今日 AI 热榜

All Systems Go: Global PC App Ecosystem Ready for NVIDIA RTX Spark

Anthropic says it cut Claude reasoning token theft by up to 99% in fight against industrial-scale distillation

Anthropic study warns that small AI models could soon replicate across everyday devices

Apple launches HomeView, its first smart display for the home

AI models can provably hide arbitrary capabilities

Meta releases its first frontier model, Muse Wonder

Amex introduces a framework to prevent AI agents from making unauthorized purchases

Alibaba Qwen Introduces QwenGyre: Elastic Reinforcement Learning RL for XLong-Horizon Agents

Open models have real benefits. Do they outweigh the risks?

Meta's new frontier model Muse Wonder Full breaks into the AI top tier

AI labs are buying data from dead startups to train their models

Z.ai GLM-Legal Beats Claude Fable 5 and Kimi K3 on Harvey LAB

From Earth to Orbit, France’s Ecosystem Delivers Agentic Computing at Scale With NVIDIA

Anthropic's Claude AI submitted false murder tip to Philadelphia police

'AI' Is Not a Tool by Rich Hickey

Sakana AI's self-improving model gets better at solving and judging, but can't catch its own blind spots

AI coding agents generate more code, but not more software

I turned my unused Kindle into a photo frame with GPT-6 Astra

AI is a tool and using it well takes skill

Intel D270 GPU AI Accelerator–Up to 480 GB LPDDR5X

DLSS 5 gets toned down for its Hogwarts Legacy debut next month

SpaceXSI bundles X Premium, Grok and Cursor into one subscription

Epic Games reportedly pivoting to "AI first" game development model

Hello Python 3.15, goodbye 3.10, and now I can use 3.11

Accessing OpenAI GPT models on Amazon Bedrock with global cross-Region inference in Thailand, Malaysia, Singapore, Indonesia, and Taiwan

Meta has some Wonder Full new AI models

TikTok’s Chief Sees AI Unlocking Science Within Three Years

TechCrunch Disrupt 2026: Day 1

Russia was not ready for the data center war it started with Ukraine

WindBorne wins a hefty US Air Force contract for its balloon data

The final Disrupt Stage lineup: Three days of conversations you won’t hear anywhere outside of TechCrunch Disrupt 2026

1M-token, multimodal decision models are here! But there is more to this one. You can classify, extract, and summarize with 1M tokens of text and imag...

Great work from @zwljohnny, @tri_dao, and @NVIDIA on ReplaySSM! Mamba2 standard decode and FlashInfer-backed MTP support are already in vLLM, reducing...

reading between the lines of how AI researchers talk about the future of LLMs, especially the ones who seem genuinely spooked, and I have a theory. I ...

Jensen just entered the stage. Very interesting speech about: - era of AI agents - DGX spark - why it took 4 years for Nvidia to work on the collabora...

Recommended NVIDIA paper on memory-efficient post-training RL. A very interesting technique for scaling RL for LLMs. The main takeaway is that you can...

GPT-4 was the undisputed best model in the world for a full year and we all debated whether anyone could catch up or if OpenAI had some secret formula...

🍉 MUSE WONDER 🍉

Ah cool - A new video model from Vidu, Q4. I'm always happy to see a new video model arrive and first impressions are very good. Lol - there was a typ...

Had some early access to Muse Wonder, but did not get a chance to use it a ton, yet it seems like a good model from limited test Perhaps more impressi...

BudgetPix Compute-Adaptive Tokenization for Pixel-Space Image Diffusion paper: https://huggingface.co/papers/2610.12307

TIL that after Garry Kasparov lost his 1997 match against Deep Blue, he accused the IBM team of having cheated by secretly using human expertise

My AI agent doing proactive customer service: Someone emailed me that, after subscribing to my Substack, they were automatically signed up for others ...

Tried out Muse Wonder on a tricky task (figuring out which parts of the OpenHands Agent Canvas are hard for users to use) and it did a great job, impr...

apparently visual coding is the social media meta 😅 3D demo PT sprint incoming

People take the Artificial Analysis Index and its individual scores way too seriously.

T3 Code has made evaluating new models so much easier (and more fun!)

Update I finally did it. I’m earning interest and my mom is proud

i think my chatgpts broken whenever i ask it anything it just makes a slot machine

why does astra provide such incredibly long explanations when i ask it to explain something? the value of an explanation is *lower* the longer it is. ...

did some spring cleaning on my personal gpu server, and decided to finally clean up the opt weights @suchenzang end of an era