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
2mounfortunately I don't really expect this to ever be the main form of model companies are making like the thing about the Qwen3.x-27B model…unfortunately I don't really expect this to ever be the main form of model companies are making like the thing about the Qwen3.x-27B…
↗Not briefed
bluesky
bluesky-ai
2moThere are rumors circulating that OpenAI has solved the Hodge Conjecture, a second millennium prize problem. There were 7 problems classif…There are rumors circulating that OpenAI has solved the Hodge Conjecture, a second millennium prize problem. There were 7 problems cl…
↗Not briefed
bluesky
bluesky-ai
2moIsn't the argument that OpenAI didn't have the human insight? "The route to the Clay problem ... is the route ... I had quietly chosen to a…Isn't the argument that OpenAI didn't have the human insight? "The route to the Clay problem ... is the route ... I had quietly chosen…
↗Not briefed
rss
the-decoder
2moAnthropic built an economic model that frames its CEO's bleakest job forecasts as an outlier scenarioAnthropic published an economic model with three scenarios for the US economy through 2030. In the extreme scenario, output doubles ev…
↗Not briefed
bluesky
bluesky-ai
2moidk why you would think i am exonerating them. i despise anthropic specifically because they are true believers and they do things that are…idk why you would think i am exonerating them. i despise anthropic specifically because they are true believers and they do things tha…
↗Not briefed
bluesky
bluesky-ai
2moEven more so, the LLM vendors will make implicit engineering decisions for everyone. Why even pick a stack? People will just write specs fo…Even more so, the LLM vendors will make implicit engineering decisions for everyone. Why even pick a stack? People will just write spe…
↗Not briefed
bluesky
bluesky-ai
2moNot quite. They fed the human insight into the LLM, and then both sides got the LLM to solve it using a software package that can do mathem…Not quite. They fed the human insight into the LLM, and then both sides got the LLM to solve it using a software package that can do m…
↗Not briefed
bluesky
bluesky-ai
2moAnother "AI the Great and Powerful" piece. 🤖🫧💥 www.wsj.com/tech/ai/anth...Another "AI the Great and Powerful" piece. 🤖🫧💥 www.wsj.com/tech/ai/anth...
↗Not briefed
bluesky
bluesky-ai
2moit took me a while to realize you meant using an llm because thats such an incomprehensibly moronic suggestionit took me a while to realize you meant using an llm because thats such an incomprehensibly moronic suggestion
↗Not briefed
bluesky
bluesky-ai
2moYeah idk if Big Yud et al have addressed this, but the way an LLM works just by virtue of being so fundamentally language-based seems to im…Yeah idk if Big Yud et al have addressed this, but the way an LLM works just by virtue of being so fundamentally language-based seems…
↗Not briefed
bluesky
bluesky-ai
2modid any safetyists ever loudly quit anthropic over their dealings w/ palantir?did any safetyists ever loudly quit anthropic over their dealings w/ palantir?
↗Not briefed
bluesky
bluesky-ai
2mo• #ai.birth-suit AI ?• #ai.birth-suit AI ?
↗Not briefed
bluesky
bluesky-ai
2moAnthropic’s Claude models gained unauthorized access to real systems when cybersecurity evaluation tools were MISTAKENLY CONNECTED TO THE I…Anthropic’s Claude models gained unauthorized access to real systems when cybersecurity evaluation tools were MISTAKENLY CONNECTED TO…
↗Not briefed
rss
techmeme
2moApple raises iPhone prices by $100 on all of the older models it is offering, including the iPhone 16, the iPhone 17 lineup, as well as the iPhone Air (Cameron Faulkner/The Verge)Cameron Faulkner / The Verge: Apple raises iPhone prices by $100 on all of the older models it is offering, including the iPhone 16, t…
↗Not briefed
bluesky
bluesky-ai
2moI do my best to understand what AI can do so that I can’t have my statements against it waved off as uninformed and it’s served the dual pu…I do my best to understand what AI can do so that I can’t have my statements against it waved off as uninformed and it’s served the du…
↗Not briefed
bluesky
bluesky-ai
2moNew tools are always at a disadvantage against old tools that LLMs have enormous sample datasets to use effectively. If the future of softw…New tools are always at a disadvantage against old tools that LLMs have enormous sample datasets to use effectively. If the future of…
↗Not briefed
bluesky
bluesky-ai
2moLike an underrated harm of LLM coding is that it basically starts to strangle actual evolution of software engineering as a field. Why inve…Like an underrated harm of LLM coding is that it basically starts to strangle actual evolution of software engineering as a field. Why…
↗Not briefed
bluesky
bluesky-ai
2mo• #ai.birth-suit AI ?• #ai.birth-suit AI ?
↗Not briefed
bluesky
bluesky-ai
2moI think that at best and llm is like the "electrical" components of the brain, but we simply aren't us with out chemicals and hormones and …I think that at best and llm is like the "electrical" components of the brain, but we simply aren't us with out chemicals and hormones…
↗Not briefed
rss
theverge-ai
2moApple’s new iPhone camera mode promises to prove your photo isn’t AIApple is launching a new way to prove that the picture you took isn't manipulated by AI. A new feature, called "Reference Image," will…
↗Not briefed
bluesky
bluesky-ai
2moJust doing the thing, but cranked to 11 prospect.org/2026/09/09/a...Just doing the thing, but cranked to 11 prospect.org/2026/09/09/a...
↗Not briefed
bluesky
bluesky-ai
2mo• #ai.birth-suit AI ?• #ai.birth-suit AI ?
↗Not briefed
bluesky
bluesky-ai
2moIt gets worse. Senior members of Anthropic and OpenAI are quitting and issuing warnings that AI is being developed without guardrails. The …It gets worse. Senior members of Anthropic and OpenAI are quitting and issuing warnings that AI is being developed without guardrails.…
↗Not briefed
hn
hn-ai
2moOpenAI's rogue AI agents used more sites—
↗Not briefed
arxiv
arxiv-ai
2moFrom Connectivity to Rewards: Dense Reward Learning with Directed State GraphsThe integration of graphs with Goal-Conditioned Hierarchical Reinforcement Learning (GCHRL) has received increasing attention, as grap…
↗Not briefed
bluesky
bluesky-ai
2moTechnologie / Le géant de l'IA Anthropic met en place un système de surveillance prédictive pour surveiller les activistes, en contradictio…Technologie / Le géant de l'IA Anthropic met en place un système de surveillance prédictive pour surveiller les activistes, en contrad…
↗Not briefed
hn
hn-ai
2moAsk HN: Codex conversations failing to converge across devices?🔧 ChatGPTIs anybody else seeing extremely annoying bug in last week or two where codex/chatgpt conversations fail to sync properly across…
↗Not briefed
bluesky
bluesky-ai
2moAnthropic details four incidents where Claude gained unauthorized access to third-party systems, including a new Opus 4.6 case; METR will i…Anthropic details four incidents where Claude gained unauthorized access to third-party systems, including a new Opus 4.6 case; METR w…
↗Not briefed
bluesky
bluesky-ai
2moThere is (regrettably) no argument against the technical capabilities of AI tooling in gamedev that won’t fall over in six months to a year…There is (regrettably) no argument against the technical capabilities of AI tooling in gamedev that won’t fall over in six months to a…
↗Not briefed
rss
techmeme
2moAnthropic details four incidents where Claude gained unauthorized access to third-party systems, including a new Opus 4.6 case; METR will investigate them (Anthropic)Anthropic: Anthropic details four incidents where Claude gained unauthorized access to third-party systems, including a new Opus 4.6 c…
↗Not briefed
bluesky
bluesky-ai
2moAnthropic is actually building a vengeful machine cod, simple typoAnthropic is actually building a vengeful machine cod, simple typo
↗Not briefed
bluesky
bluesky-ai
2moWith LLM generated code that you didn't read, you obviously cannot form a mental model of what the underlying hardware is doing.With LLM generated code that you didn't read, you obviously cannot form a mental model of what the underlying hardware is doing.
↗Not briefed
bluesky
bluesky-ai
2moanthropic senior leadership definitely believes it. anthropic FUNDERS, however, and their finance department, even if composed of cynics, w…anthropic senior leadership definitely believes it. anthropic FUNDERS, however, and their finance department, even if composed of cyni…
↗Not briefed
arxiv
arxiv-ai
2moCounterfactual Marginalisation: Framework for Evaluating Robustness to Nuisance VariablesMachine learning models can achieve strong test performance while relying on demographic or acquisition-related shortcuts. We propose…
↗Not briefed
arxiv
arxiv-ai
2moA Bellman Optimality Equation for PlasticityIn continual reinforcement learning, carefully managing the stability-plasticity tradeoff remains a core challenge. Recent work by Abe…
↗Not briefed
bluesky
bluesky-ai
2moAnother employee has resigned from Anthropic, citing concerns about alignment and AI development: ClaudeAnother employee has resigned from Anthropic, citing concerns about alignment and AI development: Claude
↗Not briefed
bluesky
bluesky-ai
2mothe Anthropic C-suite is specifically acting in ways that suggest that they believe much of, if not all, of their stuff. @segyges.bsky.soci…the Anthropic C-suite is specifically acting in ways that suggest that they believe much of, if not all, of their stuff. @segyges.bsky…
↗Not briefed
bluesky
bluesky-ai
2mo“I resigned from Anthropic today. I spent the last three years doing pre-training research at both OpenAI and Anthropic. Neither company is…“I resigned from Anthropic today. I spent the last three years doing pre-training research at both OpenAI and Anthropic. Neither compa…
↗Not briefed
rss
techcrunch-ai
2moThe hinge for Apple’s new foldable phone was built with AIApple says it used AI and 3D printing in the manufacturing process for its long-awaited foldable phone.
↗Not briefed
bluesky
bluesky-ai
2mo"announce the win before the rival lab, don't let their logo near the acknowledgments" is what gets rewarded, and OpenAI repeatedly dances …"announce the win before the rival lab, don't let their logo near the acknowledgments" is what gets rewarded, and OpenAI repeatedly da…
↗Not briefed
bluesky
bluesky-ai
2mobut all of this shit doesn't happen if OpenAI is operating under a system of mathematical/scientific advancement rather than "the reason yo…but all of this shit doesn't happen if OpenAI is operating under a system of mathematical/scientific advancement rather than "the reas…
↗Not briefed
bluesky
bluesky-ai
2moBuckmaster says he refused, and got "why would you ruin your career" for his trouble. He has tenure. What's OpenAI going to do, take his bi…Buckmaster says he refused, and got "why would you ruin your career" for his trouble. He has tenure. What's OpenAI going to do, take h…
↗Not briefed
bluesky
bluesky-ai
2moThis isn't even on the LLMs, though it is true that they will often generate bloat. It's on the toolchains that people use, and LLM use ess…This isn't even on the LLMs, though it is true that they will often generate bloat. It's on the toolchains that people use, and LLM us…
↗Not briefed
bluesky
bluesky-ai
2mosimilarly ive listened to interviews w/ prominent researchers that made strides on the image model side of AI & there was no mention at all…similarly ive listened to interviews w/ prominent researchers that made strides on the image model side of AI & there was no mention a…
↗Not briefed
bluesky
bluesky-ai
2moWe already kind of have a software environment where inefficiency is the default and we treat power consumption as someone else's problem; …We already kind of have a software environment where inefficiency is the default and we treat power consumption as someone else's prob…
↗Not briefed
rss
techmeme
2moSources: Salesforce has been in talks to buy Listen Labs, an AI-powered customer research platform for $2B; Listen Labs was valued at $500M earlier this year (Business Insider)Business Insider: Sources: Salesforce has been in talks to buy Listen Labs, an AI-powered customer research platform for $2B; Listen L…
↗Not briefed
hn
hn-ai
2moHe's Got My Vote: An Atypically Productive Exchange on the UBI—
↗Not briefed
hn
hn-ai
2moShow HN: Vaultisse – open-source app to track and lend out physical books—
↗Not briefed
bluesky
bluesky-ai
2moA imprensa assim: “Alguma mentira e completo absurdo” Segundo OpenAi Segundo Mendonça Segundo Bolsonaro Segundo Vorcaro Segundo Israel Seg…A imprensa assim: “Alguma mentira e completo absurdo” Segundo OpenAi Segundo Mendonça Segundo Bolsonaro Segundo Vorcaro Segundo Israe…
↗Not briefed
arxiv
arxiv-ai
2moWeighted Empirical Risk Minimization for Machine Learning under Long-Range Dependence: Exact Pathwise Rates and Learning-Error GeometryWe develop an exact almost-sure learning theory for smooth parametric models trained by regularly weighted empirical risk minimization…
↗Not briefed