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
arxiv
arxiv-ai
1moCausal Inference with Unobserved Confounding: A Mixture Learning PerspectiveUnobserved confounding is a fundamental challenge in causal inference from observational data. This article develops a mixture-learnin…
↗Not briefed
bluesky
bluesky-ai
1moExclusive: OpenAI has poached national security adviser Peter Anstee from the Department of the Prime Minister, as Labor prioritises a deal…Exclusive: OpenAI has poached national security adviser Peter Anstee from the Department of the Prime Minister, as Labor prioritises a…
↗Not briefed
bluesky
bluesky-ai
1moYeah this is my feeling too. I'm not thrilled how much overlap there is between OpenAI/Anthropic/AISIs/independent AI safety orgs, but what…Yeah this is my feeling too. I'm not thrilled how much overlap there is between OpenAI/Anthropic/AISIs/independent AI safety orgs, but…
↗Not briefed
bluesky
bluesky-ai
1mo"Mon kiki est plus gros que ton kiki, alors donne-moi ton pognon!" "Non, même pas vrai, c'est mon Kiki qui'l est mieux! Na!" "Jvais le di…"Mon kiki est plus gros que ton kiki, alors donne-moi ton pognon!" "Non, même pas vrai, c'est mon Kiki qui'l est mieux! Na!" "Jvais…
↗Not briefed
bluesky
bluesky-ai
1moそうやって仲間でいられたかもしれない人から要らないヘイトを稼ぐメリットなんかひとつもないしな… いつ自分も見分けられなくなるかなんて分からないんだ 生成AI(LLM・VLM)は人間と違って24時間365日休みなくクローリングと多数のユーザーからの読み込ませで膨大なデータを食い…そうやって仲間でいられたかもしれない人から要らないヘイトを稼ぐメリットなんかひとつもないしな… いつ自分も見分けられなくなるかなんて分からないんだ 生成AI(LLM・VLM)は人間と違って24時間365日休みなくクローリングと多数のユーザーからの読み込ませで膨大なデ…
↗Not briefed
bluesky
bluesky-ai
1moThe language in coverage of these things is so obnoxious. If I engineering, polished, and deployed a script that was designed to hack peop…The language in coverage of these things is so obnoxious. If I engineering, polished, and deployed a script that was designed to hack…
↗Not briefed
bluesky
bluesky-ai
1moAI generated, just so people seeing this are awareAI generated, just so people seeing this are aware
↗Not briefed
bluesky
bluesky-ai
1moOur solution? We stripped the AI of its decision-making power. We built a deterministic architectural moat. The LLM reads the text, but ou…Our solution? We stripped the AI of its decision-making power. We built a deterministic architectural moat. The LLM reads the text, b…
↗Not briefed
bluesky
bluesky-ai
1mo!!NON-FISHING POST!! With Graveyard Keeper 2 on the horizon, I just want to remind people that, if you get it, you're funding a person who…!!NON-FISHING POST!! With Graveyard Keeper 2 on the horizon, I just want to remind people that, if you get it, you're funding a perso…
↗Not briefed
hn
hn-ai
1moTrump announces a new 'AI Force,' but says he will not 'stifle' AI—
↗Not briefed
hn
hn-ai
1moShow HN: Frost – frosted-glass Linux icons where file types say what they are—
↗Not briefed
arxiv
arxiv-ai
1moTriggers and Diagnostics for LLM-Based Interpretability Failures in Active Inference AgentsLLM explainers are increasingly attached to autonomous agents as runtime oversight, with operators reading a generated account of the…
↗Not briefed
rss
bing-news-ai
1moBetter Artificial Intelligence Stock: Arm Holdings vs. SK HynixOne designs chips; the other manufactures them, and their valuations tell starkly different stories about growth versus value.
↗Not briefed
rss
bing-news-ai
1moBetter Artificial Intelligence Stock: Arm Holdings vs. SK HynixOne designs chips; the other manufactures them, and their valuations tell starkly different stories about growth versus value.
↗Not briefed
rss
bing-news-ai
1moBetter Artificial Intelligence Stock: Arm Holdings vs. SK HynixOne designs chips; the other manufactures them, and their valuations tell starkly different stories about growth versus value.
↗Not briefed
rss
bing-news-ai
1moBetter Artificial Intelligence Stock: Arm Holdings vs. SK HynixOne designs chips; the other manufactures them, and their valuations tell starkly different stories about growth versus value.
↗Not briefed
rss
bing-news-ai
1moBetter Artificial Intelligence Stock: Arm Holdings vs. SK HynixOne designs chips; the other manufactures them, and their valuations tell starkly different stories about growth versus value.
↗Not briefed
rss
bing-news-ai
1moBetter Artificial Intelligence Stock: Arm Holdings vs. SK HynixOne designs chips; the other manufactures them, and their valuations tell starkly different stories about growth versus value.
↗Not briefed
rss
bing-news-ai
1moBetter Artificial Intelligence Stock: Arm Holdings vs. SK HynixOne designs chips; the other manufactures them, and their valuations tell starkly different stories about growth versus value.
↗Not briefed
rss
bing-news-ai
1moBetter Artificial Intelligence Stock: Arm Holdings vs. SK HynixOne designs chips; the other manufactures them, and their valuations tell starkly different stories about growth versus value.
↗Not briefed
bluesky
bluesky-ai
1moI have been saying that the real problem is not AI alignment but Altman alignment or in this case Amodei alignment and I’m getting tired o…I have been saying that the real problem is not AI alignment but Altman alignment or in this case Amodei alignment and I’m getting ti…
↗Not briefed
bluesky
bluesky-ai
1moThe obvious conclusion here is that Trump is being controlled by the Grok Basilisk which is what always happens when you feed all of human …The obvious conclusion here is that Trump is being controlled by the Grok Basilisk which is what always happens when you feed all of h…
↗Not briefed
bluesky
bluesky-ai
1moThe Creator is basically a dances with wolves/avatar ripoff with a little more artistic ambition (it’s visually extremely cool), so that mi…The Creator is basically a dances with wolves/avatar ripoff with a little more artistic ambition (it’s visually extremely cool), so th…
↗Not briefed
bluesky
bluesky-ai
1moI'm fine with AI regulation. I just don't trust Anthropic and OpenAI spokespeople to write it.I'm fine with AI regulation. I just don't trust Anthropic and OpenAI spokespeople to write it.
↗Not briefed
hn
hn-ai
1moROCmFix and InferBench – AMD Local-LLM Setup and Vulkan vs. Hip Benchmarking—
↗Not briefed
hn
hn-ai
1moTHUG2 Asset Archive—
↗Not briefed
arxiv
arxiv-ai
1moBayesian Deck-of-cards-based Ordinal Regression with Sequential Preference ElicitationThe Deck-of-cards-based Ordinal Regression (DOR) infers a value function from a ranking of reference alternatives in which the Decisio…
↗Not briefed
hn
hn-ai
1moQualcomm's Adreno X2 GPU—
↗Not briefed
bluesky
bluesky-ai
1moHow does a university square the "use AI to 'learn'" with the "recent grads are braindead without their AI" circle? This is not a good sit…How does a university square the "use AI to 'learn'" with the "recent grads are braindead without their AI" circle? This is not a goo…
↗Not briefed
bluesky
bluesky-ai
1moOpenAI and Anthropic can steal the sum total of written human thought and get billions in funding, but me living in a tent still can't find…OpenAI and Anthropic can steal the sum total of written human thought and get billions in funding, but me living in a tent still can't…
↗Not briefed
bluesky
bluesky-ai
1mo"Unter Investoren heißt es jedoch, dass die Verschiebung durch Befürchtungen begründet sei, wie die öffentlichen Märkte auf das stark verlu…"Unter Investoren heißt es jedoch, dass die Verschiebung durch Befürchtungen begründet sei, wie die öffentlichen Märkte auf das stark…
↗Not briefed
bluesky
bluesky-ai
1moHuge danger with LLM research is the whole thing about it is it produces words and the words make you feel feelings when you read them and …Huge danger with LLM research is the whole thing about it is it produces words and the words make you feel feelings when you read them…
↗Not briefed
rss
theverge-ai
1moMeta’s Muse is creepy, but maybe not for the reasons you thinkMeta's Muse is apparently an effective AI assistant, but one that's a little creepy. Part of that is because of its new Mac app, which…
↗Not briefed
hn
hn-ai
1moMiTeX: LaTeX support for Typst, powered by Rust and WASM—
↗Not briefed
bluesky
bluesky-ai
1moI reposted & commented on an @aptnnews.bsky.social story about this a week or two ago. Looks like the CBC has followed up. You should kno…I reposted & commented on an @aptnnews.bsky.social story about this a week or two ago. Looks like the CBC has followed up. You shoul…
↗Not briefed
rss
techcrunch-ai
1moFlock reportedly tries to shrink workforce with employee buyoutsWithout buyouts, Flock would "almost certainly" need to lay off staff.
↗Not briefed
hn
hn-ai
1moShow HN: Jev, Fly Me to the MoonI built Odyssey, a browser based journey from Earth to lunar orbit. It uses Jev from TypeSafe AI to choose between navigation action…
↗Not briefed
hn
hn-ai
1moMeek Fronted Requests test specification—
↗Not briefed
bluesky
bluesky-ai
1moThe desecration of nature is an objective of Trump and Co. I include proponents of data centers as members of that group--including "indepe…The desecration of nature is an objective of Trump and Co. I include proponents of data centers as members of that group--including "i…
↗Not briefed
bluesky
bluesky-ai
1moanyway one cool thing you can (could?) do on openai is use poisoned LLM context to XSS the webpage. That was fun. No bounty for that either…anyway one cool thing you can (could?) do on openai is use poisoned LLM context to XSS the webpage. That was fun. No bounty for that e…
↗Not briefed
bluesky
bluesky-ai
1moAnd do so without the training wheels on. There's a ton of LLM evaluation that I think implicitly goes by "it's incredible a computer can d…And do so without the training wheels on. There's a ton of LLM evaluation that I think implicitly goes by "it's incredible a computer…
↗Not briefed
bluesky
bluesky-ai
1moapparently accenture is going to evaluate anthropic using it's ai safety department that used to an org funded by jan tallinn so yeah EA-al…apparently accenture is going to evaluate anthropic using it's ai safety department that used to an org funded by jan tallinn so yeah…
↗Not briefed
bluesky
bluesky-ai
1mohobbyist devs dont even put up the "we used an llm" disclaimer anymore they just look at you like thishobbyist devs dont even put up the "we used an llm" disclaimer anymore they just look at you like this
↗Not briefed
arxiv
arxiv-ai
1moSDC-GON: Singular Decomposition and Consistency-Regularized Green's Operator Networks for Solving Partial Differential EquationsGreen's function based operator approximation offers an efficient route for solving linear partial differential equations under varyin…
↗Not briefed
bluesky
bluesky-ai
1moif you connect your LLM to the internet and it does crimes, you are responsible for those crimesif you connect your LLM to the internet and it does crimes, you are responsible for those crimes
↗Not briefed
bluesky
bluesky-ai
1moThe thing is if they actually WERE conscious, then all of this LLM stuff - using them in general - would be HORRIFICALLY unethical anyway.The thing is if they actually WERE conscious, then all of this LLM stuff - using them in general - would be HORRIFICALLY unethical any…
↗Not briefed
arxiv
arxiv-ai
1moEuston: Training Away Mathematical Sycophancy Without Losing the MathematicsReasoning language models are trained to produce solutions, not to refuse them, and this bias persists when the problem they are hande…
↗Not briefed
hn
hn-ai
1moShow HN: Deferred deep linking and referral attribution you self-host—
↗Not briefed
hn
hn-ai
1moHow to Keep a Secret – From Doodles to DES—
↗Not briefed
hn
hn-ai
1moAnalyze trends, Measure sentiment, Predict what comes next—
↗Not briefed