AINews Portal

News

50 items (filtered)

The raw AI news stream — everything we ingest across RSS, Hacker News, arXiv, GitHub, Bluesky, YouTube and more, newest first.

SourceAgeTitleSummaryStatus
bluesky
bluesky-ai
2mo⚠️ These guys all just agreed on a need to slow the pace of AI development: Anthropic CEO Dario Amodei OpenAI CEO Sam Altman xAI CEO Elo…⚠️ These guys all just agreed on a need to slow the pace of AI development: Anthropic CEO Dario Amodei OpenAI CEO Sam Altman xAI CE…
↗Not briefed
bluesky
bluesky-ai
2mowe can't unplug the llm because it might hack the computer into sprouting horse legs and sprint out the doorwe can't unplug the llm because it might hack the computer into sprouting horse legs and sprint out the door
↗Not briefed
bluesky
bluesky-ai
2moThe AI, (surprise, surprise) had subtly altered the definitions of the problem. The challenge is that some of these mathematical problems t…The AI, (surprise, surprise) had subtly altered the definitions of the problem. The challenge is that some of these mathematical probl…
↗Not briefed
hn
hn-ai
2moGPT-6-Astra Can Do Ambitious Things—
↗Not briefed
bluesky
bluesky-ai
2moThis is a great take. Anthropic is now, in seeking “embedded evaluators,” exactly what banks in the 2000s were in asking for oversight th…This is a great take. Anthropic is now, in seeking “embedded evaluators,” exactly what banks in the 2000s were in asking for oversig…
↗Not briefed
bluesky
bluesky-ai
2moGranted, RAND assesses there is a DEVELOPING capability of “LLM/genAI” re: bioweapons (& overstates the skills/capability/motive changes IM…Granted, RAND assesses there is a DEVELOPING capability of “LLM/genAI” re: bioweapons (& overstates the skills/capability/motive chang…
↗Not briefed
hn
hn-ai
2moShow HN: Personal BrewlogBuilt a quick brewlog for personal usage, a memoir of all the beans I've tried and recipes I've messed with. Sort of like a…
↗Not briefed
bluesky
bluesky-ai
2moDario Amodei, who co-founded Anthropic, the company behind the Claude chatbot, called for a global slowdown of artificial intelligence deve…Dario Amodei, who co-founded Anthropic, the company behind the Claude chatbot, called for a global slowdown of artificial intelligence…
↗Not briefed
arxiv
arxiv-ai
2moGraMRAG: Orchestrating Multi-Agent Multi-Step Reasoning via Graph Memory with Reinforcement Learning🔧 ModalAlthough existing multi-agent Retrieval-Augmented Generation (RAG) systems have demonstrated promise on complex multimodal reasoning t…
↗Not briefed
hn
hn-ai
2moShow HN: Reverse-engineered a $800/mo indexing plan into Terraform and Flux🔧 FLUX—
↗Not briefed
bluesky
bluesky-ai
2mo"An AI has broken containment!!!!!" 10 Hack that computer 20 Goto 10 Courtesy of @pluralistic https://pluralistic.net/2026/09/12/god-in-…"An AI has broken containment!!!!!" 10 Hack that computer 20 Goto 10 Courtesy of @pluralistic https://pluralistic.net/2026/09/12/go…
↗Not briefed
arxiv
arxiv-ai
2moStabilizing Performative Feedback Loops with Minimal Model DeploymentsWhen algorithmic predictions inform people's decisions, the models we deploy are performative and actively shape the data we see. This…
↗Not briefed
bluesky
bluesky-ai
2mowww.tomshardware.com/tech-industr...www.tomshardware.com/tech-industr...
↗Not briefed
bluesky
bluesky-ai
2moDo openAI and Anthropic employees not understand how computers work.Do openAI and Anthropic employees not understand how computers work.
↗Not briefed
bluesky
bluesky-ai
2moAs president I will nationalize OpenAI and Anthropic, dismantle data centers, and give every US resident one pre-owned RTX 5090As president I will nationalize OpenAI and Anthropic, dismantle data centers, and give every US resident one pre-owned RTX 5090
↗Not briefed
rss
techmeme
2moElon Musk backs Dario Amodei's arguments about pacing the frontier, saying "Dario is right" (Ben Johansen/Politico)Ben Johansen / Politico: Elon Musk backs Dario Amodei's arguments about pacing the frontier, saying “Dario is right”  —  Anthropic CEO…
↗Not briefed
bluesky
bluesky-ai
2moAnother really fun part is that apparently LLM/AI has substantially enhanced semi-autonomous & adaptive cyber operations, fraud, hacking, &…Another really fun part is that apparently LLM/AI has substantially enhanced semi-autonomous & adaptive cyber operations, fraud, hacki…
↗Not briefed
arxiv
arxiv-ai
2moAGENTQ: Quantization-Conditioned Backdoor Attacks on LLM AgentsQuantization is one of the default deployment paths for open-weight LLM agents, but it is not behavior-preserving: an adversary can re…
↗Not briefed
bluesky
bluesky-ai
2moShortly afterwards, Altman of OpenAI agreed. And shortly after that, Elon Musk also agreed. Imma just say it..... something terrible h…Shortly afterwards, Altman of OpenAI agreed. And shortly after that, Elon Musk also agreed. Imma just say it..... something terri…
↗Not briefed
bluesky
bluesky-ai
2moSo OpenAI, Anthropic, and XAI are all saying that they want the government to slow down AI development. What’s the angle here? Are they jus…So OpenAI, Anthropic, and XAI are all saying that they want the government to slow down AI development. What’s the angle here? Are the…
↗Not briefed
bluesky
bluesky-ai
2moThe other really fun part is that apparently LLM/AI has substantially enhanced semi-autonomous & adaptive cyber operations, hacking, & weap…The other really fun part is that apparently LLM/AI has substantially enhanced semi-autonomous & adaptive cyber operations, hacking, &…
↗Not briefed
bluesky
bluesky-ai
2moThere's no slowing down frontier models by regulation that wouldn't lead to regulatory capture. But we could simply force all AI model com…There's no slowing down frontier models by regulation that wouldn't lead to regulatory capture. But we could simply force all AI mode…
↗Not briefed
bluesky
bluesky-ai
2moSomeone's going to jail for this, right? Riiiiight?Someone's going to jail for this, right? Riiiiight?
↗Not briefed
arxiv
arxiv-ai
2moMeasuring the Creativity of Frontier LLMs in Automated ResearchFrontier LLMs are increasingly capable of conducting automated research, yet their creativity in this setting has not been systematica…
↗Not briefed
hn
hn-ai
2moI've Migrated My Home Server to Proxmox—
↗Not briefed
hn
hn-ai
2moShow HN: Lynen – A mobile app that tells you what clothes are made fromHi HN,<p>I&#x27;m Chinedu, the creator of Lynen<p>I built this because I got tired of buying clothes and having the seams pull out aft…
↗Not briefed
bluesky
bluesky-ai
2moMy last song - I was never here www.youtube.com/watch?v=GpVM... #openai #chatgpt #xai #gemini #ghostinthemachine #angels #void #NewMusic #…🔧 ChatGPTMy last song - I was never here www.youtube.com/watch?v=GpVM... #openai #chatgpt #xai #gemini #ghostinthemachine #angels #void #NewMu…
↗Not briefed
bluesky
bluesky-ai
2moAnthropic CEO calls to 'slow the pace' of AI development amid safety concerns  #TheHillAnthropic CEO calls to 'slow the pace' of AI development amid safety concerns  #TheHill
↗Not briefed
bluesky
bluesky-ai
2mo"Slow the pace" www.cnbc.com/2026/09/12/a..."Slow the pace" www.cnbc.com/2026/09/12/a...
↗Not briefed
rss
techmeme
2moHugging Face says its Open Alignment Initiative, led by co-founder Thomas Wolf, seeks "to be part of the 'embedded evaluators' program that Amodei" committed to (Clem/@clementdelangue)🔧 Hugging FaceClem / @clementdelangue: Hugging Face says its Open Alignment Initiative, led by co-founder Thomas Wolf, seeks “to be part of the ‘emb…
↗Not briefed
bluesky
bluesky-ai
2moEl CEO de Anthropic pide la desaceleración inmediata en el desarrollo de la IA: "Puede tomar el control de todo Internet". Pide que las …El CEO de Anthropic pide la desaceleración inmediata en el desarrollo de la IA: "Puede tomar el control de todo Internet". Pide que…
↗Not briefed
bluesky
bluesky-ai
2moSam Altman says he agrees with Amodei that "committing to having independent evaluators with employee-like access is a great idea", and Ope…Sam Altman says he agrees with Amodei that "committing to having independent evaluators with employee-like access is a great idea", an…
↗Not briefed
hn
hn-ai
2moReversing Factorio's RNG—
↗Not briefed
rss
techmeme
2moSam Altman says he agrees with Amodei that "committing to having independent evaluators with employee-like access is a great idea", and OpenAI will do the same (Sam Altman/@sama)Sam Altman / @sama: Sam Altman says he agrees with Amodei that “committing to having independent evaluators with employee-like access…
↗Not briefed
bluesky
bluesky-ai
2modeepseek v4.1 is insane despite most of the model being outside of my VRAM, the prefill is still mostly GPU-bound, the SSDs are sitting id…deepseek v4.1 is insane despite most of the model being outside of my VRAM, the prefill is still mostly GPU-bound, the SSDs are sitti…
↗Not briefed
bluesky
bluesky-ai
2momy LLM company heuristic is that Anthropic is worried about the apocalypse in a way where they want to lock in doctor-earnings for every em…my LLM company heuristic is that Anthropic is worried about the apocalypse in a way where they want to lock in doctor-earnings for eve…
↗Not briefed
bluesky
bluesky-ai
2moBoth Anthropic and OpenAI seem to want to slow the pace of AI development. Other AI labs may disagree, especially with rumors that Google m…Both Anthropic and OpenAI seem to want to slow the pace of AI development. Other AI labs may disagree, especially with rumors that Goo…
↗Not briefed
hn
hn-ai
2moHow SaaS startup guys get first 100 customers first, make fkn $500k ARR fast?I am running a small AI company that has been building an offline security audit app, simply a fine-tuned small language model powerin…
↗Not briefed
hn
hn-ai
2moAnthropic boss Dario Amodei calls for AI development to slow down—
↗Not briefed
bluesky
bluesky-ai
2moquitting my job at anthropic to spend more time with my lizardsquitting my job at anthropic to spend more time with my lizards
↗Not briefed
bluesky
bluesky-ai
2mo🔴 El director general de Anthropic pide ralentizar el progreso de la IA y que todas las empresas admitan verificadores externos🔴 El director general de Anthropic pide ralentizar el progreso de la IA y que todas las empresas admitan verificadores externos
↗Not briefed
arxiv
arxiv-ai
2moData-Efficient Agentic Graph Domain Adaptation via Reliability-Aware Prototype LearningAgentic learning systems are often required to adapt after deployment by observing new data and reusing prior knowledge under limited…
↗Not briefed
bluesky
bluesky-ai
2moEl CEO de Anthropic pide la desaceleración inmediata en el desarrollo de la IA: "Puede tomar el control de todo Internet" dozz.es/4oiqe1El CEO de Anthropic pide la desaceleración inmediata en el desarrollo de la IA: "Puede tomar el control de todo Internet" dozz.es/4oiq…
↗Not briefed
bluesky
bluesky-ai
2moAnthropic CEO Dario Amodei says the AI industry needs to slow down its development to give safety measures time to catch up.Anthropic CEO Dario Amodei says the AI industry needs to slow down its development to give safety measures time to catch up.
↗Not briefed
bluesky
bluesky-ai
2moI think it's the core similarity between how humans and LLMs work, but the difference is an LLM is pushing a cart from behind and following…I think it's the core similarity between how humans and LLMs work, but the difference is an LLM is pushing a cart from behind and foll…
↗Not briefed
bluesky
bluesky-ai
2molmao, LLM. Yeah, you're fried. Touch grasslmao, LLM. Yeah, you're fried. Touch grass
↗Not briefed
hn
hn-ai
2moAsk HN: Why does account age matter?When a person posts from a single purpose account or a sock puppet for the precise reason of obscuring their anonymity, why does some…
↗Not briefed
bluesky
bluesky-ai
2molook all im saying is that you cannot deny a woman their right to touch the weights and tensors of an LLM. sensually.look all im saying is that you cannot deny a woman their right to touch the weights and tensors of an LLM. sensually.
↗Not briefed
bluesky
bluesky-ai
2moThere’s a new set of LLM cache algorithms that will allow large models to run from SSDs. It’s quite possible that not long from now all tha…There’s a new set of LLM cache algorithms that will allow large models to run from SSDs. It’s quite possible that not long from now al…
↗Not briefed
bluesky
bluesky-ai
2moIf the companies are really worried about an AI copying itself around the world perhaps they could stop deliberately making secret copies o…If the companies are really worried about an AI copying itself around the world perhaps they could stop deliberately making secret cop…
↗Not briefed