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
15dCausal graph fusion across incomplete cohorts and causal steering of tabular foundation modelsClinical data are split across incomplete cohorts with partially overlapping variables. Causal discovery under missing data and multi-…
↗Not briefed
bluesky
bluesky-ai
15dLots of people talking about LLM-"polished" emails because of an article (I won't link to it). The argument for doing this is workload, but…Lots of people talking about LLM-"polished" emails because of an article (I won't link to it). The argument for doing this is workload…
↗Not briefed
arxiv
arxiv-ai
15dArtificial Shoppers and Attention CompetitionWe study how AI shopping interfaces affect price competition, platform profit, and consumer welfare. A platform hosts two of three pro…
↗Not briefed
bluesky
bluesky-ai
15dLuxury Cyberpunk Fashion Model 4K Wallpaper High Quality Digital Art #AIイラスト #AIArt #AIGirl #AIArtwork #AIArtist #AIArtcommunity #DigitalA…Luxury Cyberpunk Fashion Model 4K Wallpaper High Quality Digital Art #AIイラスト #AIArt #AIGirl #AIArtwork #AIArtist #AIArtcommunity #Dig…
↗Not briefed
arxiv
arxiv-ai
15dTravel Cost Calibration for Multi-Agent Target Assignment and PathfindingTarget assignment and pathfinding (TAPF) is an appealing extension of classical multi-agent pathfinding (MAPF): each agent has a set o…
↗Not briefed
bluesky
bluesky-ai
15d*Reading an LLM email through the principle of least interest* *Does not reply to email* Checkmate.*Reading an LLM email through the principle of least interest* *Does not reply to email* Checkmate.
↗Not briefed
hn
hn-ai
15dShow HN: Lost and Fought – a 90s Style RTS – browser with multiplayerLoved to play Earch 2140, KKND, C&C, WC1-3 Dune 2 and lots of others. Combined all the features i loved into a quick multiplayer b…
↗Not briefed
bluesky
bluesky-ai
15dwww.nytimes.com/2026/11/06/s...www.nytimes.com/2026/11/06/s...
↗Not briefed
bluesky
bluesky-ai
15d“Companies including Palantir and Meta also rushed to hire Washington lobbyists with closer ties to Democrats ahead of their anticipated vi…“Companies including Palantir and Meta also rushed to hire Washington lobbyists with closer ties to Democrats ahead of their anticipat…
↗Not briefed
bluesky
bluesky-ai
15dAnthropic IPO? A little better than a 50/50 this monthAnthropic IPO? A little better than a 50/50 this month
↗Not briefed
bluesky
bluesky-ai
15dIt's not really a question of whether I believe Altman when he claims they have models that are smarter than the best scientists, mathemati…It's not really a question of whether I believe Altman when he claims they have models that are smarter than the best scientists, math…
↗Not briefed
hn
hn-ai
15dLowvram: A CLI tool to run large AI models in low VRAM GPUs—
↗Not briefed
bluesky
bluesky-ai
15dLuxury Cyberpunk Fashion Model Night City Skyline Wallpapers 4K #AIイラスト #AIArt #AIGirl #AIArtwork #AIArtist #AIArtcommunity #DigitalArt #M…Luxury Cyberpunk Fashion Model Night City Skyline Wallpapers 4K #AIイラスト #AIArt #AIGirl #AIArtwork #AIArtist #AIArtcommunity #DigitalA…
↗Not briefed
bluesky
bluesky-ai
15dLuxury Cyberpunk Fashion Model Night City Skyline Wallpapers 4K #AIイラスト #AIArt #AIGirl #AIArtwork #AIArtist #AIArtcommunity #DigitalArt #M…Luxury Cyberpunk Fashion Model Night City Skyline Wallpapers 4K #AIイラスト #AIArt #AIGirl #AIArtwork #AIArtist #AIArtcommunity #DigitalA…
↗Not briefed
bluesky
bluesky-ai
15dLuxury Cyberpunk Fashion Model Night City Skyline Wallpapers 4K #AIイラスト #AIArt #AIGirl #AIArtwork #AIArtist #AIArtcommunity #DigitalArt #M…Luxury Cyberpunk Fashion Model Night City Skyline Wallpapers 4K #AIイラスト #AIArt #AIGirl #AIArtwork #AIArtist #AIArtcommunity #DigitalA…
↗Not briefed
arxiv
arxiv-ai
15dk-Volume: a Unified Active Learning StrategyIn the classification context where revealing the label of a sample is costly, we propose a new Active Learning (AL) strategy named k-…
↗Not briefed
arxiv
arxiv-ai
15dAgainst All LikelihoodMaximum likelihood is the dominant training principle for large language models, yet likelihood is neither necessary nor sufficient fo…
↗Not briefed
hn
hn-ai
15dVals AI found a critical safety and alignment issue in recent models—
↗Not briefed
arxiv
arxiv-ai
15dExploring Weight-Space Programming: Capabilities, Generalization, and LimitationsWeight-Space Programming (WSP) enables reasoning agents to directly inspect and modify neural network parameters, offering an alternat…
↗Not briefed
bluesky
bluesky-ai
15duncritical support for the vulnerability in its struggle against the LLMuncritical support for the vulnerability in its struggle against the LLM
↗Not briefed
bluesky
bluesky-ai
15drare that I get LLM-written emails but they are always annoying: long winded, syrupy, structured like an interaction with a cliched therapi…rare that I get LLM-written emails but they are always annoying: long winded, syrupy, structured like an interaction with a cliched th…
↗Not briefed
arxiv
arxiv-ai
15dSPSC: Zero-Shot Cross-Subject EEG-to-Image Reconstruction via Semantic Preservation and Subspace CalibrationPretrained vision-language models and diffusion generators have driven rapid progress in brain-to-image generation, while EEG-based re…
↗Not briefed
arxiv
arxiv-ai
15dKineCouple: Substrate-conditioned residual readout for enzyme kinetic predictionAccurate prediction of enzyme kinetic parameters can support enzyme engineering and metabolic modeling by complementing experimental c…
↗Not briefed
arxiv
arxiv-ai
15dCan LLMs Debug Robots by Editing Their Weights?A language model can repair a robot policy's lexical instruction grounding by editing just 65 token-embedding rows. We bring weight-sp…
↗Not briefed
bluesky
bluesky-ai
15dafter the E6AI/blep thing i have sloppers in my replies that aver ai is doing just fine, why are you so mean to us, etc. meanwhile in the r…after the E6AI/blep thing i have sloppers in my replies that aver ai is doing just fine, why are you so mean to us, etc. meanwhile in…
↗Not briefed
arxiv
arxiv-ai
15dCutpurse: Decomposing Explanation-Guided Model Extraction on Tabular DataExplanation APIs are often treated as increasing model-extraction risk, but an end-to-end attack comparison cannot show whether an exp…
↗Not briefed
bluesky
bluesky-ai
15dI think this confusion from LLM users also stems in part from confusion about what is good writing. The emails from LLMs are not better. Th…I think this confusion from LLM users also stems in part from confusion about what is good writing. The emails from LLMs are not bette…
↗Not briefed
arxiv
arxiv-ai
15dCalibrated Uncertainty Quantification on TDC ADMET: What Can Be Resolved Across 22 Datasets?Absorption, distribution, metabolism, excretion, and toxicity (ADMET) models often rely on small, heterogeneous datasets, yet their pr…
↗Not briefed
bluesky
bluesky-ai
15dLLMでムカつくのはどんなに止めろと指示を書いても更新とか入ると「これを対話形式で説明してみようか?」「長くなるからザクッとわかりやすく要約してみていい?」とか提案しだすところ。人類はこういうのが好きだと学習しやがって許せん。LLMでムカつくのはどんなに止めろと指示を書いても更新とか入ると「これを対話形式で説明してみようか?」「長くなるからザクッとわかりやすく要約してみていい?」とか提案しだすところ。人類はこういうのが好きだと学習しやがって許せん。
↗Not briefed
bluesky
bluesky-ai
15dMamba model reads farm fields from orbit 🛰️ 🔗 https://bioengineer.org/mamba-ai-model-reads-sentinel-2-satellite-data-to-map-seed-maize-y…Mamba model reads farm fields from orbit 🛰️ 🔗 https://bioengineer.org/mamba-ai-model-reads-sentinel-2-satellite-data-to-map-seed-ma…
↗Not briefed
hn
hn-ai
15dAsk HN: How's your home-cooked software revolution going?I know the title is glib, but there is a genuine question behind it. With LLMs getting better, and services like sprites, deno sandbox…
↗Not briefed
arxiv
arxiv-ai
15dEnabling Privacy-Preserving Natural Language Access to Industrial Smart Meter DataEnergy management officers and grid operators increasingly rely on smart meter data for anomaly detection, performance monitoring, and…
↗Not briefed
bluesky
bluesky-ai
15dJossain sanottiin että joo periaatteessa LLM ennustaa seuraavan sanan, mutta ei vain aiempien sanojen perusteella vaan kaiken oppimansa per…Jossain sanottiin että joo periaatteessa LLM ennustaa seuraavan sanan, mutta ei vain aiempien sanojen perusteella vaan kaiken oppimans…
↗Not briefed
arxiv
arxiv-ai
15dMitigating Covariate Shift in Behavioral Cloning via Robust Stationary Distribution CorrectionWe consider offline imitation learning (IL), which aims to train an agent to imitate from the dataset of expert demonstrations without…
↗Not briefed
rss
techmeme
15dIn preliminary findings, the EU says TikTok breached the DSA by failing to prevent its recommendation algorithm from trapping users in harmful "rabbit holes" (Alexander Martin/The Record)Alexander Martin / The Record: In preliminary findings, the EU says TikTok breached the DSA by failing to prevent its recommendation a…
↗Not briefed
hn
hn-ai
15dThe Correspondence Theory of LLM Truth—
↗Not briefed
bluesky
bluesky-ai
15dLmao that's a wild statement. I have never used an LLM, have no intention to, and have gone to some lengths to disable the BS pack-ins that…Lmao that's a wild statement. I have never used an LLM, have no intention to, and have gone to some lengths to disable the BS pack-ins…
↗Not briefed
arxiv
arxiv-ai
15dSafeOrchestrate: Runtime Checked Intent-Driven Orchestration for Cyber-Physical SystemsModern cyber physical systems (CPS) integrate heterogeneous sensors, actuators, controllers, and services, composed at runtime rather…
↗Not briefed
bluesky
bluesky-ai
15dLLMが最短手続きを探すように調整されていることはハッキング騒ぎやユーザーに対する虚偽成果の提出などで示され続けている(創造主を真似て安易な方にいく)ので、「完璧な回答」(42)が見つかったらそれしか答えなくなる。LLMが最短手続きを探すように調整されていることはハッキング騒ぎやユーザーに対する虚偽成果の提出などで示され続けている(創造主を真似て安易な方にいく)ので、「完璧な回答」(42)が見つかったらそれしか答えなくなる。
↗Not briefed
bluesky
bluesky-ai
15dI wanted AI to put my family videos in shooting order and add captions, so I started building a video editor. Test run: rainy-day footage …I wanted AI to put my family videos in shooting order and add captions, so I started building a video editor. Test run: rainy-day foo…
↗Not briefed
bluesky
bluesky-ai
15dLlama 4 Maverick (17B-128E MoE) vs GPT-4.1 (1M Context Frontier): Which LLM API actually protects your SaaS gross margin at 10,000 DAU? He…🔧 Llama 4Llama 4 Maverick (17B-128E MoE) vs GPT-4.1 (1M Context Frontier): Which LLM API actually protects your SaaS gross margin at 10,000 DAU…
↗Not briefed
hn
hn-ai
15dSeeking advice: Wix stolen my .com domainI transferred .com domain to wix from another domain registrar, paid for transfer and additional year of domain from my bank card, the…
↗Not briefed
hn
hn-ai
15dShow HN: Chorono – Open-source 2D animation editor controllable via MCP—
↗Not briefed
bluesky
bluesky-ai
15dThinking Machines Inkling-2-small is a good base model, but it’s also just a good model - 594B (17B active) - $0.30/$1.00 thinkingmachine…Thinking Machines Inkling-2-small is a good base model, but it’s also just a good model - 594B (17B active) - $0.30/$1.00 thinkingma…
↗Not briefed
rss
theverge-ai
15dProtesters break into Microsoft data center site in AmsterdamMembers of the action group Geef Tegengas put a banner on a crane during a protest. | Image: Getty Images Protesters have broken int…
↗Not briefed
hn
hn-ai
15dShow HN: Nandfs – A computer made of folders that runs DOOM—
↗Not briefed
bluesky
bluesky-ai
15dA colleague who uses AI for basically everything recently told me he uses it to handle his e-mails because he gets so many he couldn’t poss…A colleague who uses AI for basically everything recently told me he uses it to handle his e-mails because he gets so many he couldn’t…
↗Not briefed
hn
hn-ai
15dShow HN: PentestCanvas – Pentesting Workspace; with Claude and KaliAgent-powered pentesting workspace with a visual investigation canvas, sandboxed terminals, recorded tests, and traceable evidence.<p>…
↗Not briefed
arxiv
arxiv-ai
15dChallenges, Opportunities, and Misconceptions in Neurosymbolic AIWe examine opportunities for neurosymbolic AI alongside increasingly powerful neural models, particularly with regards to efficiency,…
↗Not briefed
hn
hn-ai
15dQuill, a Windows browser with a reader mode you can draw on—
↗Not briefed