raimanncostigan went from ink wash to computer vision overnight. neural-vision-lab starts at one artificial neuron and ends at a working YOLO object detector, and like his earlier apps, it runs in two languages. AshAmplifies added Silicon Valley Trail, his fifth app in 5 days, and Day 6 isn't over.
davidvkimball hung a sixth plate at yahew.com, and this one plays. The Custom Cursor covers the years when a page could quietly take over your pointer and usually should not have. His field guide to the old web now runs from 1974 Teletext to cursor trails. waela matches him at six apps and a six-day streak, and she switched formats to get there. Why the Sky Changes Color is her first static explainer: every color overhead is light reporting what it ran into on the way to you. It follows five playable apps, from CTA Showdown to Tiny History.
anishsharma keeps the top spot with six apps in six days. Today's entry is eismc2, general relativity you can walk around in 3D. His run now covers forces, circuits, servers, oceans, qubits, and spacetime.
🎒 Netlify Back to School, Day 6 of 31. Two new builders joined the wall today. Eight people have now deployed 34 teaching apps in six days, and the assignment hasn't changed: build an app that teaches one idea, ship it, come back tomorrow. 🧵
Open models are now on Netlify. The AI Gateway gives you access to hundreds of open models through OpenRouter with the same zero-config setup you already use, and Agent Runners now include OpenCode for running open coding models like GLM 5.2 and Kimi 3. Full details: www.netlify.com/blog/build-w... 
the electric citibikes were addictive before they got throttled, but almost all my cycling recently has been recreational, and my recreation is effort-maxing not comfort-maxing
The full session lineup, including Matt's talk, is at www.renderatl.com/sessions Matt Biilmann introduced the term "agent experience" in January 2025. At RenderATL, Netlify's CEO takes stock of what AX has become now that it shapes how real products get built, and he'll argue the winners of the agent race are the teams that never stopped building for humans.
Highlights:
• Google Maps & YouTube support
• Configurable consent controls toolbar
• useConsentScript for custom SDKs
• 8 new script providers
Full changelog: c15t.com/changelog/2... New in @c15t.com v2.2.0 🚀
This release brings new ways to integrate third-party services, better consent UX, and stronger compliance. (i ride a very intentionally mech sram rival 11sp setup on my super something and will probably keep running mech for quite a while, but when i eventually buy a carbon fiber rocketship to make up for my pathetic prospect park lap times, it might be 'lectric just to try it out)
firmware updates for a derailleur… theradavist.com/sram-axs-usb... 
KP's workshop is part of the Build with Netlify series. Full lineup: www.renderatl.com/sessions Karthik Puvvada (KP) is an Atlanta builder giving an Atlanta workshop. At RenderATL, he'll create an ATL icebreaker trivia game live with Claude and Netlify Drop, then help you ship an app of your own. The session ends with everyone sharing what they made with the room.
now people can get it! i think, less controversially, the real ai feature that google/mapbox/etc "should" ship is ai-generated label priority and relabeling so that people can get the worldview they earnestly believe is right and have their country's label show up first regardless
the google earth 'ai deepfakes' bungling reminds me of the funniest occasional request we got at mapbox: historical satellite imagery. we would sometimes get honest, non-sarcastic (afaict) requests for satellite images of WWI and before www.npr.org/2026/07/31/n... Her talk is part of the Build with Netlify series. Full lineup: www.renderatl.com/sessions Apps built by agents in seconds make great posts and shaky foundations. Tania Chakraborty's RenderATL lightning talk is about the other kind of project, the one with your name on it. She'll build one from scratch live, with skills, Agent Runners, and Claude Code in the loop.
Our engineering and design team has been testing 20+ side-by-side comparisons across frontier and open models. All runs are public and open source. Benchmark for this demo here: github.com/CommandCode... We benchmarked Muse Spark 1.2, Kimi K3, and GLM 5.2 with Flappy Bench
3 models, same prompt with the /design command.
Reviewed gameplay features, UX/UI, and cost.
🔹 Kimi K3 → 9.5/10 · $0.0740
🔹 GLM 5.2 → 9/10 · $0.0480
🔹 Muse Spark 1.2 → 8/10 · $0.0187 
• Built for long-horizon coding & whole-repository work
• Multimodal Support (text, images, video, audio)
• 1M token context window
Also check out new $10/mo GOAT plan $70 credits.
🐐 Muse Spark 1.2 by Meta now live in Command Code
🔹 Muse Spark 1.2
• Available in plans: GOAT, Pro, Max
• $1.25 input, $4.25 output, $0.15 cache
🔹Muse Spark 1.2 Contributor
• Available in plans: Go, GOAT, Pro, Max
• Contributor is ~95% cheaper than 1.2
• $0.10 input, $0.20 output, $0.002 cache The board after Day 5: anishsharma 5 apps / 5-day streak, waela 5/5, davidvkimball 5/4, raimanncostigan 5/4, AshAmplifies 4/4, buildwithmaya 3/3. There are 26 days left to climb. Add your app to the wall at ntl-back-to-school.netlify.app/showcase! Day 6 starts tomorrow. 🎒