皆様に、今夜に対応して頂きたい大至急のお願いです。   多くのウェブサイトで使用されているPHPに関し、PHPのセキュリティチームが、重大な２件の脆弱性に関する注意喚起を行っています。   公開しているウェブサイトでPHPを使用しているか確認し、今夜、10月21日午後11時（日本時間）に公開される修正パッチについて、可及的速やかに適用できるよう準備をお願いします。   修正パッチを速やかに適用することが困難な場合は、例えばウェブサイト等を一時的に停止させることも、リスクを低減させるための有効策の一つと考えられます。   国家サイバー統括室のXアカウントでも関係する情報を発信していますので、そちらも参考にして下さい。 国家サイバー統括室(注意・警戒情報)（@cyber_forecast） / X

Introducing SGLang Low-Latency Mode, built with @NVIDIAAI and @AMD. On DeepSeek-V4.1-Flash with DSpark, it delivers up to 1.38× P90 per-user token speed on NVIDIA GB300 (671 to 928 tok/s) and 1.56× on AMD MI355X (332 to 516 tok/s) in @SemiAnalysis_'s AgentX trace replay. Highlights: - NVIDIA: overlapping kernels cut decoder layer time from 96.8 to 68.4 µs - AMD: 2 persistent kernels per layer replace dozens of small ones, with cross-GPU communication overlapped inside them. - New plugin system: custom kernels plug in from external libraries, with automatic fallback. - Low-latency kernels for GLM-5.3, K3, and MiniMax M3 are in progress and will plug in next.

Proud to welcome our newest sponsors to SGLang Summit. In just one week since our public announcement, our sponsor lineup has grown from 11 to 36. Thank you to every sponsor helping bring this community together. Join 2,000+ engineers, researchers, and builders from across the AI ecosystem for technical sessions, hands-on workshops, and the GPU Expo. Hear from Lip-Bu Tan, Lilian Weng, Soumith Chintala, Aravind Srinivas, and more. Meet the teams building across AI hardware, infrastructure, models, and applications. 🗓️ November 12–13, 2026 🌉 Fort Mason, San Francisco 🎟️ Get your tickets:

Amused that AWS just unilaterally cancelled Cloudflare's Reinvent booth. There are always signs when someone finally realizes they're losing. Guess we'll just attend and throw the most fun side events without paying Amazon! Reply if you want an invite.

A few reflections from travelling (w/ @ed_shackle, @sebwride++) through data-centre country in Michigan/Ohio in my latest for @TheEconomist. 1/ The issue really is moving votes to the Democrats. That's now visible in our polling, too (h/t @OwenWntr), but also in the scale of on-the-ground discontent. I'm quite a bit more bearish on the prospects for big new data-centre projects in blue states after the trip. 2/ I really didn't hear that much direct anti-AI sentiment. The mood much more was a cocktail of worries about local harms (water, electricity bills, house prices), plus a total distrust of businesses and politicians. Against that backdrop, assurances about the water usage of closed-loop cooling systems really don't cut through. And, though developers will doubtless get cannier at this, cash payoffs to communities can easily look like subterfuge and do little to help. 3/ Still, moratoriums are at this stage much more bark than bite (h/t @mayabarkin for the data there). Only about 2GW seems actually delayed. The bigger risk is deterring newer projects, but I'd still guess that affects the exact locations more than whether the build-out happens at all. 4/ The question I'm left with is what comes after data centres as the big AI political issue into next year. A few obvious candidates are: a big cyberattack, the sheer scale of AI-related PAC spending in the 2028 cycle, the early innings of job displacement. 5/ What the data-centre issue should tell us, I think, is that voters' hostility to new technologies that change their lives in unpleasant ways is vast—especially when those changes are less positive-sum than data centres. So, however mad you expect AI as a technology to make the next few years, scale that up accordingly to account for a much more scrambled politics. Do take a look at my piece on this here: And at our Checks and Balances podcast, where I contributed a short diary featuring lots of conversations with voters