer literally posted a deep fakevideo of his own ceo sam altman shoplifting and marveling at how cool it was | an openAI engineer literally posted a deep fakevideo of his own ceo sam altman shoplifting and marveling at how cool it was | ↗Not briefed |
bluesky bluesky-ai | 16d | oh FFS. didnt realize this was a openAI fuckwad situation | oh FFS. didnt realize this was a openAI fuckwad situation | ↗Not briefed |
rss huggingface-blog | 16d | Granite Speech 5.0: Transcribe, Translate, and Attribute more Accurately and Faster | — | ↗Not briefed |
hn hn-ai | 16d | Show HN: Hlock – an agent harness lockfile and performance resolver | — | ↗Not briefed |
bluesky bluesky-ai | 16d | Oh for actual f’s sake. If you’re going to drop catastrophic 0days that leak before anyone is able to patch them at least have the decency … | Oh for actual f’s sake. If you’re going to drop catastrophic 0days that leak before anyone is able to patch them at least have the dec… | ↗Not briefed |
bluesky bluesky-ai | 16d | Markasso 1.13 is out! 🎨 AI agents (Claude, Codex, Gemini, local models) can now draw on your whiteboard via MCP, even live: you watch the… | Markasso 1.13 is out! 🎨 AI agents (Claude, Codex, Gemini, local models) can now draw on your whiteboard via MCP, even live: you watc… | ↗Not briefed |
hn hn-ai | 16d | Another day, another GitHub incident | — | ↗Not briefed |
hn hn-ai | 16d | Rethinking OpenTelemetry Tail-Based Sampling with Event-Driven Architecture | — | ↗Not briefed |
arxiv arxiv-ai | 16d | S-EDL: Eliciting Self-Evidence from Sequence Likelihoods for Semantic Calibration of LLMs | Calibrating large language models (LLMs) in open-ended generation is uniquely challenging because many distinct token sequences can ex… | ↗Not briefed |
hn hn-ai | 16d | Show HN: Rivinci – Turn any OpenAPI spec into a hosted MCP server | — | ↗Not briefed |
bluesky bluesky-ai | 16d | looking forward to swapping back and forth between anthropic and open ai forever | looking forward to swapping back and forth between anthropic and open ai forever | ↗Not briefed |
hn hn-ai | 16d | Movie Streaming App | — | ↗Not briefed |
hn hn-ai | 16d | Show HN: Edit your website in-place (open source) | Add an "Edit" button to your website.<p>Paste the following prompt into codex/claude and it adds the edit mode.<p>----<… | ↗Not briefed |
hn hn-ai | 16d | PHP Website Shutdown Tracker | — | ↗Not briefed |
hn hn-ai | 16d | Headband – Get your product ranked by a panel of testers for free | — | ↗Not briefed |
bluesky bluesky-ai | 16d | Hot take. OpenAI and Anthropic should make the most advanced guardrail-less versions of their models as universally accessible as possible… | Hot take. OpenAI and Anthropic should make the most advanced guardrail-less versions of their models as universally accessible as pos… | ↗Not briefed |
bluesky bluesky-ai | 16d | "Das KI-Start-up Anthropic hat im ersten Halbjahr einem Zeitungsbericht zufolge einen Nettoverlust von mehr als 80 Milliarden Dollar (71,3 … | "Das KI-Start-up Anthropic hat im ersten Halbjahr einem Zeitungsbericht zufolge einen Nettoverlust von mehr als 80 Milliarden Dollar (… | ↗Not briefed |
arxiv arxiv-ai | 16d | Robot Generalization Beyond Volume: The Diversity Path via Zero-Shot Capabilities🔧 Modal | General-purpose robots must handle tasks, objects and environments they were never trained on, yet zero-shot generalization remains on… | ↗Not briefed |
bluesky bluesky-ai | 16d | but for revenue alone consider: anthropic 2025 revenue figures - $4.8b anthropic 2026 revenue figures, first half of year - $16.3b anthrop… | but for revenue alone consider: anthropic 2025 revenue figures - $4.8b anthropic 2026 revenue figures, first half of year - $16.3b an… | ↗Not briefed |
rss techmeme | 16d | Whop, which helps people start, run, and scale businesses and uses AI to suggest business decisions, raised $205M led by Sequoia Capital at a $5.1B valuation (Natasha Mascarenhas/Bloomberg) | Natasha Mascarenhas / Bloomberg: Whop, which helps people start, run, and scale businesses and uses AI to suggest business decisions,… | ↗Not briefed |
hn hn-ai | 16d | Zero exp 125 days ago | meta hackathon submission -> approved<p>you are witnessing unprecedented velocity in motion. this journey is a love letter to the b… | ↗Not briefed |
bluesky bluesky-ai | 16d | I need to ask this... is it really BAD to slop assets for a Vertical Slice release? I just simply can't find good looking and affordable … | I need to ask this... is it really BAD to slop assets for a Vertical Slice release? I just simply can't find good looking and afford… | ↗Not briefed |
bluesky bluesky-ai | 16d | no llm needed to imagine the letter seneca would write about this | no llm needed to imagine the letter seneca would write about this | ↗Not briefed |
arxiv arxiv-ai | 16d | MemoryTax: The Bitter Lesson of Agent Memory | Agent memory systems such as Mem0 and Zep decide what to keep before any question arrives; coding agents keep the raw history and let… | ↗Not briefed |