Eli

@mrdoob I think it’s too early to tell. You have to ignore the majority of shiny demos and wait for normal users to leave and not return within 6 months. A lot of amazing open software will be created, but a lot already exists, and a lot of proprietary software will be created too

We seem to be witnessing the end of the proprietary software model.

@wesbos @mrdoob where have you been? It’s the year of the Linux desktop

@mrdoob Say more?

@AdamRackis const exercisesById = {}

"jr dev who types fast"

@SoofGolan Ban mocks

@jjenzz I feel a whole branch of web dev history (especially web components) can be attributed to ‘smart Google devs misunderstanding CSS’

early in my career (00s) i went to a web conference where Google gave a talk and at one point they explained why they had to use HTML tables for their image tab, “it’s the only way” etc. the next speaker got up and before even starting their talk, they corrected them and explained how to do it w/o tables in a room full of ppl. that awkward moment has stuck with me my entire career. a constant reminder that even the devs we might put on a pedestal are still figuring things out in real time.

@ac363218 @Yuchenj_UW @karthickdotxyz Even without AI, you can create a decent ‘clone’ of almost any SaaS product with a small team of good engineers + product people. Most SaaS products also have free/open-source alternatives. Yet enterprises spend millions on the paid alternatives. Why?

@Yuchenj_UW @karthickdotxyz Easy man, they can clone, they can reduce the price of the products They win the market

Some people at frontier AI labs told me they believe startups are over. OpenAI, Anthropic, Google, xAI will absorb every industry as AGI nears. Coding today, science, medicine, and finance next. Then everything else. If they’re right, that’s a pretty boring end of the world.

@cramforce @threepointone I don’t blame the author at all, and think you were wrong to assume bad faith. To me what’s worth calling out is that CF has no processes preventing a huge block of (afaict) non-human-reviewed code being merged into an important monorepo and published under their npm namespace

Reflecting on it all, I stirred up unnecessary pain for @threepointone, and I regret it. He very much didn't deserve that, and the community didn't need it. - He is one of the most welcoming, positive, uplifting people in the JavaScript community - He's a great listener and someone who leaves every room better than he found it - He's always team web rather than team $company I'm sorry Sunil.

@lysaurs @giuseppegurgone @DavidKPiano @threepointone Why do work-related experiments on your personal account? The main thing I’m surprised by is that such a huge PR (300k LOC) would be merged without a human review. But it also speaks to it being an experiment

@document_body @giuseppegurgone @DavidKPiano @threepointone Why experiment on the company account in this specific manner? Seems too convenient to be a mistake.

Vercel wrote an article against @threepointone and they really, really should not have

@giuseppegurgone @DavidKPiano @threepointone they definitely saw the PR and could have quickly resolved/discussed it with Sunil github.com/cloudflare/age…. So the public attack really feels like they used the opportunity to push a narrative

Sunil is the best and nicest guy. I wonder if they knew it was him who forked and was experimenting - probably wouldn’t have knowingly written an implicit attack and if they did it sucks bug time. I guess forking under Cloudflare and renaming the repo didn’t help - it wouldn’t be odd if vercel misunderstood. With all tha being said, zooming out I wouldn’t engage on these matters - from the outside it is a bad look (pr-wise) for both companies.

@steveruizok When would 1000s of them be receiving new positions each frame? I would think you could almost always avoid that by using relative positions, even if it means temporarily reparenting in the case of the case of multi selection etc

genuine question for react heads, which of these is most performant? Consider that we may have thousands of them all receiving new x y props on each frame.

I made an interactive blog post about how JPEG image compression works: sophielwang.com/blog/jpeg

Vercel’s Claude Code usage just spiked

@TheAlexLichter @yacinehmito @ubugeeei @liangmiQwQ @auvred_ That’s helpful but feels a bit like an understated side note, because in other places (like the official skill mentioned in the same article skills.sh/oxc-project/ox…), oxlint is pushed as a drop-in replacement for eslint in vue projects

@yacinehmito @ubugeeei @liangmiQwQ @auvred_ > Limited support for front-end frameworks' custom file formats (e.g. Svelte, Vue, Angular) - coming later this year. oxc.rs/blog/2026-03-1…

Frankly the lack of visibility regarding when or how Vue will benefit from the new sped up tooling (oxlint, tsgo) is concerning.

@wesbos a deno/bun killer built on node/vite called something like ‘vino’

👀 taking bets for what Vite/Void is releasing next

@boshen_c Why is the oxlint vue plugin marked as equivalent to eslint-plugin-vue? The ‘lint only the script’ approach doesn’t make sense and leaves a codebase in a worse state than it was before. It would be much much better to say that .vue (.svelte + .astro) are just not supported yet

/migrate-oxlint skills.sh/oxc-project/ox…

@AdamRackis I can’t remember the last time I had to manually update an import path after moving a file. Are you using vscode? #_update-imports-on-file-move" target="_blank" rel="nofollow noopener">code.visualstudio.com/docs/typescrip…

99% sure I could have just moved all those files, and adjusted the ~12 import paths faster than the 3 minutes 45 seconds it took my agent to do so. Starting to wonder if we might be overusing this shit just a lil

@bcherny Is ‘worktree agent’ a built-in concept?

Released today: /loop /loop is a powerful new way to schedule recurring tasks, for up to 3 days at a time eg. “/loop babysit all my PRs. Auto-fix build issues and when comments come in, use a worktree agent to fix them” eg. “/loop every morning use the Slack MCP to give me a summary of top posts I was tagged in” Let us know what you think!

@csuwildcat I ask it to commit atomically and frequently, and have a requirement that all tests must pass before commit. I review a branch in github or gitkraken, and ask it to ‘recreate this branch/commit/feature with x changes’

For people writing code with just the CLI variants of LLMs: do you really just yolo everything and not review the code it generates in an IDE? I just don't understand how one can skip visual review of the code, because I still catch it generating inaccurate outputs way too often.

@zoomyzoomm It’s always been possible to dl the frontend source of a site/app, modify it, swap the endpoints. You can get a proof-of-concept clone of most web apps in a few hours by using `wget -r`. The problem is you’re going to end up with a worse version of the thing you’re copying

He’s not kidding. Took me HALF AN HOUR to vibe code Notion with Perplexity Computer. Software is legit a zero.

@levelsio @Niklas_Sikorra Try the @vite_js beta release. Dev server updates are instant, builds take a few few seconds

@Niklas_Sikorra But it takes a minute? For me it takes 0 seconds

I never fully understood the build process because I've always coded without frameworks I always felt the faster I could see the code I just typed in action, the faster my feedback loop and the faster I can ship and improve my products Waiting even a minute to build would destroy that feedback loop and make me way slower I can do lots of mini edits and see each instead of batching them and then building!