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
2moWe Trained a Model on Office Work. It Also Got Better at Coding—
↗Not briefed
hn
hn-ai
2moNovo partners with Anthropic to speed up drug development with Claude—
↗Not briefed
arxiv
arxiv-ai
2moRLLBC-Lib: An Educational Code Library for Reinforcement Learning and Learning-Based ControlReinforcement learning (RL) is an exciting concept as well as a remarkable success story worth sharing. However, RL builds on rather c…
↗Not briefed
bluesky
bluesky-ai
2moNovo partners with Anthropic to speed up drug development with Claude reut.rs/4xXQMrHNovo partners with Anthropic to speed up drug development with Claude reut.rs/4xXQMrH
↗Not briefed
bluesky
bluesky-ai
2moI mean, particularly when the people preaching all this doom cannot actually articulate the specifics of how an LLM would bring it about. L…I mean, particularly when the people preaching all this doom cannot actually articulate the specifics of how an LLM would bring it abo…
↗Not briefed
bluesky
bluesky-ai
2moBeside everything else, I’ve seen this method of deflection before and it was from Code Mystics. You pick the parts of the conversation yo…Beside everything else, I’ve seen this method of deflection before and it was from Code Mystics. You pick the parts of the conversati…
↗Not briefed
hn
hn-ai
2moShow HN: Pixel Agents – A pixel-art mission control for your Claude Code agents—
↗Not briefed
bluesky
bluesky-ai
2moobviously this solution pales in comparison to my solution of creating unmonitored swarms of LLM agents, then making one agent a commissar …obviously this solution pales in comparison to my solution of creating unmonitored swarms of LLM agents, then making one agent a commi…
↗Not briefed
bluesky
bluesky-ai
2moHonestly, reading the LLM discussions here feels like a window into some parallel universe with different physical laws. It bears no resemb…Honestly, reading the LLM discussions here feels like a window into some parallel universe with different physical laws. It bears no r…
↗Not briefed
bluesky
bluesky-ai
2mo‘After Anthropic and OpenAI raised prices for their most powerful models, the average discount offered by Chinese labs versus their America…‘After Anthropic and OpenAI raised prices for their most powerful models, the average discount offered by Chinese labs versus their Am…
↗Not briefed
bluesky
bluesky-ai
2moextremely objective llm you've got there google, i'm so excited for everything to be smashed through the fascism filter that also boils the…extremely objective llm you've got there google, i'm so excited for everything to be smashed through the fascism filter that also boil…
↗Not briefed
arxiv
arxiv-ai
2moLong-Lived Characters, Local Inference: Incremental Memory Maintenance for Game NPCsA game character should not have to reread its entire life before every conversation. For locally deployed language-model characters,…
↗Not briefed
bluesky
bluesky-ai
2moFranchement ça doit coûter cher d'être un incel. Une baie entière avec des H200 et des téraoctets de RAM, j'ai pas les moyens. Je vais fai…Franchement ça doit coûter cher d'être un incel. Une baie entière avec des H200 et des téraoctets de RAM, j'ai pas les moyens. Je vai…
↗Not briefed
bluesky
bluesky-ai
2moBut Anthropic and other companies are, supposedly, the experts in their own product. If they're programming an agent to pretend it's human …But Anthropic and other companies are, supposedly, the experts in their own product. If they're programming an agent to pretend it's h…
↗Not briefed
bluesky
bluesky-ai
2moyou kinda dont have an excuse to let your lock files rot any more bc you can just tell an LLM to upgrade it and fix any breakage lolyou kinda dont have an excuse to let your lock files rot any more bc you can just tell an LLM to upgrade it and fix any breakage lol
↗Not briefed
bluesky
bluesky-ai
2moi figure the chinese ai labs are using equally powerful llm based pentesting tools on the american side of the internet and quietly not exp…i figure the chinese ai labs are using equally powerful llm based pentesting tools on the american side of the internet and quietly no…
↗Not briefed
bluesky
bluesky-ai
2moobviously this solution pales in comparison to my solution of creating unmonitored swarms of LLM agents, giving them a credit card, and ask…obviously this solution pales in comparison to my solution of creating unmonitored swarms of LLM agents, giving them a credit card, an…
↗Not briefed
arxiv
arxiv-ai
2moSocial Laws for Multi-agent Coordination in Stochastic EnvironmentsIn multi-agent environments, coordinating agents to prevent interference and ensure robust individual performance is a critical challe…
↗Not briefed
bluesky
bluesky-ai
2moYour AI agents can now control your Google Home devicesYour AI agents can now control your Google Home devices
↗Not briefed
hn
hn-ai
2moAI Compilers Are Not Just Compilers for AI—
↗Not briefed
arxiv
arxiv-ai
2moComprehensive reconstruction of collider events with hypergraph representation learning and graph-conditioned diffusionIn particle collider experiments, event reconstruction is the task of inferring the kinematics of short-lived particles produced in th…
↗Not briefed
bluesky
bluesky-ai
2moCette semaine dans le Pavé numérique : ce n'est pas parce que les délires des techbros sur une super-intelligence exterminant l'humanité so…Cette semaine dans le Pavé numérique : ce n'est pas parce que les délires des techbros sur une super-intelligence exterminant l'humani…
↗Not briefed
rss
the-decoder
2moGoogle Deepmind launches interdisciplinary institute to tackle the big questions around AGIGoogle Deepmind has founded the Deepmind Institute (DMI), an interdisciplinary research platform focused on AGI. Led by Demis Hassabis…
↗Not briefed
bluesky
bluesky-ai
2moOpenAI endorsed a series of bipartisan bills before Congress that the company said will help to prevent AI models from accelerating the thr…OpenAI endorsed a series of bipartisan bills before Congress that the company said will help to prevent AI models from accelerating th…
↗Not briefed
rss
theverge-ai
2moGoogle will now let any AI agent run your smart homeGoogle is inviting third-party agents, including Claude and Open Claw, into Google Home. | Photo by Jennifer Pattison Tuohy / The Verg…
↗Not briefed
rss
techcrunch-ai
2moYour AI agents can now control your Google Home devices🔧 ChatGPTGoogle is launching early access to a new MCP server for Google Home, allowing AI agents like Claude, ChatGPT, and others to control c…
↗Not briefed
bluesky
bluesky-ai
2moLot of people being driven out of their gourds by LLM sycophancy, which is why I've employed Claude to spend all of its processing power to…Lot of people being driven out of their gourds by LLM sycophancy, which is why I've employed Claude to spend all of its processing pow…
↗Not briefed
bluesky
bluesky-ai
2moThat said this generally starts well before the "drafting an entire" piece part, but you can use an LLM pretty well to do research around a…That said this generally starts well before the "drafting an entire" piece part, but you can use an LLM pretty well to do research aro…
↗Not briefed
hn
hn-ai
2moCO3: Toward the Optimal FFI—
↗Not briefed
bluesky
bluesky-ai
2molarge corporations love to impose guard rails on their own pursuit of wealth and power you can just ask them to be nice and they will tha…large corporations love to impose guard rails on their own pursuit of wealth and power you can just ask them to be nice and they will…
↗Not briefed
bluesky
bluesky-ai
2moIn short, Anthropic trains its model not just to respond as a human would, but _to respond as a machine that thinks it's a human would_. It…In short, Anthropic trains its model not just to respond as a human would, but _to respond as a machine that thinks it's a human would…
↗Not briefed
bluesky
bluesky-ai
2moIs your position that OpenAI isn’t responsible? I think if I devised an autonomous lawn mowing machine and it barreled through my neighbors…Is your position that OpenAI isn’t responsible? I think if I devised an autonomous lawn mowing machine and it barreled through my neig…
↗Not briefed
hn
hn-ai
2moOmarchy on Mac—
↗Not briefed
hn
hn-ai
2moShow HN: Ivx/AI Chat, AI client that doesn't track you (with a better GUI imo)🔧 OpenRouter🔧 ChatGPTSince ChatGPT started pushing ads and somehow my instagram reels started getting more and more related to what I talked with ChatGPT,…
↗Not briefed
bluesky
bluesky-ai
2moSincerely curious: Did Claude screw up the subject-verb agreement in the sentence "My use of commas suck"? Or did Erickson add that post-AI…Sincerely curious: Did Claude screw up the subject-verb agreement in the sentence "My use of commas suck"? Or did Erickson add that po…
↗Not briefed
arxiv
arxiv-ai
2moHigher-order pruning of experts in mixture-of-experts language modelsMixture-of-Experts (MoE) language models suffer from large parameter counts, which create a significant memory bottleneck. Expert prun…
↗Not briefed
hn
hn-ai
2moAs Someone Who Has Written This Blog Post,—
↗Not briefed
bluesky
bluesky-ai
2moHe's still trivializing why it happened and how big of a deal it is in the first half, so much so that it doesn't read factually correct. …He's still trivializing why it happened and how big of a deal it is in the first half, so much so that it doesn't read factually corre…
↗Not briefed
bluesky
bluesky-ai
2moy'all partner is listening to an interview with that anthropic guy and i will now be telling him about yud and the lesswrong crowd. the qu…y'all partner is listening to an interview with that anthropic guy and i will now be telling him about yud and the lesswrong crowd. t…
↗Not briefed
bluesky
bluesky-ai
2moI feel nothing about his use of Claude and, to the extent I have any thoughts, I agree that it seems fairly responsible. It’s just the “I c…I feel nothing about his use of Claude and, to the extent I have any thoughts, I agree that it seems fairly responsible. It’s just the…
↗Not briefed
bluesky
bluesky-ai
2moWhat Suleyman calls 'circular reasoning' by Anthropic is, more flatly, con artistry. Anthropic instructs Claude to behave as if it has 'mor…What Suleyman calls 'circular reasoning' by Anthropic is, more flatly, con artistry. Anthropic instructs Claude to behave as if it has…
↗Not briefed
bluesky
bluesky-ai
2moHe’s getting very close to calling the LLM ‘Claudia’ and trying to marry it like Richard Dawkins.He’s getting very close to calling the LLM ‘Claudia’ and trying to marry it like Richard Dawkins.
↗Not briefed
bluesky
bluesky-ai
2moI've gone to dinner a couple of times with a friend over the past month and he left me a long compliment this morning about how well we bal…I've gone to dinner a couple of times with a friend over the past month and he left me a long compliment this morning about how well w…
↗Not briefed
bluesky
bluesky-ai
2mo“OpenAI and AARP will travel to 10 cities across the United States on Sept. 16 to teach older Americans how to use AI to spot scams.”“OpenAI and AARP will travel to 10 cities across the United States on Sept. 16 to teach older Americans how to use AI to spot scams.”
↗Not briefed
bluesky
bluesky-ai
2moThis is a thoughtful essay but I can't help but notice just how much it leans into the same anthropomorphism it critiques. And (I get it, i…This is a thoughtful essay but I can't help but notice just how much it leans into the same anthropomorphism it critiques. And (I get…
↗Not briefed
rss
techmeme
2moDelos Data, which develops network chips and software to connect AI chips in a data center, raised $100M from Matrix, Playground, Socratic Partners, and others (Stephen Nellis/Reuters)Stephen Nellis / Reuters: Delos Data, which develops network chips and software to connect AI chips in a data center, raised $100M fro…
↗Not briefed
arxiv
arxiv-ai
2moBeyond Outcomes: Dual-View Relational Learning for Efficient Agent BenchmarkingAgent benchmarks are substantially more costly to evaluate than conventional LLM benchmarks. Benchmark compression is therefore a natu…
↗Not briefed
arxiv
arxiv-ai
2moHow Much is a Human Right Worth? ECtHR-NPD: A Benchmark for Predicting Non-Pecuniary Damage AwardsExisting legal benchmarks cover diverse tasks, while continuous monetary remedies remain comparatively underexplored. We introduce ECt…
↗Not briefed
hn
hn-ai
2moGitMir just released Vibe, it's like Claude on steroids—
↗Not briefed
bluesky
bluesky-ai
2mo“Swarms” may have that dog, I’m not sure. I recommend not creating unmonitored swarms of LLM agents, and if you do, not giving them a credi…“Swarms” may have that dog, I’m not sure. I recommend not creating unmonitored swarms of LLM agents, and if you do, not giving them a…
↗Not briefed