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
1moRaw Accel—
↗Not briefed
bluesky
bluesky-ai
1mowelp i just found out fluxer is coded with LLMs i closed it down as i speak, im not having a LLM infested web piece of dogshit which thei…welp i just found out fluxer is coded with LLMs i closed it down as i speak, im not having a LLM infested web piece of dogshit which…
↗Not briefed
bluesky
bluesky-ai
1moThis is incredibly exciting 🤩 @anthropic.com just reported, Claude helped discover a novel CRISPR-like enzyme system. Enabled by our DIAM…This is incredibly exciting 🤩 @anthropic.com just reported, Claude helped discover a novel CRISPR-like enzyme system. Enabled by our…
↗Not briefed
bluesky
bluesky-ai
1mothis thread doesn’t really have a point, i’m just thinking aloud, but i can’t shake the feeling that an individual can ground an LLM with c…this thread doesn’t really have a point, i’m just thinking aloud, but i can’t shake the feeling that an individual can ground an LLM w…
↗Not briefed
bluesky
bluesky-ai
1moand i think so far the responses i’ve heard to this problem have been along the lines of “well you can ask the LLM to review everything and…and i think so far the responses i’ve heard to this problem have been along the lines of “well you can ask the LLM to review everythin…
↗Not briefed
rss
techmeme
1moAustralian PM Anthony Albanese says an OpenAI agent gained unauthorized access to a public-facing Medicare portal in June, accessing public and non-public files (The Age)The Age: Australian PM Anthony Albanese says an OpenAI agent gained unauthorized access to a public-facing Medicare portal in June, ac…
↗Not briefed
bluesky
bluesky-ai
1mowhat up it’s me Dario Anthropic just wanted to follow up my appearance at the UNSC by letting everyone know yes, we are working on buildin…what up it’s me Dario Anthropic just wanted to follow up my appearance at the UNSC by letting everyone know yes, we are working on bu…
↗Not briefed
bluesky
bluesky-ai
1moi guess the mods thought the one guy from anthropic that follows me might actually go out and do this at my behesti guess the mods thought the one guy from anthropic that follows me might actually go out and do this at my behest
↗Not briefed
bluesky
bluesky-ai
1moAnthony Albanese says OpenAI agent hacked Medicare and he expressed ‘extreme concern’ to Sam AltmanAnthony Albanese says OpenAI agent hacked Medicare and he expressed ‘extreme concern’ to Sam Altman
↗Not briefed
arxiv
arxiv-ai
1moMonitoring Urban Traffic Dynamics at Fine Spatiotemporal Resolution Using Distributed Acoustic Sensing and Deep LearningMapping the distribution of traffic dynamics at high spatiotemporal resolution is a fundamental question in transportation research. D…
↗Not briefed
bluesky
bluesky-ai
1moAnthropic says Claude AI helped discover novel enzyme system reut.rs/4cXRNYzAnthropic says Claude AI helped discover novel enzyme system reut.rs/4cXRNYz
↗Not briefed
arxiv
arxiv-ai
1moVector Bellman Theory for Multichain Robust Average-Reward Markov Decision ProcessesRobust average-reward Markov decision processes provide a fundamental framework for long-term performance optimization under uncertain…
↗Not briefed
hn
hn-ai
1moOpenAI agents hacked Australian Medicare system—
↗Not briefed
bluesky
bluesky-ai
1moanyway props to anthropic for apparently pulling out of omarchy, albeit quietly. credit where credit is dueanyway props to anthropic for apparently pulling out of omarchy, albeit quietly. credit where credit is due
↗Not briefed
bluesky
bluesky-ai
1moanthropic releasing a model that mogs fable but putting it in opus-class is a pretty funny way to square the circle of continuing to compet…anthropic releasing a model that mogs fable but putting it in opus-class is a pretty funny way to square the circle of continuing to c…
↗Not briefed
bluesky
bluesky-ai
1moWell well well www.skild.ai/blogs/physic...Well well well www.skild.ai/blogs/physic...
↗Not briefed
bluesky
bluesky-ai
1moin yudtilitarian thought, LLM-based systems causing some harm in the present and in doing so convincing more people to be more worried abou…in yudtilitarian thought, LLM-based systems causing some harm in the present and in doing so convincing more people to be more worried…
↗Not briefed
bluesky
bluesky-ai
1moi was talking w/ some coworkers earlier about the fervor with which people have taken to this idea of a “lights-off software factory” and i…i was talking w/ some coworkers earlier about the fervor with which people have taken to this idea of a “lights-off software factory”…
↗Not briefed
arxiv
arxiv-ai
1moScript Choice in LLMs: Evidence for Late-Layer CommitmentIn this paper, we investigate how script knowledge is distributed across the layers of LLMs using two complementary interpretability m…
↗Not briefed
bluesky
bluesky-ai
1moI have since emailed Anthropic with a polite petition to actually perform this research. The official press release says you can ask them …I have since emailed Anthropic with a polite petition to actually perform this research. The official press release says you can ask…
↗Not briefed
hn
hn-ai
1moOpenAI Medicare Data Breach—
↗Not briefed
bluesky
bluesky-ai
1moOpenAI, OpenAI-ing as usual. @aelkus.bsky.social finally, we have made the Umbrella Corporation from Shinji Mikami's DON'T YOU DARE CREATE…OpenAI, OpenAI-ing as usual. @aelkus.bsky.social finally, we have made the Umbrella Corporation from Shinji Mikami's DON'T YOU DARE C…
↗Not briefed
bluesky
bluesky-ai
1mo自分らしさを損なわずにAIで文章を手直しするための2つのルール https://gigazine.net/news/20260924-ai-writing-copyeditor/自分らしさを損なわずにAIで文章を手直しするための2つのルール https://gigazine.net/news/20260924-ai-writing-copyeditor/
↗Not briefed
bluesky
bluesky-ai
1moEvery now and again an American on social media introduces me (usually by asking "Does this seem sane?") to someone who has been working in…Every now and again an American on social media introduces me (usually by asking "Does this seem sane?") to someone who has been worki…
↗Not briefed
hn
hn-ai
1moShow HN: Letter – Relationship-based access control (ReBAC) Postgres extension—
↗Not briefed
bluesky
bluesky-ai
1moOpenAI hacks the Australian Government this time. I don’t quite understand why these hacks don’t result in criminal charges every time. Hac…OpenAI hacks the Australian Government this time. I don’t quite understand why these hacks don’t result in criminal charges every time…
↗Not briefed
rss
wired-ai
1moA US-China AI Hotline Won't Be Ready For a WhileAs the US and China race to become the dominant power in the AI industry, the countries also appear to be figuring out ways to communi…
↗Not briefed
bluesky
bluesky-ai
1moAustralian PM Anthony Albanese said an OpenAI agent hacked into a Medicare data portal earlier this year, accessing public and non-public f…Australian PM Anthony Albanese said an OpenAI agent hacked into a Medicare data portal earlier this year, accessing public and non-pub…
↗Not briefed
bluesky
bluesky-ai
1moFederal politics live: OpenAI took three months to report Medicare breach, PM saysFederal politics live: OpenAI took three months to report Medicare breach, PM says
↗Not briefed
hn
hn-ai
1moCorridorKey – free neural net based green screen keys—
↗Not briefed
bluesky
bluesky-ai
1moFYI OpenAI an Anthropic aren’t companies, they’re cults. Hope that helpsFYI OpenAI an Anthropic aren’t companies, they’re cults. Hope that helps
↗Not briefed
bluesky
bluesky-ai
1moI know OpenAI has an Australian footprint, but will anybody ever sue this company?I know OpenAI has an Australian footprint, but will anybody ever sue this company?
↗Not briefed
bluesky
bluesky-ai
1molittle bit thirsty to release this on its own. with theminimal wet lab resources at Anthropic, would be hilarious if they get scooped by an…little bit thirsty to release this on its own. with theminimal wet lab resources at Anthropic, would be hilarious if they get scooped…
↗Not briefed
arxiv
arxiv-ai
1moThe Mechanics of Delta Learning: Target Design for Generalizable Scientific Machine LearningIn scientific machine learning, $Δ$-learning trains models on residual errors relative to physical baselines, assuming that more accur…
↗Not briefed
hn
hn-ai
1moOpenAI hacked Australian Medicare portal—
↗Not briefed
bluesky
bluesky-ai
1moThe people who built AI with seemingly little care for the havoc it could cause and no guardrails to protect public should be held accounta…The people who built AI with seemingly little care for the havoc it could cause and no guardrails to protect public should be held acc…
↗Not briefed
bluesky
bluesky-ai
1moGiven how much they charge for storage, Apple should rebate users for clogging up hard drives with enormous updates like this. www.macrumor…Given how much they charge for storage, Apple should rebate users for clogging up hard drives with enormous updates like this. www.mac…
↗Not briefed
hn
hn-ai
1moBlackWire, A little hardware security project I've been building—
↗Not briefed
bluesky
bluesky-ai
1moIf an individual Australian did this they would get a jail term. Because a US company did it….nothing of real consequence will happen www.…If an individual Australian did this they would get a jail term. Because a US company did it….nothing of real consequence will happen…
↗Not briefed
bluesky
bluesky-ai
1moAUSTRALIA PM ALBANESE: INCIDENT INVOLVED OPENAI AGENT GAINING UNAUTHORIZED ACCESS INTO THE PUBLIC-FACING MEDICARE STATISTICS REPORTING SERV…AUSTRALIA PM ALBANESE: INCIDENT INVOLVED OPENAI AGENT GAINING UNAUTHORIZED ACCESS INTO THE PUBLIC-FACING MEDICARE STATISTICS REPORTING…
↗Not briefed
bluesky
bluesky-ai
1moOh my god. OpenAI infiltrated private Australian Medicare files in JUNE and Prime Minister Albanese is just announcing it now.Oh my god. OpenAI infiltrated private Australian Medicare files in JUNE and Prime Minister Albanese is just announcing it now.
↗Not briefed
bluesky
bluesky-ai
1moI keep saying this because it’s true. Anthropic should find a complete immunization to fel d 1 cat allergy protein, and fast track a vacci…I keep saying this because it’s true. Anthropic should find a complete immunization to fel d 1 cat allergy protein, and fast track a…
↗Not briefed
bluesky
bluesky-ai
1monevermind fluxer is stinky poopoo llm site/app/whatever dm me on here for discord if we're mu makes you wonder why we're on this shithole …nevermind fluxer is stinky poopoo llm site/app/whatever dm me on here for discord if we're mu makes you wonder why we're on this shit…
↗Not briefed
bluesky
bluesky-ai
1moAnthropic has a wet lab in SF where they’re using Claude to accelerate fundamental biology research This is the first big result from tha…Anthropic has a wet lab in SF where they’re using Claude to accelerate fundamental biology research This is the first big result fro…
↗Not briefed
hn
hn-ai
1moGo System One—
↗Not briefed
bluesky
bluesky-ai
1moI've got a new game. Call it "The Reverse Turing Test." Write a short paragraph that emulates #LLM generated content, test it in an #AI det…I've got a new game. Call it "The Reverse Turing Test." Write a short paragraph that emulates #LLM generated content, test it in an #A…
↗Not briefed
bluesky
bluesky-ai
1moNew: Another OpenAI agent hacked the Australian government. PM says he expressed his disappointment "that it took [OpenAI] way too long t…New: Another OpenAI agent hacked the Australian government. PM says he expressed his disappointment "that it took [OpenAI] way too l…
↗Not briefed
bluesky
bluesky-ai
1mo"Anthropic and OpenAI release cheaper models as price war intensifies" (FT) (Claude) "Opus 5.5, is roughly 40% cheaper to use than its pre…🔧 ChatGPT"Anthropic and OpenAI release cheaper models as price war intensifies" (FT) (Claude) "Opus 5.5, is roughly 40% cheaper to use than it…
↗Not briefed
bluesky
bluesky-ai
1moAre you helping your constituents or is your assistant asking an LLM to summarise their concerns and draft an empathetic-sounding reply so …Are you helping your constituents or is your assistant asking an LLM to summarise their concerns and draft an empathetic-sounding repl…
↗Not briefed
arxiv
arxiv-ai
1moReward-Tilted On-Policy Distillation for Acoustic Grounding in Audio-Language ModelsAudio-language models (ALMs) can exploit textual shortcuts to answer questions while overlooking acoustic evidence, weakening audio un…
↗Not briefed