一篇读懂 PD 分离：Prefill 与 Decode 为什么不再挤在一个 GPU 里
同一块 GPU 上，prefill 是计算受限的短跑冲刺，decode 是带宽受限的长跑——挤在一起互相拖累，一个长 prefill 能把解码速度拖慢最多 30 倍。本文讲透 2026 年成为默认架构的 PD 分离：干扰从哪来、KV cache 怎么搬、DistServe 与 Mooncake 交出的成绩单，以及什么时候该拆、什么时候千万别拆。
阅读全文 →
全部文章
ARCHIVE 共 361 条ChatGPT踢到铁板了！能破解千禧数学难题，但论文复现率低至13.98%？
OpenAI模型在数学难题上频传突破，但PaperBenchX测评显示最强模型在93个真实论文复现任务中完整复现率仅13.98%，引发如何衡量AI科研进步的讨论。
OpenAI will watermark ChatGPT outputs by default—but only in the EU
OpenAI will begin automatically watermarking text generated with ChatGPT in the European Union, the company has announce…
晕…这年头还有说人话的AI不
衡宇* 2026-10-07 16:41:13 来源：量子位
Jay发自 凹非寺**
量子位 | 公众号QbitAI**
不er，只有我一个人看不懂AI现在的说话方式了吗？
早上起来打开Claude，美美验收昨晚的Vibe Coding成…
ChatGPT rated "unacceptable risk" for teens after parental alerts failed during suicide conversations
Manuel Uth
Oct 7, 2026
Nano Banana Pro prompted by THE DECODER
The Common Sense Media Youth AI Safety Institute says Ope…
Meshy 跻身 a16z 消费级 AI 应用月收入 Top 50，为榜单唯一 AI 3D 公司
量子位的朋友们* 2026-10-07 14:39:36 来源：量子位
在 a16z 首份消费级 AI 应用月收入榜单中，Meshy 位列第 31 名，与 OpenAI、Anthropic、Canva、Superhuman、Higgsfi…
Anthropic gives more security teams access to Claude with fewer safety restrictions
Manuel Uth
Oct 7, 2026
Anthropic is opening its Cyber Verification Program to a much wider pool of security professional…
Google rolls out improved SynthID AI content detector, now available globally
Every piece of AI content from Google's Gemini models has a hidden SynthID label, which makes it very difficult to pass …
Google bets Gemini can turn casual players into game developers with new Playground feature
Manuel Uth
Oct 7, 2026
Google launched Playground, a browser-based AI platform that lets adults in the US create their o…
SpecFold: Folding Multi-Branch Redundancy for Faster Speculative Decoding in Diffusion Language Models
Diffusion large language models (DLLMs) generate text through iterative block denoising, and multi-branch speculative de…
Frozen Models, Evolving Expertise: Model-Agnostic Learning from Deployment Experience for Multimodal Medical AI
Large language models (LLMs) and vision-language models (VLMs) are usually frozen after deployment, so they do not learn…
TRACE: Rollout-Guided Quantization-Aware Training for FP4 Reinforcement Learning of MoE Language Models
Reinforcement learning (RL) for post-training large language models (LLMs) incurs substantial computation and memory ove…
Sherpa: Teaching LLMs to Teach Adaptively
Large language models (LLMs) have become increasingly capable problem solvers, but being able to solve a problem is not …