Sergei Garin 🇷🇺

Hear me out: We hated "waterfall" because after Big Upfront Planning, implementation took FOREVER, and by the time we were done, the plan was out of date But with AI, implementation is rapid (100x faster, or more, than before) So Big Upfront Planning can now make sense?

So apparently the React compiler is an amazing tool to help you rewrite your React apps to Solid

This is either a hot-take or painfully obvious: Telling your agents what you want >>>>>> telling your agents what to do

do not read books. do not watch videos. spend all of your time building stuff with ai. learn as much as possible (by asking ai).

What’s going to happen to open source? A prediction: Dependencies are over. The same Brotli vulnerability impacted .NET, Node, Bun, Envoy and more. Dependencies are too big of a target. When rewriting a dependency is one prompt away, why take the risk?

had someone opt out of our interview process because they couldn't use AI for every part of the assess all i can say to that is: if that is your mindset, you're going to be unemployed in the future we dont pay high salaries in this industry for people to prompt, we pay for domain expertise

The debate is over, and evil has won. `cursor: pointer` now means clickable. RIP my ability to tell whether ⌘-clicking something will open it in a new window. github.com/w3c/csswg-drafts/…

Another reason why you wanna use your own harness and not one where companies change what you can or can’t do at will.

А кто-нибудь из программистов, кстати, пытался уже завернуть себя в луп? - Используем Claude Routines (или вебхуки), чтобы достать новые issues из репозитория и ассайнить их на себя - Тут же последовательно запускаем агентов, чтобы все эти issue реализовывать и создавать PR - Мониторим автоматически активность в PR и последовательно итерируем/фиксим всякое - Мониторим активность в Slack — там нашему агенту уже можно и шутку пошутить, и мемасик отправить, и по работе ответить четко и быстро, чтобы создать образ максимального corporate warrior - При помощи LLM генерим отчетик для менеджера и создаем себе development goals на следующий квартал - Для онлайн-созвонов используем Dessa, который озвучивает то, что сгенерил LLM. Единственный минус — будет некоторая задержка между заданным вам вопросом и ответом со стороны вашего агента

We took a big leap: Bye bye React, and hello Remix 3! Working with the Remix team has been a blast, and this new version delivers exactly what we need: a more browser-native experience with less re-renders, hydration, javascript, and no implicit side effects. remix.run/

Stuff like this might finally kill ORMs, I’m incredibly hyped

Chrome 155 adds support for JPEG XL (.jxl) → goo.gle/4yYpSjW Expect 30-50% smaller file sizes than traditional JPEGs, native HDR, progressive rendering, and lossless JPEG transcoding. Under the hood, Chrome uses jxl-rs, a pure Rust decoder built for maximum memory safety and performance.

VIBECODERS Tell me your monthly AI bill Mine: - $120 Claude Code - $55 Grok Bot Be honest

We open sourced Void v1.1.11 and we made a music video for it. void.cloud

"It was too ambitious to build a linter and formatter so I told Claude to copy VoidZero" Congratulations!

Verification is one of the most important things for getting great results out of your agent. Clearly define "done" before code review starts. A reviewer, human or agent, asked to find bugs without this may still find something that could be improved. Give Claude checks it can run (bash commands, tests, clicking through the app in a browser) and have it pass that too before it says it's finished. When nothing blocks merge and ship.

Ultrafast is rolling out today for GPT-6.1 Sol in the API, Codex, and ChatGPT Work. Near-Astra intelligence at up to 8x faster speeds than Sol Standard, so you can build as fast as the ideas come.