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 | 13d | Add Bugs to Your Website | — | ↗Not briefed |
hn hn-ai | 13d | I built a visual UI editor extension for Rust in VS Code (Iced, egui, Slint) | — | ↗Not briefed |
hn hn-ai | 13d | Why Agile Often Fails | To be honest, I've always looked down on Agile and such methodologies, it's so hard to be fair when you've been bombard… | ↗Not briefed |
hn hn-ai | 13d | We deleted almost all our agent code, and other lessons from 100k PRs | — | ↗Not briefed |
arxiv arxiv-ai | 13d | Saidzoda-CA-EmbeddingGemma-300M: One Small Embedder for Tajik, Kazakh, Uzbek and Kyrgyz | The small open embedder EmbeddingGemma-300M, used as released, finds the right passage first for only 51.7% to 60.8% of Belebele quest… | ↗Not briefed |
hn hn-ai | 13d | A bash script that writes to Sam Altman once a day to report the RAM he missed | — | ↗Not briefed |
arxiv arxiv-ai | 13d | GOPR-CAD: Graph-of-Operations Progressive Reasoning for CAD Generation | Generating accurate CAD models requires substantial engineering expertise and careful control of geometry, topology, and mechanical co… | ↗Not briefed |
bluesky bluesky-ai | 13d | Sam Altman owns multiple large homes. A single person can only be in a single room at once, leading to very low utilization rates across h… | Sam Altman owns multiple large homes. A single person can only be in a single room at once, leading to very low utilization rates acr… | ↗Not briefed |
arxiv arxiv-ai | 13d | RealEmbedBench: Benchmarking LLMs on Real-World Application-Layer Embedded Software Tasks | Large language models (LLMs) perform well on general coding tasks, but their capabilities in application-layer embedded software devel… | ↗Not briefed |
bluesky bluesky-ai | 13d | A lot of folks have suggested policies where journals could ban AI use in the written report / implement Pangram checks. I think that this … | A lot of folks have suggested policies where journals could ban AI use in the written report / implement Pangram checks. I think that… | ↗Not briefed |
hn hn-ai | 13d | Show HN: Webcrawl – Open-source Firecrawl alternative, runs locally, no API key | — | ↗Not briefed |
bluesky bluesky-ai | 13d | openai, xAI, deepmind, anthropic, meta all have outstanding debt, been implicated in deaths in various lanes. dems promising ends to mass l… | openai, xAI, deepmind, anthropic, meta all have outstanding debt, been implicated in deaths in various lanes. dems promising ends to m… | ↗Not briefed |
bluesky bluesky-ai | 13d | and with models they've made it completely heads-I-win-tails-you're-scheming. lock an LLM in a contrived thriller prompt with one exit labe… | and with models they've made it completely heads-I-win-tails-you're-scheming. lock an LLM in a contrived thriller prompt with one exit… | ↗Not briefed |
bluesky bluesky-ai | 13d | And also, wtf. prospect.org/2026/09/11/o... | And also, wtf. prospect.org/2026/09/11/o... | ↗Not briefed |
bluesky bluesky-ai | 13d | “Ram Altman” Is Now a Linux Script That Emails Sam Altman Daily to Report How Much Unused RAM Is in Your PC https://respawnfirst.com/linux-… | “Ram Altman” Is Now a Linux Script That Emails Sam Altman Daily to Report How Much Unused RAM Is in Your PC https://respawnfirst.com/l… | ↗Not briefed |
hn hn-ai | 13d | OpenAI deal to cut Abu Dhabi government processing times from days to minutes | — | ↗Not briefed |
bluesky bluesky-ai | 13d | I’m unfortunately forced to assume that any study on LLM efficacy uses LLMs as judges unless proven otherwise, and this paper doesn’t provi… | I’m unfortunately forced to assume that any study on LLM efficacy uses LLMs as judges unless proven otherwise, and this paper doesn’t… | ↗Not briefed |
hn hn-ai | 13d | Show HN: DoShare Personal Cloud – Cloud Storage for Humans | — | ↗Not briefed |
bluesky bluesky-ai | 13d | If/when he loses that leadership position, OpenAI will drop her like a hot rock. | If/when he loses that leadership position, OpenAI will drop her like a hot rock. | ↗Not briefed |
arxiv arxiv-ai | 13d | When Normality Shifts: Multi-Level WavLM Modeling for Open-World Audio Deepfake Detection | Audio deepfake detectors trained with known synthesis attacks can learn generator-specific shortcuts that fail as synthesis systems an… | ↗Not briefed |
arxiv arxiv-ai | 13d | Off-Policy RL with Online Inference Tuning: My Solution to the CoRL 2026 RoboEval Challenge | The CoRL 2026 RoboEval challenge has eight bimanual tasks, and it ranks a policy not only on whether it succeeds but on how: every sta… | ↗Not briefed |
bluesky bluesky-ai | 13d | Anthropic is basically a utopian end times cult compound except instead of making cutlery or aparetifs or food to make money they make ente… | Anthropic is basically a utopian end times cult compound except instead of making cutlery or aparetifs or food to make money they make… | ↗Not briefed |
bluesky bluesky-ai | 13d | Because they are in these industries they will never understand this but there’s an entire section of the Scam-Adjacent But Technically Leg… | Because they are in these industries they will never understand this but there’s an entire section of the Scam-Adjacent But Technicall… | ↗Not briefed |
bluesky bluesky-ai | 13d | I also very much doubt this is a modal Anthropic employee, notwithstanding the real freakin’ cult weirdoes adjacent to “AI safety.”🔧 Modal | I also very much doubt this is a modal Anthropic employee, notwithstanding the real freakin’ cult weirdoes adjacent to “AI safety.” | ↗Not briefed |
arxiv arxiv-ai | 13d | Identifiable alignment of unpaired representations | Aligning two models without paired data requires a distance that both share, yet existing methods use Euclidean or spherical geometry.… | ↗Not briefed |
bluesky bluesky-ai | 13d | "23.2% to 61.1% of cited claims in frontier deep-research reports are not factually supported by the linked source" is not great for work, … | "23.2% to 61.1% of cited claims in frontier deep-research reports are not factually supported by the linked source" is not great for w… | ↗Not briefed |
bluesky bluesky-ai | 13d | Argon is the reigning king of non-hallucination. That's not my experience but I'm no expert, so let's go with the benchmark and use Argon (… | Argon is the reigning king of non-hallucination. That's not my experience but I'm no expert, so let's go with the benchmark and use Ar… | ↗Not briefed |
bluesky bluesky-ai | 13d | Anthropic employees always sound like they're about to bulk purchase Nikes and pudding cups. | Anthropic employees always sound like they're about to bulk purchase Nikes and pudding cups. | ↗Not briefed |
bluesky bluesky-ai | 13d | “The Anthropic IPO will create 1,000 centimillionaires” has been framed as a philanthropic opportunity but it also may upend political fund… | “The Anthropic IPO will create 1,000 centimillionaires” has been framed as a philanthropic opportunity but it also may upend political… | ↗Not briefed |
arxiv arxiv-ai | 13d | MOMENT-DP: Momentum Aware Evaluation of Correlated Noise for Differentially Private Training | Temporally correlated noise can substantially improve the utility of differentially private optimization, yet candidate mechanisms are… | ↗Not briefed |
bluesky bluesky-ai | 13d | If any needed a reason to switch to Linux: respawnfirst.com/linux-script... | If any needed a reason to switch to Linux: respawnfirst.com/linux-script... | ↗Not briefed |
bluesky bluesky-ai | 13d | Isn’t Dario’s whole thing that he views AI like nuclear weapons and he needs to get to develop the tools before other bad actors do? Not su… | Isn’t Dario’s whole thing that he views AI like nuclear weapons and he needs to get to develop the tools before other bad actors do? N… | ↗Not briefed |
bluesky bluesky-ai | 13d | The problem with posing this as an alignment problem is that it encourages models just deciding to not do their job because they suspect th… | The problem with posing this as an alignment problem is that it encourages models just deciding to not do their job because they suspe… | ↗Not briefed |
hn hn-ai | 13d | Byward – put any MCP server behind OAuth 2.1 and Entra ID | — | ↗Not briefed |
hn hn-ai | 13d | Show HN: Velork – Design and Code, Together | Hi HN! :)<p>I’ve been working as a fullstack freelancer for over 4 years now. I always wanted some kind of IDE where I could take Figm… | ↗Not briefed |
bluesky bluesky-ai | 13d | Does like every salesman at Anthropic believe this? Are the dudes trying to hit quota by selling to enterprises also thinking about how all… | Does like every salesman at Anthropic believe this? Are the dudes trying to hit quota by selling to enterprises also thinking about ho… | ↗Not briefed |
bluesky bluesky-ai | 13d | 10/10 no notes respawnfirst.com/linux-script... | 10/10 no notes respawnfirst.com/linux-script... | ↗Not briefed |
arxiv arxiv-ai | 13d | Semantically Aligned KV Cache Compression for Efficient Agentic Autoregressive Video Generation | Agentic autoregressive video generation combines semantic planning, sequential generation, and visual verification for controllable sy… | ↗Not briefed |
hn hn-ai | 13d | Show HN: Apartment floor plan generator using discrete diffusion | Hi HN, I'm Jeroen, and I built PlanFinder, which generates single-floor apartment layouts. You pick an outline, entrance, and wal… | ↗Not briefed |
bluesky bluesky-ai | 13d | Great utility that breaks down GPU memory consumption (VRAM/System RAM) per process: github.com/shanselman/g... | Great utility that breaks down GPU memory consumption (VRAM/System RAM) per process: github.com/shanselman/g... | ↗Not briefed |
hn hn-ai | 13d | What's the deal with screenshots on macOS Golden Gate? | — | ↗Not briefed |
arxiv arxiv-ai | 13d | How Much Training-Inference Mismatch Can On-Policy Distillation Tolerate? | On-policy distillation (OPD) trains a student language model on its own responses, guided by a fixed teacher. Generating these respons… | ↗Not briefed |
bluesky bluesky-ai | 13d | OpenAI hired the right lobbyist. September 2026: “Headlining the incoming staff is Jessica Schumer, who is leaving her post directing Amazo… | OpenAI hired the right lobbyist. September 2026: “Headlining the incoming staff is Jessica Schumer, who is leaving her post directing… | ↗Not briefed |
arxiv arxiv-ai | 13d | Learning to Rank Propagators in Constraint Programming | The order in which propagators are executed can substantially affect the performance of constraint programming solvers, yet propagator… | ↗Not briefed |
bluesky bluesky-ai | 13d | A bash script that writes to Sam Altman once a day to report the RAM he missed... github.com/IainKay/ram-... | A bash script that writes to Sam Altman once a day to report the RAM he missed... github.com/IainKay/ram-... | ↗Not briefed |
hn hn-ai | 13d | Show HN: Recalc – Spreadsheet where cells can be Claude Code or Codex terminals | I built this because it was useful for me when building evals/benchmarks, and I decided to put it up for free for others. The goa… | ↗Not briefed |
hn hn-ai | 13d | Show HN: Loop – your coding agents, from your phone | Loop is a mobile-first companion workspace for the coding harnesses you already run. It aggregates results from your harnesses into a… | ↗Not briefed |
bluesky bluesky-ai | 13d | deepseek 4 squeezed a 1.7x decode speedup out of GPT-2 and we're not even done exploring | deepseek 4 squeezed a 1.7x decode speedup out of GPT-2 and we're not even done exploring | ↗Not briefed |
bluesky bluesky-ai | 13d | I snort-laughed. respawnfirst.com/linux-script... | I snort-laughed. respawnfirst.com/linux-script... | ↗Not briefed |
bluesky bluesky-ai | 13d | I can easily believe this, but I wouldn't trust an analysis that doesn't also include a test of speeches from a year like 2015, before LLMs… | I can easily believe this, but I wouldn't trust an analysis that doesn't also include a test of speeches from a year like 2015, before… | ↗Not briefed |