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 |
|---|---|---|---|---|
bluesky bluesky-ai | 1mo | Shut these damn things down already | Shut these damn things down already | ↗Not briefed |
hn hn-ai | 1mo | Anthropic is "destructively scanning" books. Should you care? | — | ↗Not briefed |
rss theverge-ai | 1mo | Meta’s Muse AI sent a YouTuber’s address to a stranger | Tech YouTuber Matt Robb says that Muse gave out his home address to a total stranger this weekend, after authorizing the bot to handle… | ↗Not briefed |
bluesky bluesky-ai | 1mo | "Invest in our company to wipe out humanity" is rather more on the nose than science fiction had led me to believe was likely. May I sugg… | "Invest in our company to wipe out humanity" is rather more on the nose than science fiction had led me to believe was likely. May I… | ↗Not briefed |
bluesky bluesky-ai | 1mo | 80% of Anthropic’s half trillion in data center commitments is non-cancelable or payable regardless of usage. That’s an incredible bet on … | 80% of Anthropic’s half trillion in data center commitments is non-cancelable or payable regardless of usage. That’s an incredible be… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Co-Linguistics: AI-augmented Theory Construction in Linguistics | LLMs have been studied in recent linguistics as potential models of humans' linguistic abilities. Here we discuss an entirely differen… | ↗Not briefed |
hn hn-ai | 1mo | Investors vehemently hate AI-generated decks, how do you get around that? | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Johnson Wants AI Safety Rules to Stay Voluntary ai-beat #AISafety #AIRegulations #TechPolicy #VoluntaryGuardrails #ArtificialIntelligence | Johnson Wants AI Safety Rules to Stay Voluntary ai-beat #AISafety #AIRegulations #TechPolicy #VoluntaryGuardrails #ArtificialIntellige… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Their technical reports actually go into great deal about how Pangram works. In short, it is NOT an LLM and does not feed scanned text back… | Their technical reports actually go into great deal about how Pangram works. In short, it is NOT an LLM and does not feed scanned text… | ↗Not briefed |
arxiv arxiv-ai | 1mo | RLTL;DR: Self-improvement by Internalizing Self-generated Feedback🔧 Qwen 3 | The common paradigm of reinforcement learning with verifiable rewards (RLVR) is to let agents make multiple attempts at a task, and op… | ↗Not briefed |
arxiv arxiv-ai | 1mo | ProCTI: Prototype-Refined Global Conditioning for Diffusion-Based Time Series Imputation | Time series imputation has progressed from statistical and deep learning approaches to diffusion-based models, which have shown strong… | ↗Not briefed |
bluesky bluesky-ai | 1mo | lrt an LLM translation would never have given us "All according to keikaku. (*TL note: keikaku means plan)" though | lrt an LLM translation would never have given us "All according to keikaku. (*TL note: keikaku means plan)" though | ↗Not briefed |
hn hn-ai | 1mo | Show HN: The Other Apple – Apple in the EU vs. the US, Feature by Feature | I got annoyed years ago that the Apple experience in Europe wasn’t the same as in the US (e.g. missing services or hardware difference… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Procedural Core: A Compact Recurrent Initialization for Vision Transformers | Transformers are typically trained from random initialization, requiring all their capabilities to emerge from large-scale optimizatio… | ↗Not briefed |
bluesky bluesky-ai | 1mo | “Anthropic said nearly a quarter of its revenue came from two customers last year, and as part of its risk factors, warned that many of its… | “Anthropic said nearly a quarter of its revenue came from two customers last year, and as part of its risk factors, warned that many o… | ↗Not briefed |
arxiv arxiv-ai | 1mo | SPLASH: Switching Parallel Layouts of Attention with Seamless Handoff for LLM Serving | No single way of parallelizing attention serves large language models well under all loads. Low concurrency favors tensor parallelism,… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Correct, Don't Delete: Mitigating Emergent Misalignment with Corrective Supervision | Fine-tuning a language model on a narrow set of harmful demonstrations, such as bad medical advice, can make it broadly misaligned on… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Australian agencies warned in May of incoming ‘vulnerability storm’ ahead of OpenAI breach | Australian agencies warned in May of incoming ‘vulnerability storm’ ahead of OpenAI breach | ↗Not briefed |
rss techmeme | 1mo | In a July letter to Rep. Jamie Raskin, Apple explained it removed ICEBlock and similar ICE-tracking apps in 2025 due to agent and detainee safety concerns (Clare Duffy/CNN) | Clare Duffy / CNN: In a July letter to Rep. Jamie Raskin, Apple explained it removed ICEBlock and similar ICE-tracking apps in 2025 du… | ↗Not briefed |
bluesky bluesky-ai | 1mo | テスラ モデルSをベースに、ワイドフェンダー＋フルエアロ＋鍛造ホイールでフルカスタム。 もし実車だったら…と想像が止まりません。AIならではの夢の一台。 Starting from a Tesla Model S, given aggressive overfenders, … | テスラ モデルSをベースに、ワイドフェンダー＋フルエアロ＋鍛造ホイールでフルカスタム。 もし実車だったら…と想像が止まりません。AIならではの夢の一台。 Starting from a Tesla Model S, given aggressive overfend… | ↗Not briefed |
rss microsoft-research | 1mo | Introducing Quine: An AI research system designed for the complexity of biology | Biology doesn't operate in silos, and neither should the AI representation of it. Quine is an early-stage research effort to create a… | ↗Not briefed |
rss bing-news-ai | 1mo | Trump gambles with bold midterms move as he unveils 'Super Intelligence' despite Democrats' fears | Donald Trump is set to unveil a new name for artificial intelligence as he embraces the nascent tech and advocates for further develop… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Sorry, one more thing on AI in schools: a recent Chinese study found that using AI in classrooms raised homework scores but LOWERED test sc… | Sorry, one more thing on AI in schools: a recent Chinese study found that using AI in classrooms raised homework scores but LOWERED te… | ↗Not briefed |
arxiv arxiv-ai | 1mo | AS$^2$D: Accelerating On-Demand Audio Understanding on Mobile Devices | Speculative decoding accelerates autoregressive generation by using a smaller drafter to propose tokens for batched verification by a… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Authority Bias in Language Models: Source Deference and User Agreement Are Not Interchangeable | Language models tend to agree with whatever a user asserts, and post-training increasingly targets this sycophancy so that models eval… | ↗Not briefed |
hn hn-ai | 1mo | Anthropic's IPO prospectus shows AI vision, surging costs | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anthropic Warns AI May Pose 'Existential Risks To Humanity' In IPO Filing #HuffingtonPost | Anthropic Warns AI May Pose 'Existential Risks To Humanity' In IPO Filing #HuffingtonPost | ↗Not briefed |
hn hn-ai | 1mo | Gamma-Band Auditory Steady-State Response and Attention: A Systemic Review🔧 Gamma | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | PHASE: Multi-Regime Modeling of Incompressible Magnetohydrodynamics | Magnetohydrodynamics (MHD) is central to plasma modeling in astrophysics, space science, fusion, and engineering, but resolving multis… | ↗Not briefed |
hn hn-ai | 1mo | French public debt soars to 119% of GDP, highest level since 1946 | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | TomoTransformer: Towards a Foundation Model for CT Reconstruction | Supervised deep learning has advanced sparse-view tomographic reconstruction. However, conventional models, which typically map filter… | ↗Not briefed |
arxiv arxiv-ai | 1mo | GraphVQ: Structure-Aware Autoregressive Decoding over Context-Quantized Graph Tokens | Graph foundation models need a discrete token representation, but casting a graph as a generatable token sequence faces a structural o… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Independent Verification Paths Are Not Independent: A Case Study of Common-Mode Failure in a Satellite Catalogue Pipeline | A common safeguard for a data pipeline is redundant computation: derive each published number by two routes built on different technol… | ↗Not briefed |
rss techcrunch-ai | 1mo | With Dazzle, Marissa Mayer bets your camera roll has more info on your life than your inbox | The former Yahoo CEO is introducing a new AI personal assistant based entirely on your photos. | ↗Not briefed |
arxiv arxiv-ai | 1mo | BlenDAgger: Blended Shared Control for Interactive Imitation Learning | Robot policies are frequently trained from human corrections, yet teleoperating a robot to provide corrections is burdensome, and huma… | ↗Not briefed |
hn hn-ai | 1mo | OpenAI Halts New Model over Safety Worries | — | ↗Not briefed |
hn hn-ai | 1mo | Historical Stock Prices API: Python, JavaScript and Curl Quickstart | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | XU-RS: Explaining Credal Width in Random-Set Language Models | Uncertainty estimates tell us how unsure a model is, but not why. Without knowing which parts of an input influences a model's uncerta… | ↗Not briefed |
rss techcrunch-ai | 1mo | Meta is expanding its AI agent Muse to small businesses | The tech giant says the agent can help owners run their business and find new customers. | ↗Not briefed |
bluesky bluesky-ai | 1mo | Don't miss today's Metacurity for a concise rundown of yesterday's frenetic AI developments cyber pros need to know, along with other cruci… | Don't miss today's Metacurity for a concise rundown of yesterday's frenetic AI developments cyber pros need to know, along with other… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Credit-Guided Policy Improvement for Test-time Adaptive Vision-Language Navigation | Test-time adaptation for vision-language navigation (TTA-VLN) enables pretrained policies to adapt online to unseen environments using… | ↗Not briefed |
arxiv arxiv-ai | 1mo | FOCUS: Training-Free Decision-Preserving Context Compression for LLM Agents | LLM agents accumulate interaction histories that grow linearly with task length, causing quadratic inference cost scaling and performa… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Rational Clarification by Assistive Agents via Value-of-Information Reasoning | Users of language-based assistive agents often make ambiguous requests. In response, an assistant can either directly act on its inter… | ↗Not briefed |
arxiv arxiv-ai | 1mo | ReLMem: Learning Recurrent Memory for Longitudinal EHR Modeling | Longitudinal electronic health record (EHR) modeling requires integrating new visits with an expanding patient history. Yet the contin… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Learning as Deepfakes Evolve: RF-Prompt for Continual Audio Deepfake Detection | Continual audio deepfake detection requires learning newly emerging deepfake methods while retaining discrimination of previously enco… | ↗Not briefed |
arxiv arxiv-ai | 1mo | FedSocket: Recipient-Executable Knowledge Exchange for Heterogeneous Multimodal Federated Learning | Federated knowledge must remain usable by recipients with different modalities, private architectures, and tasks. We present FedSocket… | ↗Not briefed |
arxiv arxiv-ai | 1mo | TReVS: Integrating Textual Relevance and Visual Saliency for Efficient Vision-Language Model Token Pruning | Vision-Language Models (VLMs) excel at visual understanding and reasoning but often incur substantial inference costs due to the large… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Expecting researcher taste and judgement to become increasingly valuable (this is from Anthropic's Opus 5.5 system card) anthropic.com/clau… | Expecting researcher taste and judgement to become increasingly valuable (this is from Anthropic's Opus 5.5 system card) anthropic.com… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Ornstein-Uhlenbeck Is Hard to Beat, Yet Superlinear Drift Ships Lower Transport Costs | Brešar and Mijatović \cite{bresar2025} show that Ornstein--Uhlenbeck diffusion is hard to beat in forward convergence under assumption… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Pair Difficulty Matters: Rethinking Pairwise LLM-as-a-Judge Evaluation and Consistency | Large Language Model judges are widely used to rank texts and text-generating systems through pairwise comparison, and their reliabili… | ↗Not briefed |