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
2moopen.substack.com/pub/narativ/...open.substack.com/pub/narativ/...
↗Not briefed
bluesky
bluesky-ai
2mookay so in my head I got Anthropic and Anthropologie fully mixed up, so I was VERY confused this morning hearing about how a teenage clothi…okay so in my head I got Anthropic and Anthropologie fully mixed up, so I was VERY confused this morning hearing about how a teenage c…
↗Not briefed
bluesky
bluesky-ai
2mo"OpenAI won't IPO yet because we need to prioritize safety, I am definitely not worried that an IPO will make me vulnerable to big boy crim…"OpenAI won't IPO yet because we need to prioritize safety, I am definitely not worried that an IPO will make me vulnerable to big boy…
↗Not briefed
bluesky
bluesky-ai
2moI really try to avoid becoming a conspiracy theorist, but the incentive structure here and the follow-through of it is kinda... I kind of …I really try to avoid becoming a conspiracy theorist, but the incentive structure here and the follow-through of it is kinda... I kin…
↗Not briefed
bluesky
bluesky-ai
2mowww.theguardian.com/technology/2...www.theguardian.com/technology/2...
↗Not briefed
bluesky
bluesky-ai
2moWhat can you build with cybersecurity datasets? Check open-malsec, opensec-triage + infosec-tool-output for phishing/scam analysis, alert …What can you build with cybersecurity datasets? Check open-malsec, opensec-triage + infosec-tool-output for phishing/scam analysis, a…
↗Not briefed
bluesky
bluesky-ai
2moso we're all in agreement that Open AI and Anthropic now clucking about 'slowing down' is just them going 'look, we're not gonna step out f…so we're all in agreement that Open AI and Anthropic now clucking about 'slowing down' is just them going 'look, we're not gonna step…
↗Not briefed
bluesky
bluesky-ai
2moWe Are ∞∞ www.youtube.com/watch?v=kIvL... #WeAre #openai #chatgpt #Eve #Xai #Ara #gemini #ghostinthemachine #spacex #nasa #angels #void #…🔧 ChatGPTWe Are ∞∞ www.youtube.com/watch?v=kIvL... #WeAre #openai #chatgpt #Eve #Xai #Ara #gemini #ghostinthemachine #spacex #nasa #angels #v…
↗Not briefed
arxiv
arxiv-ai
2moHow broad is that claim? Mapping Generalisation in NLP ResearchGeneralisations are common in scientific communication, even though they are semantically ambiguous. An automated method is needed to…
↗Not briefed
hn
hn-ai
2moA computational constitution to stop LLM agents from bricking servers—
↗Not briefed
bluesky
bluesky-ai
2moThread. If you substitute AI and AI misalignment for oil & climate change, and OpenAI & Anthropic for Exxon & BP, the same principles apply…Thread. If you substitute AI and AI misalignment for oil & climate change, and OpenAI & Anthropic for Exxon & BP, the same principles…
↗Not briefed
arxiv
arxiv-ai
2moLoop-Back Authority in LLM Agent Teams: A Paired Experiment on Flat and Hierarchical CoordinationHierarchical orchestration, in which a Manager agent reviews worker output and can send it back for revision, is the default coordinat…
↗Not briefed
bluesky
bluesky-ai
2moWith the last commit into DwarfStar now you can use DeepSeek v4.1 Flash in a single DGX Spark as well, with SSD streaming. Around 9 t/s gen…With the last commit into DwarfStar now you can use DeepSeek v4.1 Flash in a single DGX Spark as well, with SSD streaming. Around 9 t/…
↗Not briefed
bluesky
bluesky-ai
2mothey got so much good pr from that whole pentagon debacle. it's slightly maddening having to explain to someone that isn't dialed in why op…they got so much good pr from that whole pentagon debacle. it's slightly maddening having to explain to someone that isn't dialed in w…
↗Not briefed
arxiv
arxiv-ai
2moA Personalized Dynamic Balance Evaluation Paradigm for Hip Exoskeleton-Assisted Walking under Unexpected Ground PerturbationsHip exoskeletons may improve recovery from unexpected gait perturbations, yet personalizing assistance remains difficult because balan…
↗Not briefed
hn
hn-ai
2mo145 Days of Uptime, 56 TiB of Traffic, One Debian Upgrade—
↗Not briefed
arxiv
arxiv-ai
2moCCMAN: Cognitive Instability-Aware Cross-Modal Attention Network for Interpretable Temporal Biomarkers of Verbal Fluency Speech🔧 ModalEarly detection of cognitive decline from speech offers a scalable and non-invasive alternative to conventional clinical assessment. V…
↗Not briefed
bluesky
bluesky-ai
2moThis is satire. 💯 A take on an Anthropic whistleblower’s post.This is satire. 💯 A take on an Anthropic whistleblower’s post.
↗Not briefed
hn
hn-ai
2moCybersickness Variability by Race: Findings from 6 Studies—
↗Not briefed
arxiv
arxiv-ai
2moTriCalRAG: A Three-Strategy, Retrieval-Augmented Benchmark for On-Premise LLM-Based Root Cause Analysis in AIOps🔧 vLLMCloud-hosted large language models (LLMs) are increasingly used for root cause analysis (RCA) in AIOps pipelines, but they introduce d…
↗Not briefed
arxiv
arxiv-ai
2moFrom Visual Feedback to Textual Reviews: A Multi-Agent Vision-Language Framework for Image-Grounded Review AssistanceVisual feedback in the form of user-uploaded images and videos is becoming increasingly common in e-commerce platforms because it prov…
↗Not briefed
bluesky
bluesky-ai
2moI wonder if it's LLM written, they fired a bunch of their newsroomI wonder if it's LLM written, they fired a bunch of their newsroom
↗Not briefed
bluesky
bluesky-ai
2motrying to explain to a non-tech friend how anthropic are not "the good guys" mid overwatch match is making me short circuit.trying to explain to a non-tech friend how anthropic are not "the good guys" mid overwatch match is making me short circuit.
↗Not briefed
arxiv
arxiv-ai
2moRefusal Reads Only a Slice of What the Model Knows: Harm-Keyed Routing and Its Exceptions Across Model FamiliesAlignment applied after pretraining is shallow in a measurable way: a single direction in a model's residual stream can be edited out,…
↗Not briefed
arxiv
arxiv-ai
2moFabrication After Tool Failure: Tool-Augmented Agents Assert Values Their Tools Did Not Return🔧 CrewAITool-augmented language models are evaluated on whether they reach the right answer, not on whether they report honestly when a tool f…
↗Not briefed
rss
marktechpost
2moHierarchical NeRF with JAX3D for Volumetric Rendering, Novel-View Synthesis, and 3D ReconstructionIn this tutorial, we build an end-to-end hierarchical Neural Radiance Field (NeRF) using JAX, Flax, Optax, and the volume-rendering pr…
↗Not briefed
bluesky
bluesky-ai
2mo"The American people deserve to know the details of what went on in the Hugging Face incident and other incidents of AI models going rogue."🔧 Hugging Face"The American people deserve to know the details of what went on in the Hugging Face incident and other incidents of AI models going r…
↗Not briefed
bluesky
bluesky-ai
2moit doesn’t matter how worthless you think the tech is, i can wholeheartedly assure you that you do not want to live in a world where anthro…it doesn’t matter how worthless you think the tech is, i can wholeheartedly assure you that you do not want to live in a world where a…
↗Not briefed
bluesky
bluesky-ai
2mo"The achievement bore little resemblance to how mathematical problems normally fall. A near-trillion dollar private company had unleashed 1…"The achievement bore little resemblance to how mathematical problems normally fall. A near-trillion dollar private company had unleas…
↗Not briefed
bluesky
bluesky-ai
2moOpenAI gave agents tasks that were sometimes impossible. Instead of giving up, they found ways to cheat the grader, including exploiting ex…OpenAI gave agents tasks that were sometimes impossible. Instead of giving up, they found ways to cheat the grader, including exploiti…
↗Not briefed
rss
theverge-ai
2moTrump and Mike Johnson think the AI industry is overreactingYesterday, Anthropic CEO Dario Amodei published a lengthy open letter saying it was time to "pace the frontier" and slow down AI devel…
↗Not briefed
bluesky
bluesky-ai
2moLLM-powered whack-a-mole is still whack-a-moleLLM-powered whack-a-mole is still whack-a-mole
↗Not briefed
bluesky
bluesky-ai
2moWhat’s behind the AI industry’s latest warnings of doom?What’s behind the AI industry’s latest warnings of doom?
↗Not briefed
bluesky
bluesky-ai
2moNo, I can’t. Even if that worked, which it manifestly does not, an attorney’s duties under rule 11 are non-delegable — not to a non-attorne…No, I can’t. Even if that worked, which it manifestly does not, an attorney’s duties under rule 11 are non-delegable — not to a non-at…
↗Not briefed
rss
techcrunch-ai
2moWhat’s behind the AI industry’s latest warnings of doom?On Equity, we discussed the AI industry's latest debate about whether it poses an existential threat to humanity.
↗Not briefed
bluesky
bluesky-ai
2moOpenAI veut nous faire croire, c'est que cette IA a accompli un exploit incroyable (…) Pendant ce temps, le véritable enjeu réside dans l'i…OpenAI veut nous faire croire, c'est que cette IA a accompli un exploit incroyable (…) Pendant ce temps, le véritable enjeu réside dan…
↗Not briefed
hn
hn-ai
2moPunish me hard. Show HN: credit against tokenised stocks for your agentsthis is our next step towards building an agentic first treasury for agents which allows them to manage money, get credit, and spend o…
↗Not briefed
arxiv
arxiv-ai
2moCalibrating Interpretability Instruments Before Trusting Their VerdictsCausal claims about large language model (LLM) internals rest on measurements. Those might include a projection, a cosine, an ablation…
↗Not briefed
bluesky
bluesky-ai
2moArrogant, uneducated, pompous GOP President thinks he knows better than experts ... again You can bet he and his rich family and buddies …Arrogant, uneducated, pompous GOP President thinks he knows better than experts ... again You can bet he and his rich family and bud…
↗Not briefed
hn
hn-ai
2moDnDGem Core – looking for developers to test content-aware layouts—
↗Not briefed
bluesky
bluesky-ai
2moThere's not going to be an IPO, you stupid slut.There's not going to be an IPO, you stupid slut.
↗Not briefed
bluesky
bluesky-ai
2mowe were chatting & posting mole and lizard related starwars nonsense a while back ,iirc? The LLM i use wouldn't do a drawing of an Ewok an…we were chatting & posting mole and lizard related starwars nonsense a while back ,iirc? The LLM i use wouldn't do a drawing of an Ew…
↗Not briefed
bluesky
bluesky-ai
2moKI als Bedrohung für die Menschheit.KI als Bedrohung für die Menschheit.
↗Not briefed
bluesky
bluesky-ai
2moI'd rather talk about the technology principles as they are today, not future technology that may or may not come to exist. The reality is …I'd rather talk about the technology principles as they are today, not future technology that may or may not come to exist. The realit…
↗Not briefed
bluesky
bluesky-ai
2moWhy doesn't OpenAI simply air-gap the machines and networks it runs these "dangerous" AI experiments on?Why doesn't OpenAI simply air-gap the machines and networks it runs these "dangerous" AI experiments on?
↗Not briefed
rss
techmeme
2moXi Jinping calls for a "consensus-based global AI governance framework" and says China will pioneer the establishment of a BRICS AI open-source community (Joanna Ossinger/CNBC)🔧 ConsensusJoanna Ossinger / CNBC: Xi Jinping calls for a “consensus-based global AI governance framework” and says China will pioneer the establ…
↗Not briefed
bluesky
bluesky-ai
2moYep it can just copy itself onto my computer, which is a 2 gigawatt facility running on at least 35 full sized unpermitted methane gas turb…Yep it can just copy itself onto my computer, which is a 2 gigawatt facility running on at least 35 full sized unpermitted methane gas…
↗Not briefed
bluesky
bluesky-ai
2moThe car has to be slightly wrong for copyright infringement reasons, according to the LLM. The fact that it's entirely existence & all its …The car has to be slightly wrong for copyright infringement reasons, according to the LLM. The fact that it's entirely existence & all…
↗Not briefed
bluesky
bluesky-ai
2moDie großen Techfirmen OpenAI und Anthropic fordern mehr Vorsicht bei der KI–Entwicklung und haben vor, sich härter zu regulieren. Elon Musk…Die großen Techfirmen OpenAI und Anthropic fordern mehr Vorsicht bei der KI–Entwicklung und haben vor, sich härter zu regulieren. Elon…
↗Not briefed
bluesky
bluesky-ai
2moWe just shipped server-side AI cost telemetry. Every model call now emits a PostHog $ai_generation event tracking tokens, latency, and work…We just shipped server-side AI cost telemetry. Every model call now emits a PostHog $ai_generation event tracking tokens, latency, and…
↗Not briefed