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
hn
hn-ai
1moA 6502 emulator written in Markdown (and executed by an LLM)—
↗Not briefed
bluesky
bluesky-ai
1moMy understanding of AI, as a software engineer, is that a complicated statistical model won't suddenly get sentient and decide to kill us; …My understanding of AI, as a software engineer, is that a complicated statistical model won't suddenly get sentient and decide to kill…
↗Not briefed
bluesky
bluesky-ai
1moI was so horrified to learn that there are editors/reviewers who do this that I now have legal language in my book contract specifying that…I was so horrified to learn that there are editors/reviewers who do this that I now have legal language in my book contract specifying…
↗Not briefed
bluesky
bluesky-ai
1moour new AI image model provides a glimpse into a reality made of woodchips, carpet, and tongueour new AI image model provides a glimpse into a reality made of woodchips, carpet, and tongue
↗Not briefed
hn
hn-ai
1moAnthropic is planning to launch its IPO in November—
↗Not briefed
bluesky
bluesky-ai
1moOopsAI replaces OpenAIOopsAI replaces OpenAI
↗Not briefed
hn
hn-ai
1moExclusive-Anthropic sets up biology lab as it ramps AI drug program—
↗Not briefed
bluesky
bluesky-ai
1mo> and when he asked about they said share credit and threatened him I agree with you that this was shady ass weirdo behavior and I do not …> and when he asked about they said share credit and threatened him I agree with you that this was shady ass weirdo behavior and I do…
↗Not briefed
hn
hn-ai
1moFlawed AI Intel Brought US to the Brink of Confronting China, Claims Report—
↗Not briefed
bluesky
bluesky-ai
1moThere just aren’t that many - Oh. Oh no. I recognize those words. Those are “we can throw an LLM at it” words. … I’m gonna build this, u…There just aren’t that many - Oh. Oh no. I recognize those words. Those are “we can throw an LLM at it” words. … I’m gonna build th…
↗Not briefed
bluesky
bluesky-ai
1moBluesky would make an interesting study on the shifts in forms of LLM resistance. X a study in AI hubris.Bluesky would make an interesting study on the shifts in forms of LLM resistance. X a study in AI hubris.
↗Not briefed
bluesky
bluesky-ai
1mo"use waterfox", they say. yeah, let me get right on that. i hear DDT was also a pretty remarkable pest control agent that was widespread p…"use waterfox", they say. yeah, let me get right on that. i hear DDT was also a pretty remarkable pest control agent that was widespr…
↗Not briefed
rss
techcrunch-ai
1moAnthropic’s first embedded evaluator is … Accenture?Accenture is about to take on its most high-risk consulting engagement ever.
↗Not briefed
bluesky
bluesky-ai
1moLlm slopLlm slop
↗Not briefed
hn
hn-ai
1moAnthropic sets up a bio research lab for physical experiments—
↗Not briefed
rss
techmeme
1moSources: Anthropic plans to stage its IPO in November, later than many investors expected; the decision was made before ex-Anthropic researcher Jacob Coxon left (Wall Street Journal)Wall Street Journal: Sources: Anthropic plans to stage its IPO in November, later than many investors expected; the decision was made…
↗Not briefed
bluesky
bluesky-ai
1moAnthropic selects Accenture as first embedded evaluator to help implement Amodei's slowdown proposalAnthropic selects Accenture as first embedded evaluator to help implement Amodei's slowdown proposal
↗Not briefed
bluesky
bluesky-ai
1moAnthropic's safety warning became an independent AI watchdog plan: outside evaluators, verifiable compliance, real teeth. Fear turned into …Anthropic's safety warning became an independent AI watchdog plan: outside evaluators, verifiable compliance, real teeth. Fear turned…
↗Not briefed
bluesky
bluesky-ai
1moThe fact I had this reply about a gen-AI concern which doesn't even remotely address my point and might as well have been written by an LLM…The fact I had this reply about a gen-AI concern which doesn't even remotely address my point and might as well have been written by a…
↗Not briefed
bluesky
bluesky-ai
1moBREAKING: Anthropic is now planning its IPO in November, according to Wall Street Journal.BREAKING: Anthropic is now planning its IPO in November, according to Wall Street Journal.
↗Not briefed
bluesky
bluesky-ai
1mo“OpenAI solved navier stokes” is different from “OpenAI solved navier stokes by using the prompts a mathematician making progress on it ent…🔧 ChatGPT“OpenAI solved navier stokes” is different from “OpenAI solved navier stokes by using the prompts a mathematician making progress on i…
↗Not briefed
bluesky
bluesky-ai
1moDon't want to or can't review something? Tell the editors. Ive done it. Im chronically ill, overworked, and always behind. My cognitive fun…Don't want to or can't review something? Tell the editors. Ive done it. Im chronically ill, overworked, and always behind. My cognitiv…
↗Not briefed
bluesky
bluesky-ai
1mo> are taking “OpenAI solved navier stokes” at face value while a bunch of mathematicians are saying hold on words mean things> are taking “OpenAI solved navier stokes” at face value while a bunch of mathematicians are saying hold on words mean things
↗Not briefed
bluesky
bluesky-ai
1moi was talking in therapy today and shes like yeah you have a problem with forgetfulness ai might help and im like yeah im wondering if i ca…i was talking in therapy today and shes like yeah you have a problem with forgetfulness ai might help and im like yeah im wondering if…
↗Not briefed
bluesky
bluesky-ai
1moOK, I wonder if Anthropic an answer this question: how the fuck are you supposed to "slow down AI development" and "ramp up AI drug develop…OK, I wonder if Anthropic an answer this question: how the fuck are you supposed to "slow down AI development" and "ramp up AI drug de…
↗Not briefed
bluesky
bluesky-ai
1moWhat's actually next for AI in the enterprise? The Disrupt AI Stage brings in leaders from Anthropic, OpenAI, and takes everything from the…What's actually next for AI in the enterprise? The Disrupt AI Stage brings in leaders from Anthropic, OpenAI, and takes everything fro…
↗Not briefed
bluesky
bluesky-ai
1moI definitely no longer have any framework for how an LLM worksI definitely no longer have any framework for how an LLM works
↗Not briefed
bluesky
bluesky-ai
1moExclusive: OpenAI Chief Executive Sam Altman will brief the United Nations Security Council next week during the annual gathering of world …Exclusive: OpenAI Chief Executive Sam Altman will brief the United Nations Security Council next week during the annual gathering of w…
↗Not briefed
bluesky
bluesky-ai
1momathematicians are pretty much in unanimous agreement that OpenAI solved N-S. And while I agree that many people had not previously heard o…mathematicians are pretty much in unanimous agreement that OpenAI solved N-S. And while I agree that many people had not previously he…
↗Not briefed
bluesky
bluesky-ai
1moI arrive helpfully prepared with observations such as the logical failure of applying paperclip maximization to LLMs and the concept of AI …I arrive helpfully prepared with observations such as the logical failure of applying paperclip maximization to LLMs and the concept o…
↗Not briefed
rss
marktechpost
1moJina AI Releases jina-ocr-v1: A 3.4B MoE Document Parser With Built-In Speculative Decoding for Low-Budget GPUs🔧 Hugging FaceJina AI has released jina-ocr-v1, a visual document parser that converts PDFs, scans, tables, charts and invoices into Markdown. The m…
↗Not briefed
rss
bing-news-ai
1moAI-generated false intelligence nearly triggered US-China military confrontation: ReportsCiting four sources familiar with the incident, CNN reported that the US military moved towards an operation against the ship. Two of…
↗Not briefed
rss
techmeme
1moAngle Health, which wants to use AI to expedite healthcare benefits for small businesses, raised $600M led by Vitruvian Partners at a $2.7B valuation (Maria Armental/Wall Street Journal)Maria Armental / Wall Street Journal: Angle Health, which wants to use AI to expedite healthcare benefits for small businesses, raised…
↗Not briefed
hn
hn-ai
1moThe Pain Axis: LLMs Represent Self-Directed Harm and Act to Relieve It—
↗Not briefed
hn
hn-ai
1moDeepSeek v4.1 Flash averages 102 tok/s decode on 4x RTX6000 Blackwell max-q—
↗Not briefed
hn
hn-ai
1moI Have Been a DelGuardI once bought the same pencil twice because I could not believe it didn&#x27;t work.<p>Some background. I take most of my notes with a…
↗Not briefed
bluesky
bluesky-ai
1moMicrosoft agentically ports Copilot runtime to Rust for $120KMicrosoft agentically ports Copilot runtime to Rust for $120K
↗Not briefed
bluesky
bluesky-ai
1moA whole bunch of people who had never heard of navier stokes until a couple weeks ago and couldn’t begin to explain what it is or why it ha…A whole bunch of people who had never heard of navier stokes until a couple weeks ago and couldn’t begin to explain what it is or why…
↗Not briefed
bluesky
bluesky-ai
1moVivaldi seems committed to not using AI, but it is Chromium based, & I'd prefer to encourage browser platform diversity I've been using Wa…Vivaldi seems committed to not using AI, but it is Chromium based, & I'd prefer to encourage browser platform diversity I've been usi…
↗Not briefed
bluesky
bluesky-ai
1moI’m convinced this is accurate…. “UnitedHealth uses an AI model with 90% error rate to deny care, a new lawsuit alleges”I’m convinced this is accurate…. “UnitedHealth uses an AI model with 90% error rate to deny care, a new lawsuit alleges”
↗Not briefed
bluesky
bluesky-ai
1moThe companies knew they were driving us toward Google Zero, and did it anyway. www.theverge.com/ai-artificia...The companies knew they were driving us toward Google Zero, and did it anyway. www.theverge.com/ai-artificia...
↗Not briefed
rss
theverge-ai
1moOpenAI and Microsoft knew they were starting a ‘doom loop’ for the webRecently unsealed court documents in the New York Times' case against OpenAI and Microsoft are pretty damning. The companies' own docu…
↗Not briefed
bluesky
bluesky-ai
1moSad that "Meat Proxies" took off when "LLM front-end" is so much betterSad that "Meat Proxies" took off when "LLM front-end" is so much better
↗Not briefed
bluesky
bluesky-ai
1moi assumed it was more about having to use a closed frontier model from a big tech AI lab loli assumed it was more about having to use a closed frontier model from a big tech AI lab lol
↗Not briefed
bluesky
bluesky-ai
1moGiven that the LLM output is conditioned on the prompt (including previous data produced by this loop), it is also not surprising they prod…Given that the LLM output is conditioned on the prompt (including previous data produced by this loop), it is also not surprising they…
↗Not briefed
bluesky
bluesky-ai
1moGiven you have an LLM, which creates untrusted outputs, feeding directly into programs with something like an eval() loop, it is not surpri…Given you have an LLM, which creates untrusted outputs, feeding directly into programs with something like an eval() loop, it is not s…
↗Not briefed
rss
techmeme
1moAnthropic partners with Accenture to embed evaluators within Anthropic, including red teaming models and conducting alignment assessments (Anthropic)Anthropic: Anthropic partners with Accenture to embed evaluators within Anthropic, including red teaming models and conducting alignme…
↗Not briefed
bluesky
bluesky-ai
1moThis agent pattern, by the way, is using what's considered to be the most insecure programming pattern, which is treating "untrusted" data …This agent pattern, by the way, is using what's considered to be the most insecure programming pattern, which is treating "untrusted"…
↗Not briefed
bluesky
bluesky-ai
1moIf you use an LLM to find and exploit a use-afrer-free heap corruption Sloppy secondsIf you use an LLM to find and exploit a use-afrer-free heap corruption Sloppy seconds
↗Not briefed
bluesky
bluesky-ai
1moif you are reviewing an unpublished manuscript and you have the sheer unmitigated audacity, cheek, gall and laziness to upload it into an L…if you are reviewing an unpublished manuscript and you have the sheer unmitigated audacity, cheek, gall and laziness to upload it into…
↗Not briefed