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
10dToolTurns: Benchmarking LLM Agents in Multi-Turn DialogueTool-using language models are increasingly deployed in conversational settings, yet most large-scale tool-use benchmarks focus on iso…
↗Not briefed
bluesky
bluesky-ai
10dThat’s what I don’t get about the “If LLM had the answer, some other small indie studio solved it already” - MAYBE. Or maybe it just connec…That’s what I don’t get about the “If LLM had the answer, some other small indie studio solved it already” - MAYBE. Or maybe it just c…
↗Not briefed
arxiv
arxiv-ai
10dEdge-Native Multimodal RAG: Hardware-Aware Quantization and Architectural Optimization for Consumer GPUsDeployment of retrieval-augmented generation (RAG) for Edge-AI is constrained by memory and bandwidth limitations. A standard 7B param…
↗Not briefed
arxiv
arxiv-ai
10dA Transferable Pretrained 3D Model for Impedance Inversion with Sparse-Well Prior ConditioningDeep-learning-based seismic impedance inversion commonly requires retraining for each target survey, which may generalize poorly beyon…
↗Not briefed
bluesky
bluesky-ai
10dJust subpoena their LLM queriesJust subpoena their LLM queries
↗Not briefed
bluesky
bluesky-ai
10dI think I hate anthropic a little more than the rest because they larp being the vanguard of ethicsI think I hate anthropic a little more than the rest because they larp being the vanguard of ethics
↗Not briefed
youtube
youtube-mattvidpro
10dGemini 4 Argon is Cheaper Than Opus... But the Catch?What's up, guys? So, I've had a stomach bug for like 5 days. It has been killing me, but I'm feeling better. I'm back into it. Gemini…
↗Not briefed
rss
techmeme
10dHow US-based game publishers are losing ground in the $121B mobile gaming market to Chinese rivals, which can make games quicker, cheaper, and more profitably (Bloomberg)Bloomberg: How US-based game publishers are losing ground in the $121B mobile gaming market to Chinese rivals, which can make games qu…
↗Not briefed
hn
hn-ai
10dAsk HN: Why doesn't Ramp AI Index include OpenRouter or other aggregators?🔧 OpenRouterI spend a lot of time looking at these two open data resources, which I&#x27;m sure many of you do as well:<p>Open Router Rankings, es…
↗Not briefed
arxiv
arxiv-ai
10dRethinking "RL Generalizes, SFT Memorizes": The Role of SFT DataLarge language models trained with Reinforcement Learning (RL) with verifiable rewards exhibit strong reasoning ability and broad gene…
↗Not briefed
hn
hn-ai
10dAnthropic Software Engineer Salaries—
↗Not briefed
arxiv
arxiv-ai
10dMuEdit: An efficient multi-task editing method towards inter-domain knowledge conflictsLarge language models (LLMs) require frequent knowledge updates across heterogeneous domains, but existing editing methods fail when a…
↗Not briefed
hn
hn-ai
10dGSD-LQ: Zero-calibration sub-4-bit compression and 7B in 3.78 GB VRAM—
↗Not briefed
arxiv
arxiv-ai
10dHy ASR 3.0 Technical ReportWe present Hy ASR 3.0, a large language model (LLM)-based automatic speech recognition (ASR) system that delivers strong recognition a…
↗Not briefed
bluesky
bluesky-ai
10dWhat we should ultimately take from the heated debate around consulting an LLM for a handful of lines of code is not that we need to direct…What we should ultimately take from the heated debate around consulting an LLM for a handful of lines of code is not that we need to d…
↗Not briefed
bluesky
bluesky-ai
10dthe LLM thinks I'm a Zhou dynasty failsonthe LLM thinks I'm a Zhou dynasty failson
↗Not briefed
bluesky
bluesky-ai
10dReally? I heard they were looking at a "conservative" (still insane) one trillion dollars. I mean, Anthropic is married to Nvidia. And Nvi…Really? I heard they were looking at a "conservative" (still insane) one trillion dollars. I mean, Anthropic is married to Nvidia. An…
↗Not briefed
arxiv
arxiv-ai
10dEpigraph Form Reinforcement Learning-based Safe Trajectory Tracking Control for Connected and Autonomous VehiclesSafe trajectory tracking (STT) aims to enable autonomous vehicles to follow a reference trajectory accurately while avoiding collision…
↗Not briefed
hn
hn-ai
10dClaude usage limit reset is a gambler's rushI have a humble Pro Claude sub for my personal projects, and I use it on Android.<p>So I used up my weekly limit on Monday, and I was…
↗Not briefed
bluesky
bluesky-ai
10dBREAKING: Anthropic bankers prepare to increase IPO valuation to $1.5 trillion or more, per FTBREAKING: Anthropic bankers prepare to increase IPO valuation to $1.5 trillion or more, per FT
↗Not briefed
bluesky
bluesky-ai
10dThe backstory of OpenAI really does piss me off Basically, Google was doing slow, incremental work on LLMs and Sam Altman just...walked ou…The backstory of OpenAI really does piss me off Basically, Google was doing slow, incremental work on LLMs and Sam Altman just...walk…
↗Not briefed
arxiv
arxiv-ai
10dUnderstanding the Effects of Multi-Agent Deliberation for Human-AI Decision-MakingMulti-agent deliberation (MAD), where multiple LLMs engage in structured discussion, improves AI performance on many benchmarks, but w…
↗Not briefed
bluesky
bluesky-ai
10dI would rather be called slurs than have to endure a fuckass LLM sucking up to meI would rather be called slurs than have to endure a fuckass LLM sucking up to me
↗Not briefed
hn
hn-ai
10dClaude Code is accidentally exposing secrets all over GitHub—
↗Not briefed
bluesky
bluesky-ai
10dFuck Anthropic. Using publicly funded science producing publicly available data used to train their algorithms for privatised profit. Fuck …Fuck Anthropic. Using publicly funded science producing publicly available data used to train their algorithms for privatised profit.…
↗Not briefed
bluesky
bluesky-ai
10dit looks like every time you go to your analytics anywhere in YouTube Studio now, it prompts the site's LLM to generate a "channel summary"…it looks like every time you go to your analytics anywhere in YouTube Studio now, it prompts the site's LLM to generate a "channel sum…
↗Not briefed
bluesky
bluesky-ai
10dalso the conversation originally was a group of IGF jurors who didn't want to vote for a studio using LLMs , which they are allowed to do. …also the conversation originally was a group of IGF jurors who didn't want to vote for a studio using LLMs , which they are allowed to…
↗Not briefed
arxiv
arxiv-ai
10dNested Ensemble Learning with Topological Data Analysis for Graph Classification and RegressionWe propose a nested ensemble learning framework that utilizes Topological Data Analysis (TDA) to extract and integrate topological fea…
↗Not briefed
bluesky
bluesky-ai
10dToday's GitHub trending: agent skills again. - mattpocock/skills: small, composable skills for any model (+1,225 today) - homeassistant-ai/…Today's GitHub trending: agent skills again. - mattpocock/skills: small, composable skills for any model (+1,225 today) - homeassistan…
↗Not briefed
bluesky
bluesky-ai
10da heartwarming github issue https://github.com/openai/codex/issues/59060a heartwarming github issue https://github.com/openai/codex/issues/59060
↗Not briefed
arxiv
arxiv-ai
10dSocial Engineering Attack Detection in Corporate CommunicationsSocial Engineering attacks continue to pose a significant risk to the corporate world with a growing level of sophistication. It is di…
↗Not briefed
bluesky
bluesky-ai
10dhi mr. p. boat, long time reader! the code generated from their LLM usage came from other coders who are not seeing any of that big walk re…hi mr. p. boat, long time reader! the code generated from their LLM usage came from other coders who are not seeing any of that big wa…
↗Not briefed
hn
hn-ai
10dShow HN: Kotlean: write Android apps in Lean—
↗Not briefed
hn
hn-ai
10dGetting Started with OpenIndiana—
↗Not briefed
bluesky
bluesky-ai
10dit's so cool that despite everything being llm nowadays no one seems to be able to manage to actually care about bundle sizesit's so cool that despite everything being llm nowadays no one seems to be able to manage to actually care about bundle sizes
↗Not briefed
arxiv
arxiv-ai
10dMachine Learning Moment Closure for the Boltzmann Transport Equation: Enforcing HyperbolicityThe Boltzmann transport equation describes particle transport across collisional regimes, but resolving its velocity dependence is exp…
↗Not briefed
bluesky
bluesky-ai
10dLLM usage for coding is also bad.LLM usage for coding is also bad.
↗Not briefed
youtube
youtube-wesroth
10dAnthropic just built it...A few days ago, Anthropic updated the S1 filings to say that recursive self-improvement is imminent and they're about to meet the resp…
↗Not briefed
arxiv
arxiv-ai
10dCan an AI Agent Be Epistemically Akratic? Higher-Order Evidence & the Limits of Self-CorrectionWhen we judge that our reasoning was defective, it seems plausible that we should abandon the conclusion it produced. This thought has…
↗Not briefed
arxiv
arxiv-ai
10dCharacterizing Local Breadth in Language GenerationLarge language models sharpen the validity--breadth trade-off: a generator should avoid hallucinations without collapsing onto a narro…
↗Not briefed
hn
hn-ai
10dMoofPort: An Open-Source WiFi Ethernet Driver for the Wi-Fi on Classic Mac—
↗Not briefed
youtube
youtube-aiexplained
10d‘We may release models we can’t control’: Amodei, Altman … and Gemini CarbonOkay, it is true that half the internet seemed to get hacked in the last week, but for me, being totally honest, this counts as a quie…
↗Not briefed
bluesky
bluesky-ai
10dNow, if only I could put my money where my mouth is and use Claude for this purpose on my own papers I just hate AI so much... 😢 (lots of…Now, if only I could put my money where my mouth is and use Claude for this purpose on my own papers I just hate AI so much... 😢 (lo…
↗Not briefed
bluesky
bluesky-ai
10dI'm making an old-school RPG, Star Stones, entirely with AI. Very early development. Code by Claude, 3D models from Tripo and Meshy, music …🔧 ElevenLabsI'm making an old-school RPG, Star Stones, entirely with AI. Very early development. Code by Claude, 3D models from Tripo and Meshy, m…
↗Not briefed
bluesky
bluesky-ai
10d"The $50 million commitment is equivalent to about 0.0036% of Anthropic’s expected $1.4 trillion valuation. Anthropic’s donation comes one …"The $50 million commitment is equivalent to about 0.0036% of Anthropic’s expected $1.4 trillion valuation. Anthropic’s donation comes…
↗Not briefed
bluesky
bluesky-ai
10dtbh i think this is why open models are extremely important. i've had a great experience with deepseek v4.1-1017 flash in reverse engineeri…tbh i think this is why open models are extremely important. i've had a great experience with deepseek v4.1-1017 flash in reverse engi…
↗Not briefed
bluesky
bluesky-ai
10d3/ Run your own workload parameters & interactive FinOps stress test in our full engineering guide: https://tokenmargin.dev/guides/openrout…🔧 OpenRouter3/ Run your own workload parameters & interactive FinOps stress test in our full engineering guide: https://tokenmargin.dev/guides/ope…
↗Not briefed
bluesky
bluesky-ai
10dveterans day is tomorrow and my roommate's out of town and i'm about 50% through a claude token budget that refreshes 9am thursday so i t…veterans day is tomorrow and my roommate's out of town and i'm about 50% through a claude token budget that refreshes 9am thursday s…
↗Not briefed
arxiv
arxiv-ai
10dMLIRBench - A Benchmark and Dataset for Compiler and Hardware-Aware ReasoningIn recent years, large language models (LLMs) have become increasingly popular in software engineering and have performed well on diff…
↗Not briefed
arxiv
arxiv-ai
10dFOE-Dreamer: Deployment-Efficient Learning of Cyber Defense Policies in Operational NetworksAutonomous cyber defense is widely studied with deep reinforcement learning, but most work trains policies inside simulators because l…
↗Not briefed