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
2moFly Arena: Simulating a Fruit Fly with a Connectome and Physics—
↗Not briefed
bluesky
bluesky-ai
2mo【2ch】中華AI、すべからく裏でClaudeに丸投げしていたことが暴露されて国民がパニックに陥る【2ch】中華AI、すべからく裏でClaudeに丸投げしていたことが暴露されて国民がパニックに陥る
↗Not briefed
bluesky
bluesky-ai
2moThose who say that AI is going to destroy humanity within the next few years owe it to the rest of us to explain how. What exactly is it th…Those who say that AI is going to destroy humanity within the next few years owe it to the rest of us to explain how. What exactly is…
↗Not briefed
bluesky
bluesky-ai
2moOpenAI’s Sam Altman says it would be ‘ill-advised’ to go public in 2026OpenAI’s Sam Altman says it would be ‘ill-advised’ to go public in 2026
↗Not briefed
bluesky
bluesky-ai
2mo"When you are racing towards a cliff, you don’t just ease up on the gas pedal. You hit the brakes.""When you are racing towards a cliff, you don’t just ease up on the gas pedal. You hit the brakes."
↗Not briefed
rss
techcrunch-ai
2moOpenAI’s Sam Altman says it would be ‘ill-advised’ to go public in 2026While OpenAI has filed confidentially for an IPO, the company will not be going public this year, according to CEO Sam Altman.
↗Not briefed
hn
hn-ai
2moI built a harness and now it is my daily driver for local agents—
↗Not briefed
bluesky
bluesky-ai
2moAI is definitely taking & changing jobs in those fields right now, but the results (except in some areas of medicine) are often worse than …AI is definitely taking & changing jobs in those fields right now, but the results (except in some areas of medicine) are often worse…
↗Not briefed
arxiv
arxiv-ai
2moExact Finite Attention Responses From RoPE DerivativesWe derive exact local responses for attention interventions, allowing candidate edits to be scored from a cached baseline and one back…
↗Not briefed
bluesky
bluesky-ai
2mo@horsedisc.bsky.social isn‘t even based on an LLM & it’s some of the best material on blue sky.@horsedisc.bsky.social isn‘t even based on an LLM & it’s some of the best material on blue sky.
↗Not briefed
bluesky
bluesky-ai
2moOpenAI is calling for a Public Wealth Fund UBI so that everyone, not just investors, owns a piece of the AI future. This is now mainstream…OpenAI is calling for a Public Wealth Fund UBI so that everyone, not just investors, owns a piece of the AI future. This is now mains…
↗Not briefed
bluesky
bluesky-ai
2mo“We must slow the pace at which we improve the capabilities of AI models,” Dario Amodei, co-founder of AI giant Anthropic, said in an essay…“We must slow the pace at which we improve the capabilities of AI models,” Dario Amodei, co-founder of AI giant Anthropic, said in an…
↗Not briefed
bluesky
bluesky-ai
2moLike, you have to admit, OpenAi being the only company that AI ends up destroying would be an incredible bitLike, you have to admit, OpenAi being the only company that AI ends up destroying would be an incredible bit
↗Not briefed
bluesky
bluesky-ai
2moYou don't have to imagine future "AI" models doing harm. You don't have to resort to science-fictional ideas. *This* is happening *now*: m…You don't have to imagine future "AI" models doing harm. You don't have to resort to science-fictional ideas. *This* is happening *now…
↗Not briefed
bluesky
bluesky-ai
2moThis is my opinions on calls to "pause" research on "superintelligence"This is my opinions on calls to "pause" research on "superintelligence"
↗Not briefed
bluesky
bluesky-ai
2moSorry, I didn’t realize your context window was so small. I’ll rephrase: Where in the process do you see LLM legal research saving time ove…Sorry, I didn’t realize your context window was so small. I’ll rephrase: Where in the process do you see LLM legal research saving tim…
↗Not briefed
arxiv
arxiv-ai
2moTo do($x$) or not to do($x$): Medical Image Counterfactuals for Dataset AugmentationMedical image analysis is often hindered by biased datasets, which can lead to biased models and limited clinical applicability. A pro…
↗Not briefed
bluesky
bluesky-ai
2moHow did he pass is job interview? The same reason he's all over the press & everyone in the "left" is captured by these people. I'm gon…How did he pass is job interview? The same reason he's all over the press & everyone in the "left" is captured by these people. I'…
↗Not briefed
bluesky
bluesky-ai
2mo💥🧵 #BREAKING | Anthropic CEO Warns AI Agents Could Threaten The Internet Within Months The chief executive of San Francisco-based Anthro…💥🧵 #BREAKING | Anthropic CEO Warns AI Agents Could Threaten The Internet Within Months The chief executive of San Francisco-based A…
↗Not briefed
bluesky
bluesky-ai
2moOpenAI has terminated their Pro 20x $200/mo plan. I suspect the price wars have begun, and they're going upOpenAI has terminated their Pro 20x $200/mo plan. I suspect the price wars have begun, and they're going up
↗Not briefed
bluesky
bluesky-ai
2moAre they bad at it, or is an LLM designed to literally fabricate words from an immense dartboard simply not an appropriate tool to use for …Are they bad at it, or is an LLM designed to literally fabricate words from an immense dartboard simply not an appropriate tool to use…
↗Not briefed
bluesky
bluesky-ai
2moIt seems like 🇺🇸 & the 🌎 at Large has had a problem of late, or at least a certain portion of our populations does with listening to exp…It seems like 🇺🇸 & the 🌎 at Large has had a problem of late, or at least a certain portion of our populations does with listening t…
↗Not briefed
bluesky
bluesky-ai
2moI really don't like the idea of being killed by Claude. It sounds ridiculous. If an AI is to kill me, it should be called something like T…I really don't like the idea of being killed by Claude. It sounds ridiculous. If an AI is to kill me, it should be called something l…
↗Not briefed
bluesky
bluesky-ai
2mowe already have this openrouter.ai/deepseek/dee... (and before you mention it: none of the third party providers have any incentive to sel…🔧 OpenRouterwe already have this openrouter.ai/deepseek/dee... (and before you mention it: none of the third party providers have any incentive t…
↗Not briefed
bluesky
bluesky-ai
2moOn va passer la semaine pro sur le sujet, désolé pour les indisponibilités du moment, c'est intolérable. On va devoir faire une page "capt…On va passer la semaine pro sur le sujet, désolé pour les indisponibilités du moment, c'est intolérable. On va devoir faire une page…
↗Not briefed
bluesky
bluesky-ai
2mothe problem is not that she is making the pretty uncontroversial point that OpenAI is being incredibly reckless, but that she's taking the …the problem is not that she is making the pretty uncontroversial point that OpenAI is being incredibly reckless, but that she's taking…
↗Not briefed
hn
hn-ai
2moShow HN: AgentRuleBench, does AI agents violate inferred architecture rules?—
↗Not briefed
bluesky
bluesky-ai
2moWhich is relevant here. Where do you see LLM research saving time over human research?Which is relevant here. Where do you see LLM research saving time over human research?
↗Not briefed
arxiv
arxiv-ai
2moSemantic Knowledge Technologies: what the Semantic Web lost sight of, and what it never hadThe Semantic Web set out to give information a machine-interpretable form so that software could integrate and reason over it. Its sta…
↗Not briefed
hn
hn-ai
2moOpenAI delaying IPO amid AI safety concerns, Sam Altman says—
↗Not briefed
hn
hn-ai
2moShow HN: An independent directory of AI misalignment reportsHey HN,<p>In light of the concerns regarding AI safety, I created a website to track all known cases of misalignment.<p>I will keep th…
↗Not briefed
rss
gary-marcus
2moCould rogue agent swarms take over the entire internet in the next six months?Dario Amodei seems to think so.
↗Not briefed
bluesky
bluesky-ai
2moEven my LLM generated avatar feels called outEven my LLM generated avatar feels called out
↗Not briefed
bluesky
bluesky-ai
2moOpenAI’dan Hükümetlere Çağrı: "Yapay Zeka Güvenlik Yasaları Zorunlu Hale Getirilmeli!" www.airehber.com.tr/timeline/ope... #openai #aisa…OpenAI’dan Hükümetlere Çağrı: "Yapay Zeka Güvenlik Yasaları Zorunlu Hale Getirilmeli!" www.airehber.com.tr/timeline/ope... #openai…
↗Not briefed
bluesky
bluesky-ai
2mo"[An LLM] might give you a reasonable facsimile, and the most dangerous thing about a facsimile is that it will eventually convince you tha…"[An LLM] might give you a reasonable facsimile, and the most dangerous thing about a facsimile is that it will eventually convince yo…
↗Not briefed
bluesky
bluesky-ai
2moyou're imagining a hostile hacker AI when there's real hostile human hackers already using AI tools, AI somehow acting autonomously (which …you're imagining a hostile hacker AI when there's real hostile human hackers already using AI tools, AI somehow acting autonomously (w…
↗Not briefed
bluesky
bluesky-ai
2moMeanwhile - futurism.com/artificial-i...Meanwhile - futurism.com/artificial-i...
↗Not briefed
bluesky
bluesky-ai
2moI’m totally sure eddie is correct here and the anthropic S1 filing will actually just be fraud somehowI’m totally sure eddie is correct here and the anthropic S1 filing will actually just be fraud somehow
↗Not briefed
bluesky
bluesky-ai
2moit's actually kinda funny how much AI has cooked the brains of everyone employed at bluesky. i genuinely don't know if anyone actually work…it's actually kinda funny how much AI has cooked the brains of everyone employed at bluesky. i genuinely don't know if anyone actually…
↗Not briefed
arxiv
arxiv-ai
2moTalking to Me or Someone Else? Rethinking Talk-to-Me Detection in Egocentric VideosOnline understanding of who is talking to the camera wearer is a key capability for egocentric social interaction. However, existing t…
↗Not briefed
hn
hn-ai
2moUsing Elm and Rust for one 3D web app—
↗Not briefed
bluesky
bluesky-ai
2mothis makes the sympathetic reading of Anthropic "can we all please slow down public releases, use our cutting edge models to patch giant se…this makes the sympathetic reading of Anthropic "can we all please slow down public releases, use our cutting edge models to patch gia…
↗Not briefed
bluesky
bluesky-ai
2moOpenAI let it be known back in June their IPO would be "postponed." And it wasn't for fucking "safety" reasons. www.nytimes.com/2026/06/25…OpenAI let it be known back in June their IPO would be "postponed." And it wasn't for fucking "safety" reasons. www.nytimes.com/2026/…
↗Not briefed
bluesky
bluesky-ai
2moIf I'm being charitable I would say the people working on LLMs are suffering from LLM induced psychosis.If I'm being charitable I would say the people working on LLMs are suffering from LLM induced psychosis.
↗Not briefed
bluesky
bluesky-ai
2moIf Dario Amodei is truly concerned about slowing AI development, one of the easiest ways to do that would be by pausing or scaling back his…If Dario Amodei is truly concerned about slowing AI development, one of the easiest ways to do that would be by pausing or scaling bac…
↗Not briefed
bluesky
bluesky-ai
2moI will take Kathryn over an LLM every single day of the weekI will take Kathryn over an LLM every single day of the week
↗Not briefed
bluesky
bluesky-ai
2mo* Iran used Claude to target US Navy in Middle East, Anthropic says ~ Navy Times 9-11-26 www.navytimes.com/news/your-na...* Iran used Claude to target US Navy in Middle East, Anthropic says ~ Navy Times 9-11-26 www.navytimes.com/news/your-na...
↗Not briefed
bluesky
bluesky-ai
2mo100%. tbh i think it’s very silly to think that US laws are going to stop the true believers at ant or openai from working on this stuff. t…100%. tbh i think it’s very silly to think that US laws are going to stop the true believers at ant or openai from working on this stu…
↗Not briefed
bluesky
bluesky-ai
2moDeepSeek's Flash 4.1: bigger and leaner DeepSeek unveiled Flash V4.1, a 763B parameter model 2.5x larger, yet with significantly lower mem…DeepSeek's Flash 4.1: bigger and leaner DeepSeek unveiled Flash V4.1, a 763B parameter model 2.5x larger, yet with significantly lowe…
↗Not briefed
hn
hn-ai
2moShow HN: SimTower, Decompiled, Rewritten, and in the BrowserI wanted to re-play SimTower but found it too difficult to get running through emulation. Instead, I worked for 6 months with Claude C…
↗Not briefed