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 | 13d | "Holy crap Lois! This is worse than the time the European Research Council used AI to turn me into a pile of poorly rendered boxes and then… | "Holy crap Lois! This is worse than the time the European Research Council used AI to turn me into a pile of poorly rendered boxes and… | ↗Not briefed |
arxiv arxiv-ai | 13d | Actions Should EARN Their Credit: Test-Time Adaptation and Route Consolidation for Frozen LLM Agents | LLM agents can use feedback from repeated attempts to improve their decisions without updating model parameters. However, positive act… | ↗Not briefed |
bluesky bluesky-ai | 13d | AI models that can be run locally and are open weight/open source are some of the best tech disability aids we've seen in years. Screen rea… | AI models that can be run locally and are open weight/open source are some of the best tech disability aids we've seen in years. Scree… | ↗Not briefed |
bluesky bluesky-ai | 13d | HMM…CORRUPTION IS RUNNING RAMPANT…HERE IS ANOTHER LOW COVERED STORY EXPLAINED BY @jscandid.bsky.social …SMH…A SPY ORGANIZATION IN THE U.S….… | HMM…CORRUPTION IS RUNNING RAMPANT…HERE IS ANOTHER LOW COVERED STORY EXPLAINED BY @jscandid.bsky.social …SMH…A SPY ORGANIZATION IN THE… | ↗Not briefed |
bluesky bluesky-ai | 13d | RE: the repost, so far I've managed to avoid using any GenAI/LLM tools at my development job and no one has really made a stink to me about… | RE: the repost, so far I've managed to avoid using any GenAI/LLM tools at my development job and no one has really made a stink to me… | ↗Not briefed |
rss theverge-ai | 13d | Apple is racing to make Siri more agentic | Bloomberg's Mark Gurman says CEO John Ternus is pressuring AI czar Craig Federighi to accelerate the Siri roadmap and deliver a powerf… | ↗Not briefed |
bluesky bluesky-ai | 13d | Ich liebe das Internet: “Ram Altman” Is Now a Linux Script That Emails Sam Altman Daily to Report How Much Unused RAM Is in Your PC respa… | Ich liebe das Internet: “Ram Altman” Is Now a Linux Script That Emails Sam Altman Daily to Report How Much Unused RAM Is in Your PC… | ↗Not briefed |
bluesky bluesky-ai | 13d | Amazon Nova Pro v1 vs Gemini 1.5 Pro (2M Context): Which LLM API actually protects your SaaS gross margin at 10,000 DAU? Here is the live … | Amazon Nova Pro v1 vs Gemini 1.5 Pro (2M Context): Which LLM API actually protects your SaaS gross margin at 10,000 DAU? Here is the… | ↗Not briefed |
hn hn-ai | 13d | Show HN: Models Engineering Index | Mei—or, the Models Engineering Index—calculates language models' practical performance on engineering tasks. | ↗Not briefed |
bluesky bluesky-ai | 13d | Digging into this, it looks like it's something made by two people affiliated with the Internet archive and was contributed to by Claude, b… | Digging into this, it looks like it's something made by two people affiliated with the Internet archive and was contributed to by Clau… | ↗Not briefed |
bluesky bluesky-ai | 13d | Schumer is controlled opposition & Democrats will lose the 2028 elections if he isn't replaced. There are plenty of options other than Van … | Schumer is controlled opposition & Democrats will lose the 2028 elections if he isn't replaced. There are plenty of options other than… | ↗Not briefed |
hn hn-ai | 13d | My Local Coding Agent | — | ↗Not briefed |
arxiv arxiv-ai | 13d | From Sound to Occupancy: Privacy-Aware Audio Sensing for Buildings | Occupancy-based controls regulate lighting and HVAC according to whether a space is in use. Passive infrared (PIR) sensors, the most c… | ↗Not briefed |
bluesky bluesky-ai | 13d | @MatthewBerman on YT! "RSI already started..." www.youtube.com/watch?v=lVbQ... #RSI #AI ##RSL #LLM #SOTA #Anthropic 11/8/2026 | @MatthewBerman on YT! "RSI already started..." www.youtube.com/watch?v=lVbQ... #RSI #AI ##RSL #LLM #SOTA #Anthropic 11/8/2026 | ↗Not briefed |
arxiv arxiv-ai | 13d | HydroTuring: A physics probe benchmark testing mass, energy, and momentum conservation in hydrologic models🔧 FLUX | Hydrologic models, including deep learning systems now deployed for operational flood forecasting, are evaluated primarily by how clos… | ↗Not briefed |
bluesky bluesky-ai | 13d | Desert Wheatears don't look like that! I assume this was edited with a LLM? ebird.org/news/threat-... | Desert Wheatears don't look like that! I assume this was edited with a LLM? ebird.org/news/threat-... | ↗Not briefed |
arxiv arxiv-ai | 13d | UTOPI: Efficient Egocentric Long-Video Understanding in AR via User-Guided Token Pre-Compression | As long-video understanding gains increasing attention, augmented reality (AR) devices are emerging as a compelling deployment platfor… | ↗Not briefed |
bluesky bluesky-ai | 13d | Well, Sports Illustrated shot its latest Swimsuit Issue at Marcia's Garden and, well... guess who they asked to model for 'em? 😎👙 #AInsa… | Well, Sports Illustrated shot its latest Swimsuit Issue at Marcia's Garden and, well... guess who they asked to model for 'em? 😎👙 #… | ↗Not briefed |
bluesky bluesky-ai | 13d | Last year I used AI to donate to Sherrod Brown. matthodges.com/posts/2025-0... | Last year I used AI to donate to Sherrod Brown. matthodges.com/posts/2025-0... | ↗Not briefed |
arxiv arxiv-ai | 13d | LiveAsset: Learning Intent Assets from Corrective Interactions for Agent Runtime Control | Large language model (LLM) agents acting across web applications and online services often need multi-turn corrections to align their… | ↗Not briefed |
bluesky bluesky-ai | 13d | we're at the "sir, it's only a model.. I said hop in" stage of AI development at this point | we're at the "sir, it's only a model.. I said hop in" stage of AI development at this point | ↗Not briefed |
rss techmeme | 13d | THORChain, which handled ~$1B of stolen Bybit funds in 2025 for $12M+ in fees, began moving the Binance loot despite FBI pleas, saying it can't stop transfers (Wall Street Journal) | Wall Street Journal: THORChain, which handled ~$1B of stolen Bybit funds in 2025 for $12M+ in fees, began moving the Binance loot desp… | ↗Not briefed |
bluesky bluesky-ai | 13d | "Hi Sam, Please don't be alarmed, but there is 8.6 GiB of RAM on my desk that isn't in a data centre. That's 56.2% of what's installed her… | "Hi Sam, Please don't be alarmed, but there is 8.6 GiB of RAM on my desk that isn't in a data centre. That's 56.2% of what's installe… | ↗Not briefed |
bluesky bluesky-ai | 13d | ON THE RADAR: Claude Code underperforms open-source rivals on consumer GPUs, contradicting expectations of frontier-model dominance. A 16GB… | ON THE RADAR: Claude Code underperforms open-source rivals on consumer GPUs, contradicting expectations of frontier-model dominance. A… | ↗Not briefed |
arxiv arxiv-ai | 13d | InfraRSI: Harnessing Recursive Self-Improvement for Training Infrastructure | While Recursive Self-Improvement (RSI) accelerates isolated algorithm and kernel synthesis, distributed LLM training depends on tightl… | ↗Not briefed |
hn hn-ai | 13d | Sketching software architecture with Svugy | — | ↗Not briefed |
bluesky bluesky-ai | 13d | Corruption Was the number one issue in exit polls Schumers daughter is a lobbyist for openAI prospect.org/2026/09/11/o... | Corruption Was the number one issue in exit polls Schumers daughter is a lobbyist for openAI prospect.org/2026/09/11/o... | ↗Not briefed |
arxiv arxiv-ai | 13d | Kimodo-Policy: Learning Kinematic Action Models for Humanoid Loco-Manipulation | Humanoid loco-manipulation requires continuous coordination of high-dimensional whole-body motion to realize task objectives. While vi… | ↗Not briefed |
hn hn-ai | 13d | Canada's AI translation tool keeps making hilarious mistakes | — | ↗Not briefed |
bluesky bluesky-ai | 13d | i have been accused of using ai to generate a patreon essay, with the attached screenshot as evidence. as though chat gpt or claude could e… | i have been accused of using ai to generate a patreon essay, with the attached screenshot as evidence. as though chat gpt or claude co… | ↗Not briefed |
arxiv arxiv-ai | 13d | When Should Diffusion Language Models Reconsider Their Commitments? | Masked Diffusion Models (MDMs) are well known for fast text generation via parallel decoding. However, independent parallel sampling c… | ↗Not briefed |
hn hn-ai | 13d | Craft: A Tendon-Driven Hand with Hybrid Hard-Soft Compliance | — | ↗Not briefed |
arxiv arxiv-ai | 13d | Not All Branches Are Equal: Adaptive Semantic ID Construction for Generative Recommendation | Generative recommendation models represent items as sequences of discrete tokens known as semantic IDs. Existing approaches derive the… | ↗Not briefed |
bluesky bluesky-ai | 13d | Liebe alles daran. | Liebe alles daran. | ↗Not briefed |
bluesky bluesky-ai | 13d | Bye Bye Silicon Valley | Bye Bye Silicon Valley | ↗Not briefed |
bluesky bluesky-ai | 13d | Sounds like they scaled up way too fast just assuming they can trust whatever the LLM spurts out, didn’t watch out for things like duplicat… | Sounds like they scaled up way too fast just assuming they can trust whatever the LLM spurts out, didn’t watch out for things like dup… | ↗Not briefed |
hn hn-ai | 13d | Show HN: LLM-cost-watch – Ask Claude why your LLM bill jumped | — | ↗Not briefed |
bluesky bluesky-ai | 13d | It helps to put rules in AGENTS.md. Early on an LLM (that we were required to use) was writing code with memory leaks. It didn't seem to … | It helps to put rules in AGENTS.md. Early on an LLM (that we were required to use) was writing code with memory leaks. It didn't see… | ↗Not briefed |
bluesky bluesky-ai | 13d | Lightweight AI spots forest fires fast 🔥 A new model called YOLOv8-CNR detects forest fires with higher accuracy and speed than standard … | Lightweight AI spots forest fires fast 🔥 A new model called YOLOv8-CNR detects forest fires with higher accuracy and speed than stan… | ↗Not briefed |
hn hn-ai | 13d | Show HN: Hackertube for Roku TV | Karpathy recently built Hackertube.<p>I watch a lot of Youtube on my TV and wanted to replace that with hackertube, so I built a roku… | ↗Not briefed |
bluesky bluesky-ai | 13d | It's The Guardian; they decided to let the picture desk go and replaced it with an LLM | It's The Guardian; they decided to let the picture desk go and replaced it with an LLM | ↗Not briefed |
hn hn-ai | 13d | Show HN: Elo Ratings for Tests, ODIs, and T20Is | — | ↗Not briefed |
arxiv arxiv-ai | 13d | Safeguarding Agentic AI in Radiology Workflows: A Multimodel Evaluation of Rule-based and Large Language Model-based Verification | Purpose: To estimate how often tool-using large language model (LLM) agents record clinically relevant false or unsupported informatio… | ↗Not briefed |
bluesky bluesky-ai | 13d | This is why Schumer refused to help Mamdani win. Schumer's daughter went from working for Amazon to Open AI. He choose his daughters job o… | This is why Schumer refused to help Mamdani win. Schumer's daughter went from working for Amazon to Open AI. He choose his daughters… | ↗Not briefed |
hn hn-ai | 13d | What Should We Learn? What LLMs Taught Me: Understanding Isn't Producing | — | ↗Not briefed |
bluesky bluesky-ai | 13d | Lol | Lol | ↗Not briefed |
arxiv arxiv-ai | 13d | PRISM: Causal Counterfactual Minimization of Agent Tool Privileges via Information Bottleneck Role Induction | This paper shifts the object of role engineering from human users to large language model (LLM) agents, and proposes PRISM to address… | ↗Not briefed |
arxiv arxiv-ai | 13d | Weight-Space Programming of Variational Quantum Circuits: Large Language Models Write Quantum Circuit Parameters Without an Optimizer | Variational quantum algorithms (VQAs) optimize circuit angles through repeated quantum-device evaluations, often facing vanishing grad… | ↗Not briefed |
bluesky bluesky-ai | 13d | Load bearing is auth boundary and retry semantics, not the server. I put one internal tool behind an MCP server and spent a day on scoped t… | Load bearing is auth boundary and retry semantics, not the server. I put one internal tool behind an MCP server and spent a day on sco… | ↗Not briefed |
arxiv arxiv-ai | 13d | Ostinato: Fast CPU-GPU MoE Inference for Local Agents via Resident Experts in Every Forward Pass | Language-model agents increasingly run on workstations with a few consumer GPUs, where large mixture-of-experts (MoE) models keep part… | ↗Not briefed |