News
The raw AI news stream — everything we ingest across RSS, Hacker News, arXiv, GitHub, Bluesky, YouTube and more, newest first.
| Source | Age | Title | Summary | Status |
|---|---|---|---|---|
arxiv arxiv-ai | 26d | Behavioral Forgetting Is Not Privacy in Machine Unlearning | Machine unlearning aims to remove targeted information from large language models, but existing evaluations often measure forgetting o… | ↗Not briefed |
arxiv arxiv-ai | 26d | Cross-Embodiment Causal Transfer from Human Videos to Robot Manipulation | Scaling robot manipulation models requires large amounts of interaction data, yet collecting diverse real-robot experience remains cos… | ↗Not briefed |
bluesky bluesky-ai | 26d | The main way the LLM boom has affected my life is to make a lot of things so much harder. It makes information filtering an even more const… | The main way the LLM boom has affected my life is to make a lot of things so much harder. It makes information filtering an even more… | ↗Not briefed |
bluesky bluesky-ai | 26d | One sees LLM-interested people elsewhere on here saying things like "how has AI been changing YOUR life?" And it's sort of an interesting … | One sees LLM-interested people elsewhere on here saying things like "how has AI been changing YOUR life?" And it's sort of an interes… | ↗Not briefed |
hn hn-ai | 26d | Show HN: Wordwonk – a new wordgame mechanic everyday | Claude invents and publishes a new word game mechanic everyday with an example. I do nothing. Checkout the first one | ↗Not briefed |
arxiv arxiv-ai | 26d | Single-stage joint diffusion of sky images and clear-sky index for ultra-short-term irradiance forecasting: what the frame-to-irradiance stage discards | Sky-image irradiance forecasting is dominated by a two-stage design: a video-prediction model forecasts future sky frames, and a separ… | ↗Not briefed |
bluesky bluesky-ai | 26d | Recording device hidden in toilet paper found in top-secret OpenAI meeting room arstechnica.com/tech-policy/... | Recording device hidden in toilet paper found in top-secret OpenAI meeting room arstechnica.com/tech-policy/... | ↗Not briefed |
bluesky bluesky-ai | 26d | Exceptional-thing-enjoyers don't enjoy LLM outputs Tuned-Randomized-Statistical-Average is Definitionally Unexceptional -- unless that is,… | Exceptional-thing-enjoyers don't enjoy LLM outputs Tuned-Randomized-Statistical-Average is Definitionally Unexceptional -- unless tha… | ↗Not briefed |
arxiv arxiv-ai | 26d | RSIFlow: Adaptive Self-Improvement Workflows with an Evolving Meta-Agent | Large language model (LLM) agents have demonstrated strong capabilities in complex tasks, including code generation, tool use, and mul… | ↗Not briefed |
rss techmeme | 26d | Nvidia's RTX Spark Superchip in the Surface Laptop Ultra outperforms Apple's M5 Max MacBook Pro by ~25% in agentic AI completion benchmarks using Qwen models (Jeffrey Kampman/Tom's Hardware) | Jeffrey Kampman / Tom's Hardware: Nvidia's RTX Spark Superchip in the Surface Laptop Ultra outperforms Apple's M5 Max MacBook Pro by ~… | ↗Not briefed |
arxiv arxiv-ai | 26d | In the Guise of Correction: How Framing and Cultural Adaptation Erode LLM Safety on Real-World Criminal Cases🔧 Elicit | Existing safety benchmarks rely largely on synthetic prompts with loosely grounded taxonomies and offer little focused analysis of sev… | ↗Not briefed |
bluesky bluesky-ai | 26d | Today's news roundup: - Trump says people who leak military plans should be shot - Eight states sue OpenAI alleging the company misled user…🔧 ChatGPT | Today's news roundup: - Trump says people who leak military plans should be shot - Eight states sue OpenAI alleging the company misled… | ↗Not briefed |
bluesky bluesky-ai | 26d | That Claude Mythos found so many bugs in Bank of America’s software stack they needed 3,000 employees and 1,000 contractors to fix them is … | That Claude Mythos found so many bugs in Bank of America’s software stack they needed 3,000 employees and 1,000 contractors to fix the… | ↗Not briefed |
bluesky bluesky-ai | 26d | The program for this year's Causal Data Science Meeting #CDSM26 is out! 38 talks over two days, Nov 4–5: double machine learning, diff-in-… | The program for this year's Causal Data Science Meeting #CDSM26 is out! 38 talks over two days, Nov 4–5: double machine learning, dif… | ↗Not briefed |
bluesky bluesky-ai | 26d | I'm so angry about this. There's plenty to be said about the inadequacy of the pipeline model but anthropic has probably spent more money f… | I'm so angry about this. There's plenty to be said about the inadequacy of the pipeline model but anthropic has probably spent more mo… | ↗Not briefed |
hn hn-ai | 26d | I do not use LLMs more | — | ↗Not briefed |
bluesky bluesky-ai | 26d | I am genuinely appalled that we're multiple years into this LLM age and we're still seeing these base-level flaws with using LLMs crop up i… | I am genuinely appalled that we're multiple years into this LLM age and we're still seeing these base-level flaws with using LLMs crop… | ↗Not briefed |
arxiv arxiv-ai | 26d | DINO-FIL: Multi-Layer Transformer Fusion for Generative Inpainting Forgery Localization | We propose DINO-FIL, a transformer-based framework for pixel-level localization of generative image forgeries. The method leverages mu… | ↗Not briefed |
arxiv arxiv-ai | 26d | Same Spectrum, Harder First-Order Training: Phase-Locking in Single- versus Multi-Qubit Re-Uploading Circuits | Angle-encoded variational quantum circuits implement truncated Fourier series. For one-dimensional inputs, a single-qubit re-uploading… | ↗Not briefed |
bluesky bluesky-ai | 26d | People claiming LLMs are sentient should be treated exactly like people claiming their calculator is sentient, because there is no fundamen… | People claiming LLMs are sentient should be treated exactly like people claiming their calculator is sentient, because there is no fun… | ↗Not briefed |
bluesky bluesky-ai | 26d | US FTC subpoenas OpenAI and Anthropic over consumer harms, Axios reports reut.rs/4dzSgAA | US FTC subpoenas OpenAI and Anthropic over consumer harms, Axios reports reut.rs/4dzSgAA | ↗Not briefed |
rss techmeme | 26d | Ledger unveils Atom, an $899 data-only Android-based smartphone companion device launching in summer 2027, as it pauses some sales following hardware tampering (Dominic Preston/The Verge) | Dominic Preston / The Verge: Ledger unveils Atom, an $899 data-only Android-based smartphone companion device launching in summer 2027… | ↗Not briefed |
bluesky bluesky-ai | 26d | Not only can LLM agents NOT distinguish between self and environment, they cannot distinguish between fact and fiction. They do not exper… | Not only can LLM agents NOT distinguish between self and environment, they cannot distinguish between fact and fiction. They do not… | ↗Not briefed |
rss bing-news-ai | 26d | Apple Intelligence Gets China Approval: Siri AI Left Out | Apple Intelligence may reach China in December with iOS 27.3, but Siri AI and other features will be missing. See what users can expec… | ↗Not briefed |
rss bing-news-ai | 26d | Apple Intelligence Gets China Approval: Siri AI Left Out | Apple Intelligence may reach China in December with iOS 27.3, but Siri AI and other features will be missing. See what users can expec… | ↗Not briefed |
bluesky bluesky-ai | 26d | The worst part for me is that I’m genuinely excited by the prospect of people who had not previously coded anything being able to create th… | The worst part for me is that I’m genuinely excited by the prospect of people who had not previously coded anything being able to crea… | ↗Not briefed |
hn hn-ai | 26d | Show HN: deCDN – a peer-to-peer CDN where every piece is BLAKE3-verified | — | ↗Not briefed |
arxiv arxiv-ai | 26d | CF-AVSS: A Coarse-to-Fine Audio-Visual Speech Separation Framework with Target-Speaker Refinement and Post-Fusion for Real-World AVSE Challenge | Audio-visual speech enhancement (AVSE) leverages visual cues to recover target speech from mixtures corrupted by reverberation, noise,… | ↗Not briefed |
hn hn-ai | 26d | Show HN: Face, a browser window over agents that keeps the plan in files | Hi HN, I'm the author. FACE is a single Go binary. You run `face` in the folder that holds your projects and it serves a page on… | ↗Not briefed |
bluesky bluesky-ai | 26d | O que mais me surpreende é uma empresa arrombada igual a OpenAI notificar a polícia federal sobre isso | O que mais me surpreende é uma empresa arrombada igual a OpenAI notificar a polícia federal sobre isso | ↗Not briefed |
hn hn-ai | 26d | gea3D – Three.js-style native software 3D renderer for tiny devices | — | ↗Not briefed |
bluesky bluesky-ai | 26d | Isso não deveria ser divulgado. Esses bilolê acham que ninguém tem acesso ao que escrevem pra essas LLM. Deixa eles se entregarem mais. | Isso não deveria ser divulgado. Esses bilolê acham que ninguém tem acesso ao que escrevem pra essas LLM. Deixa eles se entregarem mais. | ↗Not briefed |
hn hn-ai | 26d | Show HN: Multi persona library to eradicate Spec Hallucination | Spec hallucination is a real challenges, spec writing is a skill influenced by the individual experience and org culture. Tier 1 tech… | ↗Not briefed |
bluesky bluesky-ai | 26d | Yeah I am not participating in their humiliation kink. | Yeah I am not participating in their humiliation kink. | ↗Not briefed |
arxiv arxiv-ai | 26d | From Research Goals to Scientific Closure: Evaluating Agents Through Evidence-Supported Physics Research | Can AI agents organize scientific investigations that lead to well-supported conclusions? To evaluate this capability, we introduce Ph… | ↗Not briefed |
hn hn-ai | 26d | Embedded analytics engine for .NET 8|9|10 | — | ↗Not briefed |
bluesky bluesky-ai | 26d | 物理学者、失業へ。OpenAI「物理学の未解決問題TOP50のうち49問を解きました」 https://fire5ch.com/503651 | 物理学者、失業へ。OpenAI「物理学の未解決問題TOP50のうち49問を解きました」 https://fire5ch.com/503651 | ↗Not briefed |
arxiv arxiv-ai | 26d | Q($γ_H$): Entropy-Guided Horizon Mixing in Multi-Discount Q-Learning | A fixed discount factor in Q-learning imposes a single planning horizon throughout training, despite the differing demands of early ex… | ↗Not briefed |
arxiv arxiv-ai | 26d | Bias Mitigation in Audio-Enabled LLMs through Adversarial Training and Distillation Methods | Audio-enabled large language models (AuLLMs) allow users to interact with them simply by talking to them, without the need for an inte… | ↗Not briefed |
hn hn-ai | 26d | PdfCraft v0.8.0 | — | ↗Not briefed |
bluesky bluesky-ai | 26d | I have more sympathy for Pikachu locked in a N64 cart than I ever will for a LLM | I have more sympathy for Pikachu locked in a N64 cart than I ever will for a LLM | ↗Not briefed |
hn hn-ai | 26d | Open source for 27 years, unread: what LLMs changed about the Xanadu code | — | ↗Not briefed |
bluesky bluesky-ai | 26d | These folks have emotional attachment with an LLM. They may rely on it for friendship, love, therapy, or suchlike. Having it be alive or co… | These folks have emotional attachment with an LLM. They may rely on it for friendship, love, therapy, or suchlike. Having it be alive… | ↗Not briefed |
bluesky bluesky-ai | 26d | AI agents now run entire cyber kill chains—threats from Claude models growing fast. #AI #Cybersecurity #Anthropic #AIagents #ThreatIntel #M… | AI agents now run entire cyber kill chains—threats from Claude models growing fast. #AI #Cybersecurity #Anthropic #AIagents #ThreatInt… | ↗Not briefed |
bluesky bluesky-ai | 26d | A fun experiment I've been working on: how green is Brussels really, seen from the street rather than from above? I ran 42,037 Street View… | A fun experiment I've been working on: how green is Brussels really, seen from the street rather than from above? I ran 42,037 Street… | ↗Not briefed |
arxiv arxiv-ai | 26d | DeciCAD: Harness-Grounded Learning of Action Policies for CAD Agents | Large Language Model agents operate through closed-loop action policies whose decision space is defined by an agent harness. We invest… | ↗Not briefed |
hn hn-ai | 26d | Open-sourced helm chart for run:AI | — | ↗Not briefed |
hn hn-ai | 26d | ESP32-iPhone-agent: Let an AI model use an iPhone via ESP32, AirPlay, USB HID | — | ↗Not briefed |
bluesky bluesky-ai | 26d | This is the first thing I had to see this morning? Another clown on the Internet who wants you to believe that AI is sentient? Splendid. A… | This is the first thing I had to see this morning? Another clown on the Internet who wants you to believe that AI is sentient? Splendi… | ↗Not briefed |
rss techmeme | 26d | Amazon launches two AI wearables, the $75 Bee Frontier wristband and $100 Bee Eclipse magnetic phone-attachment recorder, with a $12/month Bee Pro subscription (Bloomberg) | Bloomberg: Amazon launches two AI wearables, the $75 Bee Frontier wristband and $100 Bee Eclipse magnetic phone-attachment recorder, w… | ↗Not briefed |