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
2moPeople say you'll be left behind if you dont adopt an LLM right now, but that doesnt at all guarantee you WONT be left behind if you do. Th…People say you'll be left behind if you dont adopt an LLM right now, but that doesnt at all guarantee you WONT be left behind if you d…
↗Not briefed
rss
techmeme
2moApple unveils the 2nm A20 Pro, featuring a 6-core CPU with two "desktop class" super cores, a 7-core GPU up to 40% faster, and 32 Neural Engine cores in total (Hartley Charlton/MacRumors)Hartley Charlton / MacRumors: Apple unveils the 2nm A20 Pro, featuring a 6-core CPU with two “desktop class” super cores, a 7-core GPU…
↗Not briefed
arxiv
arxiv-ai
2moIBIB: A Protocol for Measuring Enterprise AI Systems by Serving Route, Not Model IdentifierEnterprises deploy systems, not checkpoints. Usable capability depends jointly on weights, serving route, precision, output contract,…
↗Not briefed
bluesky
bluesky-ai
2mothat's not an LLM tho we do need better terminology, because AI is literally just a marketing term, and there is machine learning I don't …that's not an LLM tho we do need better terminology, because AI is literally just a marketing term, and there is machine learning I d…
↗Not briefed
hn
hn-ai
2moCodex GPT-5.6-sol Performance Tracker🔧 GPT-5—
↗Not briefed
bluesky
bluesky-ai
2moMichigan Democratic Senate candidate Abdul El-Sayed is calling for a pause on AI development after warnings from Anthropic researchers that…Michigan Democratic Senate candidate Abdul El-Sayed is calling for a pause on AI development after warnings from Anthropic researchers…
↗Not briefed
bluesky
bluesky-ai
2moI just heard someone refer to hand written non LLM generated software as trad code, and I feel very ill now.I just heard someone refer to hand written non LLM generated software as trad code, and I feel very ill now.
↗Not briefed
arxiv
arxiv-ai
2moLearning with Covariance Matrices: Principal Component Analysis Meets Learning with GraphsThis feature article provides an overview of the theoretical foundations for coVariance neural networks (VNNs), i.e., graph neural net…
↗Not briefed
bluesky
bluesky-ai
2moif you're a normie, if you've heard about them at all, Anthropic is a company lead by Anndrreesseenn and Peter Thiel and exists to replace …if you're a normie, if you've heard about them at all, Anthropic is a company lead by Anndrreesseenn and Peter Thiel and exists to rep…
↗Not briefed
hn
hn-ai
2moShieldCrash – Explained and Defender Tips—
↗Not briefed
bluesky
bluesky-ai
2moif one more vendor based technical support representative hits me with the "let me put your question into our internal LLM and see what it …if one more vendor based technical support representative hits me with the "let me put your question into our internal LLM and see wha…
↗Not briefed
bluesky
bluesky-ai
2moLearning today that a lot of people didn’t realize what Anthropic employees believeLearning today that a lot of people didn’t realize what Anthropic employees believe
↗Not briefed
bluesky
bluesky-ai
2moThe ultimate endorsement of Jacob Coxon’s thread has come via a repost by Eliezer Yudkowsky. (Yudkowsky sounded the alarm on AI alignment …The ultimate endorsement of Jacob Coxon’s thread has come via a repost by Eliezer Yudkowsky. (Yudkowsky sounded the alarm on AI align…
↗Not briefed
bluesky
bluesky-ai
2moin their defense -- and you know my opinion on Anthropic -- I do not believe this to be true at all they have a lot of true believers. der…in their defense -- and you know my opinion on Anthropic -- I do not believe this to be true at all they have a lot of true believers…
↗Not briefed
arxiv
arxiv-ai
2moNonmaximal sums of maximally monotone operators under Rockafellar's constraint qualificationWe construct counterexamples to Rockafellar's sum conjecture in which two maximally monotone operators satisfy the interior-domain con…
↗Not briefed
bluesky
bluesky-ai
2moThe only thing that makes it not COMPLETELY stupid is that OpenAI also believes this they just think the best case solution is that they wi…The only thing that makes it not COMPLETELY stupid is that OpenAI also believes this they just think the best case solution is that th…
↗Not briefed
bluesky
bluesky-ai
2moGood-faith dialogue with people that use Claude for historical work… Nonsense, that’s not the dialogue we need. We need AI literacy in his…Good-faith dialogue with people that use Claude for historical work… Nonsense, that’s not the dialogue we need. We need AI literacy i…
↗Not briefed
hn
hn-ai
2moUse 1Password to sign in to websites with Claude—
↗Not briefed
bluesky
bluesky-ai
2moI don’t know how I live in a world where a company called Anthropic is saying that humans might die from their products and all that’s miss…I don’t know how I live in a world where a company called Anthropic is saying that humans might die from their products and all that’s…
↗Not briefed
bluesky
bluesky-ai
2mocommuniqué from bitdizzycommuniqué from bitdizzy
↗Not briefed
bluesky
bluesky-ai
2moThe thing you have to understand to get why they haven't all resigned is that they think they're in a MAD scenario with OpenAI where they h…The thing you have to understand to get why they haven't all resigned is that they think they're in a MAD scenario with OpenAI where t…
↗Not briefed
hn
hn-ai
2moQwen 3.8 follows GPT-5.5 Pro reasoning prefills🔧 GPT-5🔧 Qwen 3—
↗Not briefed
bluesky
bluesky-ai
2moNobody thought they were slop. AI slop is a particular thing. Podcasts didn't have a clear funding model, people had been listening to ta…Nobody thought they were slop. AI slop is a particular thing. Podcasts didn't have a clear funding model, people had been listening…
↗Not briefed
bluesky
bluesky-ai
2moWhen the first OpenAI models were released, people were arguing that it was bad because it would never be able to generate competent images…When the first OpenAI models were released, people were arguing that it was bad because it would never be able to generate competent i…
↗Not briefed
bluesky
bluesky-ai
2moJust had this conversation at work today. Every single CEO needs to read this story. If OpenAI can do this to these researchers, they can a…Just had this conversation at work today. Every single CEO needs to read this story. If OpenAI can do this to these researchers, they…
↗Not briefed
bluesky
bluesky-ai
2moA robot running fast IS NOT IMPRESSIVE, because OF COURSE you can make a machine go fast. That's the whole point of machines. To do things …A robot running fast IS NOT IMPRESSIVE, because OF COURSE you can make a machine go fast. That's the whole point of machines. To do th…
↗Not briefed
rss
techmeme
2moApple announces the iPhone 18 Pro and iPhone 18 Pro Max, with a similar design to 17 Pro and a new A20 Pro chip, in deep black, silver, glacier, and burgundy (Jay Peters/The Verge)Jay Peters / The Verge: Apple announces the iPhone 18 Pro and iPhone 18 Pro Max, with a similar design to 17 Pro and a new A20 Pro chi…
↗Not briefed
bluesky
bluesky-ai
2moIt’s curious how they never bring up NFTs, which I think LLMs are most like. Also, all the NFT bros quickly became LLM bros.It’s curious how they never bring up NFTs, which I think LLMs are most like. Also, all the NFT bros quickly became LLM bros.
↗Not briefed
bluesky
bluesky-ai
2moNobody actually believes a LLM is going to become a machine god and end humanity. Its a stupid marketing ploy to dupe greedy gullible rich …Nobody actually believes a LLM is going to become a machine god and end humanity. Its a stupid marketing ploy to dupe greedy gullible…
↗Not briefed
bluesky
bluesky-ai
2moAi fascists: ok we paused model development to comply with the law Ai fascists continued: oh em gee bro the models are developing themselv…Ai fascists: ok we paused model development to comply with the law Ai fascists continued: oh em gee bro the models are developing the…
↗Not briefed
bluesky
bluesky-ai
2moDeepMind’s new genome ‘atlas’ charts effects of all nine billion human gene mutations www.nature.com/articles/d41...DeepMind’s new genome ‘atlas’ charts effects of all nine billion human gene mutations www.nature.com/articles/d41...
↗Not briefed
arxiv
arxiv-ai
2moDeep Learning-Based Detection of Electrical Faults and Power Quality Disturbances in Aerospace Power SystemsMore Electric Aircraft require fast and reliable monitoring of high-frequency electrical networks, yet most power quality disturbance…
↗Not briefed
bluesky
bluesky-ai
2moit is absolutely, legitimately a good-faith belief on almost everyone's part. it's why Amodei quit OpenAI!it is absolutely, legitimately a good-faith belief on almost everyone's part. it's why Amodei quit OpenAI!
↗Not briefed
bluesky
bluesky-ai
2moOpenAI names top US AI safety official to foundation board ft.trib.al/uDFXT3FOpenAI names top US AI safety official to foundation board ft.trib.al/uDFXT3F
↗Not briefed
bluesky
bluesky-ai
2moBut idk like yeah let’s just take it seriously. One of the people responsible for “safety” at Anthropic says that the chance Claude kills e…But idk like yeah let’s just take it seriously. One of the people responsible for “safety” at Anthropic says that the chance Claude ki…
↗Not briefed
bluesky
bluesky-ai
2mo"In October, over 600 current and former employees at OpenAI were able to sell their shares for a total of $6.6 billion. Around 75 of them …"In October, over 600 current and former employees at OpenAI were able to sell their shares for a total of $6.6 billion. Around 75 of…
↗Not briefed
bluesky
bluesky-ai
2moThey’re losing crucial synapses every day bc they’re cerebrum-deep in LLM usage.They’re losing crucial synapses every day bc they’re cerebrum-deep in LLM usage.
↗Not briefed
bluesky
bluesky-ai
2moFolks - if you want to use AI privately, and secure - DuckDuckGo has the goods: duck.ai/chat?t=iphoneFolks - if you want to use AI privately, and secure - DuckDuckGo has the goods: duck.ai/chat?t=iphone
↗Not briefed
hn
hn-ai
2moProcedural Graphs: Self-Evolving Execution Structures for LLM Agents—
↗Not briefed
bluesky
bluesky-ai
2moRegarding the Anthropic "we're all doomed" story: if a safety advisor at a car company said there's a 1-in-10 chance our product will kill …Regarding the Anthropic "we're all doomed" story: if a safety advisor at a car company said there's a 1-in-10 chance our product will…
↗Not briefed
hn
hn-ai
2moIs OpenAI Taking Everyone for Fools?—
↗Not briefed
bluesky
bluesky-ai
2moThe fact that LLM output is (and is regarded as) less reliable than wikipedia does not mean that people have forgotten that wikipedia gener…The fact that LLM output is (and is regarded as) less reliable than wikipedia does not mean that people have forgotten that wikipedia…
↗Not briefed
bluesky
bluesky-ai
2moSo, alignment science lead for Anthropic Evan Hubinger said today they ‘ earnestly believe AI could kill all humans!’ with ‘>10% within the…So, alignment science lead for Anthropic Evan Hubinger said today they ‘ earnestly believe AI could kill all humans!’ with ‘>10% withi…
↗Not briefed
arxiv
arxiv-ai
2moSemigroup-JEPA: Latent Dynamics Consistency for Zero-Shot Physics GeneralizationJoint-Embedding Predictive Architecture (JEPA) world models learn a compact latent representation of the world that supports predictio…
↗Not briefed
bluesky
bluesky-ai
2moThis is bad! These people are insane cultists lmao. prospect.org/2026/09/09/a...This is bad! These people are insane cultists lmao. prospect.org/2026/09/09/a...
↗Not briefed
rss
theverge-ai
2moMicrosoft has new AI privacy rules for schoolsMicrosoft agreed to a set of safety and privacy principles for AI in schools a week after two major school systems announced a ban on…
↗Not briefed
bluesky
bluesky-ai
2mowhat if anthropic people were full of shit on that and just decided they could make more money launching a competitor to altman?what if anthropic people were full of shit on that and just decided they could make more money launching a competitor to altman?
↗Not briefed
bluesky
bluesky-ai
2moWhat did people expect the *head of alignment at Anthropic* to believe about AI risk?What did people expect the *head of alignment at Anthropic* to believe about AI risk?
↗Not briefed
hn
hn-ai
2moShow HN: Rdltr – Inbox zero for your reading list🔧 ChatGPTHi HN, for the past few months I’ve been working on a read later app called RDLTR (<a href="https:&#x2F;&#x2F;rdltr.app" rel="nofollow…
↗Not briefed
bluesky
bluesky-ai
2moThe company was founded because they felt OpenAI was betraying its promise, now someone feels that the promise is being betrayed yet againThe company was founded because they felt OpenAI was betraying its promise, now someone feels that the promise is being betrayed yet a…
↗Not briefed