Skip to content

Comment on Scaling long-running autonomous coding

Comments

You can tell they're not just wrapping an existing rendering engine because of those very obvious rendering glitches, but the pages are legible and look mostly correct.

Honestly expected more from a sw article. How is this proof, when it could easily be the agents including packages to do the heaving lifting and then screwing it up on top?

Show me the code that's using other packages for the heavy lifting of implementing the layout engine.

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.