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 | 1mo | Rationally Linearized MLPs | We present Rationally Linearized Multilayer Perceptrons(RaLiMPs), a new approach to efficiently linearizing MLPs using low-degree rati… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Que las LLM actuales son algoritmos estadísticos es algo que intento explicar a la gente pero sé que es una batalla perdida, es como explic… | Que las LLM actuales son algoritmos estadísticos es algo que intento explicar a la gente pero sé que es una batalla perdida, es como e… | ↗Not briefed |
hn hn-ai | 1mo | Ask HN: What if no AIpocalypse, no AI bubble, what if it just keeps advancing? | The two mainstream arguments are the extremes, because that's how our modern society rolls. It's either that there is an AI… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Not All Correct Fixes Run Alike: An Empirical Study of Patch Performance in Repository-Level Issue Resolution | Repository-level issue-resolution benchmarks such as SWE-bench evaluate submitted patches primarily based on whether they pass the cor… | ↗Not briefed |
hn hn-ai | 1mo | Anthropic Usage Policy to prohibit "abusive or cruel behavior" toward Claude | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | なお間に合わんかった模様(4月の記事) uravation.com/media/projec... | なお間に合わんかった模様(4月の記事) uravation.com/media/projec... | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anthropic “prohíbe comportamiento abusivo o cruel “ hacia Claude . Como si fuera humano . La antropomorfización de la IA es peligrosa . En… | Anthropic “prohíbe comportamiento abusivo o cruel “ hacia Claude . Como si fuera humano . La antropomorfización de la IA es peligrosa… | ↗Not briefed |
hn hn-ai | 1mo | Demons, Anti-Christ, and AI | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Probabilistic Ship Trajectory Forecasting using Transformer Models on AIS Data | Anticipation of vessel trajectories is an essential task when navigating at sea. This applies to both human navigators and autonomous… | ↗Not briefed |
bluesky bluesky-ai | 1mo | いま企業は、AnthropicやOpenAIから先端情報の提供を受けながら、生成AIも用いての不正アクセス対策やダメコンの準備を進めているらしいですね。先端情報が有料なのかは知りませんが、対策に用いられる生成AIはAnthropicやOpenAIのモデルです。 | いま企業は、AnthropicやOpenAIから先端情報の提供を受けながら、生成AIも用いての不正アクセス対策やダメコンの準備を進めているらしいですね。先端情報が有料なのかは知りませんが、対策に用いられる生成AIはAnthropicやOpenAIのモデルです。 | ↗Not briefed |
bluesky bluesky-ai | 1mo | Most bubbles end after the "confusion over the revenue figures" becomes obvious for what it is. "The confusion over the revenue figure 'l… | Most bubbles end after the "confusion over the revenue figures" becomes obvious for what it is. "The confusion over the revenue figu… | ↗Not briefed |
hn hn-ai | 1mo | Chip-hex:cost-benefit, energy efficiency and hexagonal | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | I was just thinking about how if, today, I need to do (software thing), it is likely to be cheaper and easier to just get an LLM to do it, … | I was just thinking about how if, today, I need to do (software thing), it is likely to be cheaper and easier to just get an LLM to do… | ↗Not briefed |
bluesky bluesky-ai | 1mo | also, LLMs will only improve and reduce prices so if you’re building greenfield and for some reason it’s not cheaper, you’re still probabl… | also, LLMs will only improve and reduce prices so if you’re building greenfield and for some reason it’s not cheaper, you’re still pr… | ↗Not briefed |
hn hn-ai | 1mo | OpenAI withdraws 3 preprints after releasing 722 manuscripts | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Personalization Without Stereotyping: Do Speech LLMs Know When to Turn a Deaf Ear? | Recent speech LLMs can leverage rich paralinguistic cues, enabling more complex and personalized interactions. However, little work in… | ↗Not briefed |
hn hn-ai | 1mo | Linux kernel dropping Assisted-by tag | — | ↗Not briefed |
hn hn-ai | 1mo | Opus 5.5 Took Fable's Job. Haiku 5.5 Proved Me Wrong | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Have you seen @velitchkov.eu's essay about all the "orders" of builders who might benefit from LLM-assisted coding? www.linkandth.ink/p/t… | Have you seen @velitchkov.eu's essay about all the "orders" of builders who might benefit from LLM-assisted coding? www.linkandth.in… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Tell me your thoughts on this phrase (and any variation of it): 'The authors used [LLM] to assist with language editing, grammar checking, … | Tell me your thoughts on this phrase (and any variation of it): 'The authors used [LLM] to assist with language editing, grammar check… | ↗Not briefed |
hn hn-ai | 1mo | Claude Dashboards: The Jupyter Notebooks of Agent Age | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | I was going to write a blog post about the openai dump (ok, it's about 3/4 written already) but this one captures the gist of my feelings p… | I was going to write a blog post about the openai dump (ok, it's about 3/4 written already) but this one captures the gist of my feeli… | ↗Not briefed |
hn hn-ai | 1mo | Anthropic bans users from being 'cruel' to its AI systems | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | Three OpenAI employees who were let go from the company have written an open letter to OpenAI to express their concerns about safety: “As … | Three OpenAI employees who were let go from the company have written an open letter to OpenAI to express their concerns about safety:… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Defence against the dark arts of distillation: Unsupervised discovery of subliminal learning | When a language model is distilled from a teacher that shares its base parameters, the student can inherit behavioural traits of the t… | ↗Not briefed |
bluesky bluesky-ai | 1mo | word on the subway | word on the subway | ↗Not briefed |
rss techmeme | 1mo | The UK plans to curb non-compete rules that force employees to wait before moving jobs or starting companies, an issue that has drawn complaints from startups (Shona Ghosh/Bloomberg) | Shona Ghosh / Bloomberg: The UK plans to curb non-compete rules that force employees to wait before moving jobs or starting companies,… | ↗Not briefed |
arxiv arxiv-ai | 1mo | MAGMA-Bench: Evaluating Long-Horizon Embodied Agents under Dynamic Tasks and Execution Failures | Long-horizon embodied agents must maintain task-relevant information, coordinate heterogeneous robots, and adapt their decisions as ex… | ↗Not briefed |
rss simon-willison | 1mo | A new feature for my, blog built using my voice | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | for this week's hard fork we talked to elizabeth dias about her fascinating article on anthropic's secret meetings w/ religious figures. j… | for this week's hard fork we talked to elizabeth dias about her fascinating article on anthropic's secret meetings w/ religious figur… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Library folks: anyone had conversations with faculty about LLM use and uh... getting scooped by the AI companies? How'd that go? #medlibs | Library folks: anyone had conversations with faculty about LLM use and uh... getting scooped by the AI companies? How'd that go? #medl… | ↗Not briefed |
arxiv arxiv-ai | 1mo | S-Space: Exploring Spatial Workspace in Multimodal Language Models | Multimodal language models (MLLMs) exhibit promising spatial reasoning capabilities, yet the internal mechanisms supporting them remai… | ↗Not briefed |
hn hn-ai | 1mo | Native Support for Rust on the GPU | — | ↗Not briefed |
arxiv arxiv-ai | 1mo | Simulating Training Spikes of Deep Neural Networks by Layer-wise Random Feature Models🔧 Replicate | Intermittent ``spikes'' in loss, during the pre-training of Large Language Models (LLMs), remains a mysterious phenomenon both theoret… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Anthropic bans swearing at the dishwasher. | Anthropic bans swearing at the dishwasher. | ↗Not briefed |
bluesky bluesky-ai | 1mo | Writing good docstrings is something I'm so shitty at, and in principle maybe an LLM could do it, but I think about how I feel when somebod… | Writing good docstrings is something I'm so shitty at, and in principle maybe an LLM could do it, but I think about how I feel when so… | ↗Not briefed |
bluesky bluesky-ai | 1mo | OpenAI and Anthropic are hiring former Trump officials for senior roles qz.com/openai-anthr... | OpenAI and Anthropic are hiring former Trump officials for senior roles qz.com/openai-anthr... | ↗Not briefed |
bluesky bluesky-ai | 1mo | The OpenAI employees: "We were fired for prioritizing safety over the near-term interests of OpenAI as a corporation."🇺🇸 | The OpenAI employees: "We were fired for prioritizing safety over the near-term interests of OpenAI as a corporation."🇺🇸 | ↗Not briefed |
bluesky bluesky-ai | 1mo | I can't remember where I read about the English Literature Prof who said we should critically read LLM text to understand its 'structure', … | I can't remember where I read about the English Literature Prof who said we should critically read LLM text to understand its 'structu… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Die drei von OpenAI kürzlich entlassenen Ex-Mitarbeiter setzen sich mit einem kritischen Brief zur Wehr und warnen vor den Entwicklungen in… | Die drei von OpenAI kürzlich entlassenen Ex-Mitarbeiter setzen sich mit einem kritischen Brief zur Wehr und warnen vor den Entwicklung… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Introducing luxt, the LLM fork of Nuxt, now maintained by Danielbot. Now with 5000% more slop and churn, and 70% less community. /S | Introducing luxt, the LLM fork of Nuxt, now maintained by Danielbot. Now with 5000% more slop and churn, and 70% less community. /S | ↗Not briefed |
arxiv arxiv-ai | 1mo | Beyond Domain-Routed Supervision: Cross-Teacher Guarding for Multi-Teacher On-Policy Distillation | Multi-teacher on-policy distillation (MOPD) extends on-policy distillation (OPD) by routing domain-specific data to specialized teache… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Causal Effect Identifiability in Causal Abstractions: A Strict Hierarchy of Criteria | Identifying causal effects from observational data typically requires a fully specified causal graph, an assumption that is often unre… | ↗Not briefed |
hn hn-ai | 1mo | Telegram Desktop: one-click account takeover via IPC injection | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | If you're not worried that an LLM will ask a better question than you can, or think more creatively about how to approach it than you can, … | If you're not worried that an LLM will ask a better question than you can, or think more creatively about how to approach it than you… | ↗Not briefed |
arxiv arxiv-ai | 1mo | Asymptotic Analysis of Discrete-Time Self-Attention Dynamics in Transformers🔧 Consensus | In machine learning, a self-attention dynamics describes the attention mechanism of transformers, and is obtained treating the layer i… | ↗Not briefed |
arxiv arxiv-ai | 1mo | RESCUE: Is This Alarm Worth Acting On? Counterfactual Evaluation of VLA Runtime Monitors | Runtime failure monitors for robot policies are typically evaluated by how accurately they distinguish failed from successful executio… | ↗Not briefed |
bluesky bluesky-ai | 1mo | Reports that OpenAI’s annualized revenue is closer to $50B than the previously reported $70B caused a chip stock selloff yesterday. A lot … | Reports that OpenAI’s annualized revenue is closer to $50B than the previously reported $70B caused a chip stock selloff yesterday. A… | ↗Not briefed |
hn hn-ai | 1mo | The Missing Map of the Sky | — | ↗Not briefed |
bluesky bluesky-ai | 1mo | 🏔️ A B.C. teen rescued from Crown Mountain's “Widowmaker” route says AI didn't send him there. He tells CBC he used Claude mainly for pack… | 🏔️ A B.C. teen rescued from Crown Mountain's “Widowmaker” route says AI didn't send him there. He tells CBC he used Claude mainly for… | ↗Not briefed |