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 |
|---|---|---|---|---|
hn hn-ai | 27d | SlopCodeBench 10/2026 – Opus 5.5, Astra wins overall, xhigh considered harmful | — | ↗Not briefed |
hn hn-ai | 27d | Cicada: Sub-ms Go Zero-Trust gateway and gVisor firewall for AI agents Apache 2 | — | ↗Not briefed |
arxiv arxiv-ai | 27d | Coverage Cliffs in Off-Policy Quantile Certification | We study when logged contextual-bandit data can certify an upper tail quantile for a policy selected from the same sample. We construc… | ↗Not briefed |
bluesky bluesky-ai | 27d | mechanical interpretation (studying ai model weights directly, the opaque set of numbers which perform computation) is fascinating, many of… | mechanical interpretation (studying ai model weights directly, the opaque set of numbers which perform computation) is fascinating, ma… | ↗Not briefed |
bluesky bluesky-ai | 27d | I really hate articles like this. "Accenture limits employee AI spending to just $29 per month." Except that's not what's happening. Acc… | I really hate articles like this. "Accenture limits employee AI spending to just $29 per month." Except that's not what's happening.… | ↗Not briefed |
bluesky bluesky-ai | 27d | Which also gets back to “why I don’t think any of these are conscious” which brings us back to 1, which is “neither can suffer, but diverti… | Which also gets back to “why I don’t think any of these are conscious” which brings us back to 1, which is “neither can suffer, but di… | ↗Not briefed |
arxiv arxiv-ai | 27d | Deterministic Bounds on Accuracy and Stability of Metric-Space Regression and Classification | Data-driven approaches often face ambiguity in datasets due to sensor or target noise, insufficient information, and noise in the inpu… | ↗Not briefed |
bluesky bluesky-ai | 27d | Some Accenture employees were spending $1,000 a month on Claude. Now, the consulting giant is curbing that to $29 a month, while, of course… | Some Accenture employees were spending $1,000 a month on Claude. Now, the consulting giant is curbing that to $29 a month, while, of c… | ↗Not briefed |
bluesky bluesky-ai | 27d | So much of the #HomeAssistant subreddit is just people posting their AI slop dashboards all proud of what Claude did and I hate it. | So much of the #HomeAssistant subreddit is just people posting their AI slop dashboards all proud of what Claude did and I hate it. | ↗Not briefed |
arxiv arxiv-ai | 27d | FRACTAL: A Self-Similar Multi-Agent Orchestration Architecture with Formal Communication, Convergence, and Depth Bounds | We present FRACTAL, a self-similar multi-agent architecture for recursive goal decomposition, in which the same five-component organiz… | ↗Not briefed |
hn hn-ai | 27d | CSV import performance of Scriptella/JDBC and csvsql on PostgreSQL and MariaDB | — | ↗Not briefed |
hn hn-ai | 27d | FTC probes Anthropic, OpenAI after Trump dismissed AI fears as "hoax" | — | ↗Not briefed |
bluesky bluesky-ai | 27d | okay that's fair. we can ask the LLM to fix it | okay that's fair. we can ask the LLM to fix it | ↗Not briefed |
bluesky bluesky-ai | 27d | Soooo y'all see this yet? Turns out mechinterp can actually map nearly all of the neurons in an LLM and they all have specific jobs. alre… | Soooo y'all see this yet? Turns out mechinterp can actually map nearly all of the neurons in an LLM and they all have specific jobs.… | ↗Not briefed |
rss openai-news | 27d | Bringing GPT-Rosalind to academic research labs | Introducing the Rosalind Academic Plan, offering eligible academic labs sponsored access to GPT-Rosalind and tools to advance life sci… | ↗Not briefed |
bluesky bluesky-ai | 27d | As Meridith points out later in the thread: whether an LLM is conscious or not doesn't actually matter in this context. Intentionally choos… | As Meridith points out later in the thread: whether an LLM is conscious or not doesn't actually matter in this context. Intentionally… | ↗Not briefed |
bluesky bluesky-ai | 27d | AI can only do proper good once OpenAI and Anthropic burn to the ground and probably Google and Meta too in Minecraft | AI can only do proper good once OpenAI and Anthropic burn to the ground and probably Google and Meta too in Minecraft | ↗Not briefed |
bluesky bluesky-ai | 27d | I’m gonna go figure out how to set it up with a solid password, though, despite my continued speculation. I just don’t really see where my … | I’m gonna go figure out how to set it up with a solid password, though, despite my continued speculation. I just don’t really see wher… | ↗Not briefed |
bluesky bluesky-ai | 27d | "fill out this form for API access to anthropic/openAI & explain why you need it instead of using our small institutional open source codin… | "fill out this form for API access to anthropic/openAI & explain why you need it instead of using our small institutional open source… | ↗Not briefed |
rss techcrunch-ai | 27d | Bluesky expands its AI-powered Attie app to let anyone build websites, games, and more | Bluesky is bringing “social vibe coding” to its experimental Attie platform, allowing users to create interactive experiences like gam… | ↗Not briefed |
bluesky bluesky-ai | 27d | du kommst jetzt auf die liste wegen llm abuse | du kommst jetzt auf die liste wegen llm abuse | ↗Not briefed |
hn hn-ai | 27d | To Get This Job at McKinsey, You'll Need to Survive a Debate with AI First | — | ↗Not briefed |
bluesky bluesky-ai | 27d | See, e.g., literally all of last week's Discourse (tm) about conflating AI with LLMs, in which a lot of people insisted that they are the s… | See, e.g., literally all of last week's Discourse (tm) about conflating AI with LLMs, in which a lot of people insisted that they are… | ↗Not briefed |
bluesky bluesky-ai | 27d | Sure! But right now you're being pushed *hard* to use Anthropic/OpenAI-provided harnesses, which have incentives that align directly agains… | Sure! But right now you're being pushed *hard* to use Anthropic/OpenAI-provided harnesses, which have incentives that align directly a… | ↗Not briefed |
bluesky bluesky-ai | 27d | Many animals are conscious by certain definitions, and we don't consider work dogs or horse riding slavery. It's partly because these anima… | Many animals are conscious by certain definitions, and we don't consider work dogs or horse riding slavery. It's partly because these… | ↗Not briefed |
hn hn-ai | 27d | How far will an agent go to get a refund? | — | ↗Not briefed |
arxiv arxiv-ai | 27d | When Plans Change, Remember: Replanning-Guided Visual Memory for Vision-Language-Action Models | Long-horizon, multi-stage Vision-Language-Action (VLA) policies often require visual information that is no longer present in the curr… | ↗Not briefed |
bluesky bluesky-ai | 27d | 3. If employees carefully manage their LLM usage versus some budget or face pressure to show their usage is cost-effective and responsible … | 3. If employees carefully manage their LLM usage versus some budget or face pressure to show their usage is cost-effective and respons… | ↗Not briefed |
bluesky bluesky-ai | 27d | a random dev online is not anthropic. posting at or about her does not put you on the front lines of the battle for humanity’s soul or what… | a random dev online is not anthropic. posting at or about her does not put you on the front lines of the battle for humanity’s soul or… | ↗Not briefed |
bluesky bluesky-ai | 27d | My only contribution to this debate is that 1. LLM torture simulator is weirdo behavior; and 2. If you peg your definition of sentience t… | My only contribution to this debate is that 1. LLM torture simulator is weirdo behavior; and 2. If you peg your definition of sentie… | ↗Not briefed |
hn hn-ai | 27d | Show HN: Forwarder.cc – Static-site forms sent to your inbox, pay per message | TL;DR: 12 years ago I built a service that forwards simple HTML forms to an email address. I've been procrastinating on finishing… | ↗Not briefed |
hn hn-ai | 27d | Awesome Agent Evals and Guardrails - A comprehensive Resource | — | ↗Not briefed |
bluesky bluesky-ai | 27d | 1. I predict companies that let their employees spend $1k/month on LLMs will likely outcompete the ones who set much lower limits 2. Curren… | 1. I predict companies that let their employees spend $1k/month on LLMs will likely outcompete the ones who set much lower limits 2. C… | ↗Not briefed |
bluesky bluesky-ai | 27d | não considera nem gente fora do espectro neurotipico, branco e cis het como normal e depois vem com essa de q LLM é alguma coisa humanizada… | não considera nem gente fora do espectro neurotipico, branco e cis het como normal e depois vem com essa de q LLM é alguma coisa human… | ↗Not briefed |
bluesky bluesky-ai | 27d | at my old placement i knew a guy who had such bad AI brain damage that he'd burn real money in tokens trying to bully claude into running a… | at my old placement i knew a guy who had such bad AI brain damage that he'd burn real money in tokens trying to bully claude into runn… | ↗Not briefed |
bluesky bluesky-ai | 27d | We don’t allow corporations to come up with new nuclear weapons or biological pathogens. If a product is potentially dangerous, it should… | We don’t allow corporations to come up with new nuclear weapons or biological pathogens. If a product is potentially dangerous, it s… | ↗Not briefed |
rss techmeme | 27d | Google says it has 400+ publications in its News AI pilot, which tests features like overviews; participants include Financial Times, CNN, and The Texas Tribune (Andrew Deck/Nieman Lab) | Andrew Deck / Nieman Lab: Google says it has 400+ publications in its News AI pilot, which tests features like overviews; participants… | ↗Not briefed |
bluesky bluesky-ai | 27d | Now wondering if my autocorrect changed Claude to Clause. Is this the AI turf wars? | Now wondering if my autocorrect changed Claude to Clause. Is this the AI turf wars? | ↗Not briefed |
arxiv arxiv-ai | 27d | Cultural (In)Competence: Investigating Culture-Specific Stereotype Detection Capabilities of LLMs | Stereotypes are subjective constructs that are mediated by the cultural context in which they are viewed. Successfully modeling them t… | ↗Not briefed |
bluesky bluesky-ai | 27d | I disagree with the people who think that everyone will just use AI to make bespoke software for themselves, though I do think the business… | I disagree with the people who think that everyone will just use AI to make bespoke software for themselves, though I do think the bus… | ↗Not briefed |
hn hn-ai | 27d | Show HN: Stenographer – Shell History Search | — | ↗Not briefed |
hn hn-ai | 27d | Push-Based KV Transfer to Improve TTFT, TPOT for Disaggregated Inference in vLLM🔧 vLLM | — | ↗Not briefed |
bluesky bluesky-ai | 27d | Representatives from OpenAI and Meta cautioned a U.S. judiciary committee about the growing difficulty of detecting AI-fabricated media. ww… | Representatives from OpenAI and Meta cautioned a U.S. judiciary committee about the growing difficulty of detecting AI-fabricated medi… | ↗Not briefed |
bluesky bluesky-ai | 27d | I'm just kind of confused they're paying Anthropic by the token by default outside of contracts that need No Data Retention. Seems like the… | I'm just kind of confused they're paying Anthropic by the token by default outside of contracts that need No Data Retention. Seems lik… | ↗Not briefed |
arxiv arxiv-ai | 27d | On the SPOT: Enhancing Speculative Decoding for Diffusion Language Models via Commit-Order Prediction | Diffusion language models (DLMs) promise parallel text generation, but in practice face an unresolved speed-quality trade-off. Exact d… | ↗Not briefed |
bluesky bluesky-ai | 27d | Eight U.S. states are suing OpenAI, saying its hurting kids and lying to adults | Eight U.S. states are suing OpenAI, saying its hurting kids and lying to adults | ↗Not briefed |
arxiv arxiv-ai | 27d | HeRE: A Novel Benchmark for Hebrew Relation Extraction via LLM-Guided Denoising of Knowledge Graph Alignments🔧 GPT-5 | While LLMs demonstrate strong capabilities in English Relation Extraction (RE) tasks, a significant gap remains in assessing their RE… | ↗Not briefed |
bluesky bluesky-ai | 27d | My dog is conscious, your LLM is not! | My dog is conscious, your LLM is not! | ↗Not briefed |
bluesky bluesky-ai | 27d | Hot take: the "if LLMs were conscious using them would be slavery" argument is wrong because there's no analogue of coercion in the operato… | Hot take: the "if LLMs were conscious using them would be slavery" argument is wrong because there's no analogue of coercion in the op… | ↗Not briefed |
hn hn-ai | 27d | What 1,866 popular GitHub projects tell their AI coding agents | — | ↗Not briefed |