I scraped the schedule for Open Sauce 2025 this morning and built an alternative schedule interface with the option to add everything to your calendar (via ICS)... working entirely on my iPhone, using OpenAI Codex and Claude Artifacts

I guess you could call this "vibe scraping"? OpenAI Codex turns out to be great at writing custom scrapers if you give it internet access and tell it to download and install Playwright

Prompts + transcripts: https://simonwillison.net/2025/Jul/17/vibe-scraping/

I continue to be entirely unafraid that these tools are going to obsolete my skills as a software engineer

In a truly beautiful cautionary tale about the dangers of vibe-coding on a phone, it turns out the schedule page I shipped weighed 130MB because it included unoptimized profile photos of all 170+ speakers!

I've ditched the images now, reducing the page weight to a much less horrifying 93.58 KB. https://simonwillison.net/2025/Jul/17/vibe-scraping/#update-i-removed-the-speaker-avatars

Vibe scraping and vibe coding a schedule app for Open Sauce 2025 entirely on my phone

This morning, working entirely on my phone, I scraped a conference website and vibe coded up an alternative UI for interacting with the schedule using a combination of OpenAI Codex …

Simon Willison’s Weblog
@simon did you also vibe code the performance check?

@simon Vibe coding was not the problem here, the original website is all gigantic PNGs.

(I fetched those results on desktop and they're super slow b/c I am on a public wifi, but switching to a mobile user-agent doesn't result in smaller thumbnails)

@neilk @simon Well vibe coding kind of was the problem here, insofar as it was done on a mobile phone. Simon says in his blog his phone had no access to performance profiling tools.
@jimgar @neilk with hindsight, I should have run something like PageSpeed Insights against the page to check before I shared the link with anyone!

@simon Are you not afraid that those skills will erode over time?

Being able to do a task consists of both the skill to do the work and the confidence that you are able to do it. If one continously take the easy route and leave more and more to the AIs, I think both sides of that equation will deteriorate, possibly one faster than the other.

Not to mention, how do you expect to gain new skills? Will what you know now, be equally helpfull 10 years from now?

@mapcar I actually think I'm gaining new skills more rapidly thanks to generative AI

I learned how to test headings using VoiceOver on my iPhone 20 minutes ago, thanks to a quick search-assisted prompt to o3

The reduction in friction for all sorts of tiny bits and pieces means I'm spending more of my cognitive energy trying and learning new things!

@mapcar it's very hard for me to generalize that though, since I have 25 years of existing dxpeiand once that was accumulated the hard way

How much of the frustration and toil of figuring things out without AI assistance is necessary to develop a deep enough understanding of things for it to stick?

@simon It for sure is the practicing of a skill that is essential to actually acquiring it.

Doing something new is always hard and frustrating, it takes a lot of self-discipline to stick with the manual practice if one is sitting next to an escape hatch.

Of course skills comes in various shapes and sizes, but just having an LLM show you something is worth little, if you do not go out and do the stuff by hand afterwards.

@simon Again, you aren't trying to find work right now and you don't know what it's like out here.
@simon It doesn't make much of a difference to my landlord if AI can really reduce engineering headcount or if that's just what they're telling people.
@simon I'm reading all these posts like they're dispatched from an alternate and increasingly divergent reality.

@3psboyd do you think the tough hiring market is mainly due to AI, or mainly due to a bunch of FANG companies flooding the market by laying off 100,000s of engineers, or a combination of the two?

I've heard it sucks, but I don't have a good feel for the underlying principle reasons

@simon I think it's more the latter, but also jobs where you need 4-6 years of experience with LLMs seem to be the only thing getting capital investment.
@simon As is the great and long tradition of machine learning, it's terrific at exacerbating existing harms.