Comment by Aldipower

2 days ago

It is a task that could be _easily_ done manually in much shorter time without AI, probably by developers who even love to develop. The reaction on this shouldn't be misjudged as anti-AI. A lot of people, including me, just do not get it! For scientific purposes? Ok, fair enough. But what is the further meaning of this exercise?

The point is that if we agree that this task is truly a one shot, as long as you agree it’s faster to prompt than code, then while you “easily” do this task in around an hour (or however long you say it will take you), I’ll prompt Claude in around 5 minutes, and get a few more things done while I let it run in the background. What am I missing from your argument?

  • Reading the blog post, prompting Claude setting up Playwright etc. takes at least one hour maybe more? Not seeing where your 5 minutes coming from.

    • author here -- it took like 5 minutes of actual attention from me? I'm not sure why you are counting reading the blog post or setting up playwright. I guess I did read the blog post, but im not sure that should count. And claude set up playwright, not me.

      1 reply →

    • “Setting up playwright” is about two sentences of a prompt to Claude. I know this because I’ve done it many times. The authors prompt in the post is only a few hundred words. (Most of the post is just LLM output.) I’m certain I could type that out in 5 minutes.

      Do we really need another post where I time how long it takes to prompt Claude to create the Space Jam website?

      6 replies →

    • When I did it, vanilla Claude (Max5, default Opus 4.5, no special skills loaded, etc) had playwright up and running for screenshots in minutes (after intervening to tell it to use python 3.13 after noticing python 3.14 seems to be missing a lot of wheels and uv was rebuilding numpy for some reason) without me telling how to screenshot at all.

      https://news.ycombinator.com/threads?id=fluidcruft#46185996

      (One fun wrinkle I enjoyed watching was I created the target screenshot in Firefox, and Claude was using playwright with Chrome. Ultimately, I have no idea whether either Firefox or Chrome has the correct actual fonts and I'm not a webdev and don't remember how to figure that all out)

For me it's more that I'm not a web developer and it would definitely take me way longer to research all the parts of doing this. I have booksmarts (at best) about basic CSS and have given up trying to keep up with javascript anything.

It's seemingly an experiment to see how an LLM performs when the task is just outside of its milieu. The answer is not very well.