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
2moData quality, not architecture, drove 12 times the compute-efficiency gains in AI pretraining from 2019 to 2025, a new study findsData quality, not architecture, drove 12 times the compute-efficiency gains in AI pretraining from 2019 to 2025, a new study finds
↗Not briefed
bluesky
bluesky-ai
2moI'd love to contribute but I'm having issues with the latest Falcon LLM model, can't get any responses.I'd love to contribute but I'm having issues with the latest Falcon LLM model, can't get any responses.
↗Not briefed
bluesky
bluesky-ai
2moThe openai full court press just leading to people who have had a sour experience about Sebastian coming out and saying "ya no, he does thi…The openai full court press just leading to people who have had a sour experience about Sebastian coming out and saying "ya no, he doe…
↗Not briefed
arxiv
arxiv-ai
2moAdaptive Distributed Physical-Layer Authentication and Attack Detection in 6G Non-Terrestrial Networks via Causal Meta-LearningPhysical-layer authentication (PLA) in non-terrestrial networks (NTNs) is challenged by severe Doppler shifts, long delays, and fast c…
↗Not briefed
bluesky
bluesky-ai
2moAmerican Mathematical Society rewriting history: "story began with Navier, Stokes, Leray, and Ladyzhenskaya and has culminated in the rece…American Mathematical Society rewriting history: "story began with Navier, Stokes, Leray, and Ladyzhenskaya and has culminated in the…
↗Not briefed
bluesky
bluesky-ai
2moi ran the timeline through chat—consensus is oai may have started work in response to the initial Buckmaster email, but unlikely they clea…🔧 Consensusi ran the timeline through chat—consensus is oai may have started work in response to the initial Buckmaster email, but unlikely they…
↗Not briefed
hn
hn-ai
2moSauron: A free and open source terminal development environment—
↗Not briefed
bluesky
bluesky-ai
2moadded mathjax rendering through the kitty graphics protocol to my LLM CLI in 5 minutes. wild github.com/david-crespo...added mathjax rendering through the kitty graphics protocol to my LLM CLI in 5 minutes. wild github.com/david-crespo...
↗Not briefed
hn
hn-ai
2moShow HN: Auto-Rig Web – Zero-click 3D humanoid rigging in a Web Worker—
↗Not briefed
bluesky
bluesky-ai
2moOpenAI is basically an entire company dedicated to the idea that real products would be better if they were scams.OpenAI is basically an entire company dedicated to the idea that real products would be better if they were scams.
↗Not briefed
bluesky
bluesky-ai
2mo"in which language is an LLM intelligent" seems like a really poor gotcha but a really interesting empirical question about relative cross-…"in which language is an LLM intelligent" seems like a really poor gotcha but a really interesting empirical question about relative c…
↗Not briefed
bluesky
bluesky-ai
2moI'm sure others have brought this up but it seems to beggar belief a little that OpenAI claims they solved the problem with minimal human i…I'm sure others have brought this up but it seems to beggar belief a little that OpenAI claims they solved the problem with minimal hu…
↗Not briefed
bluesky
bluesky-ai
2moClaude用11天的时间将威尔斯对费马大定律的证明形式化,写了1300万行代码,并验证了威尔斯的证明100%正确。如果人来做这事,要花10年时间。以后数学证明工作的确可以由AI部分取代。Claude用11天的时间将威尔斯对费马大定律的证明形式化,写了1300万行代码,并验证了威尔斯的证明100%正确。如果人来做这事,要花10年时间。以后数学证明工作的确可以由AI部分取代。
↗Not briefed
rss
simon-willison
2moIntroducing ChatGPT Images 2.5🔧 ChatGPT—
↗Not briefed
bluesky
bluesky-ai
2mo【GPT-Image2.5にご用心を】 OpenAIの画像生成AIがバージョンアップした模様です。 ・肖像権・著作権を含んだ大量の画像無断使用 ・違法コピー、海賊版使用 ・有料フォント無断使用 これらのリスクを抱えていますので、画像生成AI全般の使用はお止めください。【GPT-Image2.5にご用心を】 OpenAIの画像生成AIがバージョンアップした模様です。 ・肖像権・著作権を含んだ大量の画像無断使用 ・違法コピー、海賊版使用 ・有料フォント無断使用 これらのリスクを抱えていますので、画像生成AI全般の使用はお止めください。
↗Not briefed
bluesky
bluesky-ai
2moThis implies the existence of a Yugoslavian non-aligned LLM leading an alternative road to Socialism with Adriatic beaches and cafe culture…This implies the existence of a Yugoslavian non-aligned LLM leading an alternative road to Socialism with Adriatic beaches and cafe cu…
↗Not briefed
bluesky
bluesky-ai
2moopenai managed to recruit Priority Disputes Georg to quarterback navier-stokes. amazingopenai managed to recruit Priority Disputes Georg to quarterback navier-stokes. amazing
↗Not briefed
rss
techmeme
2moQ&A with Mark Zuckerberg on Muse, the data center backlash, Meta's recent youth-safety settlement, privacy concerns about Meta's smart glasses, and more (Alex Heath/Sources)Alex Heath / Sources: Q&A with Mark Zuckerberg on Muse, the data center backlash, Meta's recent youth-safety settlement, privacy conce…
↗Not briefed
arxiv
arxiv-ai
2moInductive Biases in Field-Level Cosmological Inference from Galaxy CatalogsWe perform field-level likelihood-free inference of the matter density parameter $Ω_m$ from simulated galaxy catalogs using machine le…
↗Not briefed
bluesky
bluesky-ai
2moI don’t know of any significant AI LLM model that is not, at its core, based on stolen IP or unethically-acquired IP. If you know of one, …I don’t know of any significant AI LLM model that is not, at its core, based on stolen IP or unethically-acquired IP. If you know of…
↗Not briefed
bluesky
bluesky-ai
2moDear OpenAI, I’m about to publish a proof of this break, coauthored with anthropic. Do with this information what you willDear OpenAI, I’m about to publish a proof of this break, coauthored with anthropic. Do with this information what you will
↗Not briefed
bluesky
bluesky-ai
2moSometimes I miss the "investigating research misconduct" part of my former job. Others I'm grateful I got out before the field turned into …Sometimes I miss the "investigating research misconduct" part of my former job. Others I'm grateful I got out before the field turned…
↗Not briefed
bluesky
bluesky-ai
2moOpenAI announces gyre breakthrough. Axial stretching lets the centre reach unbounded speed on finite energy. Things fall apart, as required…OpenAI announces gyre breakthrough. Axial stretching lets the centre reach unbounded speed on finite energy. Things fall apart, as req…
↗Not briefed
hn
hn-ai
2moShow HN: Spatial memory for document OCR pipelines (zero dependencies)—
↗Not briefed
bluesky
bluesky-ai
2moOpenAI was able to turn a mostly solved million dollar problem and several million dollars into a solved million dollar problem and you're …OpenAI was able to turn a mostly solved million dollar problem and several million dollars into a solved million dollar problem and yo…
↗Not briefed
hn
hn-ai
2moStop making swap partitions–use swap files instead—
↗Not briefed
bluesky
bluesky-ai
2mo"yeah so ai makes child porn" "deeply unserious if you think LLMs do that" sorry but if you actually dont have the intellect to think tha…"yeah so ai makes child porn" "deeply unserious if you think LLMs do that" sorry but if you actually dont have the intellect to thin…
↗Not briefed
bluesky
bluesky-ai
2moAcho que não sabe ler. Ela se referiu diretamente ao chat gpt. Uma LLM. Inteligencia artificial é usada a anos e não tem nada a ver com o l…Acho que não sabe ler. Ela se referiu diretamente ao chat gpt. Uma LLM. Inteligencia artificial é usada a anos e não tem nada a ver co…
↗Not briefed
bluesky
bluesky-ai
2mo1. Researchers outside of OpenAI enter their research into an OpenAI "AI"-assisted product. 2. OpenAI publishes research seemingly based o…1. Researchers outside of OpenAI enter their research into an OpenAI "AI"-assisted product. 2. OpenAI publishes research seemingly ba…
↗Not briefed
bluesky
bluesky-ai
2moI think it’s pretty obvious they injected some artificial time pressure for themselves here - because they knew “OpenAI solves the problem …I think it’s pretty obvious they injected some artificial time pressure for themselves here - because they knew “OpenAI solves the pro…
↗Not briefed
bluesky
bluesky-ai
2moThere's a personal note in here about what the Anthropic mess feels like right now for us authors. A commenter read it and wrote: "Thank …There's a personal note in here about what the Anthropic mess feels like right now for us authors. A commenter read it and wrote: "T…
↗Not briefed
bluesky
bluesky-ai
2moOne of the academics came pretty damn close to directly accusing them, it's remarkable and entirely believableOne of the academics came pretty damn close to directly accusing them, it's remarkable and entirely believable
↗Not briefed
arxiv
arxiv-ai
2moThe Mutations of Machine SpeechAlgorithmic outputs now populate the digital environments through which contemporary life is organized. The role of law in facilitatin…
↗Not briefed
bluesky
bluesky-ai
2moYou know how you can coordinate 10,000+ AI agents? Answer: Message Board OpenAI’s internal model group arrived at the Navier–Stokes solut…You know how you can coordinate 10,000+ AI agents? Answer: Message Board OpenAI’s internal model group arrived at the Navier–Stokes…
↗Not briefed
bluesky
bluesky-ai
2moSebastien Bubeck responds to allegations specifically, i think this greatly clarifies the, “why would you ruin your career over this?” qu…Sebastien Bubeck responds to allegations specifically, i think this greatly clarifies the, “why would you ruin your career over this…
↗Not briefed
bluesky
bluesky-ai
2moWait until future generations find out that this creepy Sam Altman biopic, the creepy Elon Musk documentary, and the creepy Elizabeth Holme…Wait until future generations find out that this creepy Sam Altman biopic, the creepy Elon Musk documentary, and the creepy Elizabeth…
↗Not briefed
bluesky
bluesky-ai
2mo🤔 money laundering. besides the obvious financial crimes keeping them running, aren't LLM services using the same kind of money-launderi…🤔 money laundering. besides the obvious financial crimes keeping them running, aren't LLM services using the same kind of money-lau…
↗Not briefed
bluesky
bluesky-ai
2moAnti AI people metastacized "your model was trained on data used without consent" to "a pro-AI person does not care about consent in genera…Anti AI people metastacized "your model was trained on data used without consent" to "a pro-AI person does not care about consent in g…
↗Not briefed
bluesky
bluesky-ai
2moa very anti-social but also pretty funny thing you could do with this website given how open it is would be to make a bot that 'clones' use…a very anti-social but also pretty funny thing you could do with this website given how open it is would be to make a bot that 'clones…
↗Not briefed
bluesky
bluesky-ai
2moOpenAI likely commited the most shameless theft today and there is nothing indicating the exposure wil stop them. We are all lab hamsters f…OpenAI likely commited the most shameless theft today and there is nothing indicating the exposure wil stop them. We are all lab hamst…
↗Not briefed
hn
hn-ai
2moShow HN: SagaShield – ACID transactions and security guardrails for AI agents—
↗Not briefed
bluesky
bluesky-ai
2moAnthropic invented bluetooth but goodAnthropic invented bluetooth but good
↗Not briefed
bluesky
bluesky-ai
2moQuick, spread some rumors about other really hard problems that Anthropic is on the verge of solving.Quick, spread some rumors about other really hard problems that Anthropic is on the verge of solving.
↗Not briefed
bluesky
bluesky-ai
2moNew: No more free Ransomware Readiness Assessments for you, Tiny Tim! 150M in VeriScan ID breach, Tesla Cybercab warns “not every high volt…New: No more free Ransomware Readiness Assessments for you, Tiny Tim! 150M in VeriScan ID breach, Tesla Cybercab warns “not every high…
↗Not briefed
arxiv
arxiv-ai
2moEfficient Fairness Auditing Across Guidance Scales in Text-to-Image Diffusion Models via Causal Abstraction🔧 Stable DiffusionFairness auditing of text-to-image diffusion models often requires generating large numbers of images across sampling configurations,…
↗Not briefed
bluesky
bluesky-ai
2moI'm fully in the "if a game looks like it's trying to be AAA then an llm touched it" mindset era can't trust shit anymoreI'm fully in the "if a game looks like it's trying to be AAA then an llm touched it" mindset era can't trust shit anymore
↗Not briefed
bluesky
bluesky-ai
2moEvery time OpenAI makes a claim about mathematics, it's lying. Every single time I've investigated, there's been a giant asterisk like "sto…Every time OpenAI makes a claim about mathematics, it's lying. Every single time I've investigated, there's been a giant asterisk like…
↗Not briefed
bluesky
bluesky-ai
2moreckon we could have avoided this whole LLm psychosis thing if more people had watched Pontypoolreckon we could have avoided this whole LLm psychosis thing if more people had watched Pontypool
↗Not briefed
arxiv
arxiv-ai
2moInfra-Bench CLS: A Global, Open-Source Benchmark for Critical Infrastructure Classification with Earth Observation Foundation ModelsCritical infrastructure location data is often incomplete and unevenly distributed globally, especially in developing regions. Earth o…
↗Not briefed
bluesky
bluesky-ai
2mosometimes i think i don't like video games anymore but really it's more that the video games that get all the marketing are slop. not in th…sometimes i think i don't like video games anymore but really it's more that the video games that get all the marketing are slop. not…
↗Not briefed