Jamie Brandon's Speed Matters post has aged really... | Reilly Wood

Reilly's posts Post Back to profile

Jamie Brandon's Speed Matters post has aged really well. It's also pretty much what @simon is observing about LLMs

*ducks*

Quote from Jamie Brandon's Speed Matters blog post: https://www.scattered-thoughts.net/writing/speed-matters/

Being 10x faster also changes the kinds of projects that are worth doing.

Last year I spent something like 100 hours writing a text editor. That would be 2-4 weeks if had been working on it full-time. I learned the basics of graphics programming, I got to try out some unusual architecture ideas, I get the satisfaction of working every day with a tool that I made myself and I now have a small familiar codebase that I can use for other experiments. That's a pretty good deal for a few weeks work.

If I was 10x slower it would have been 20-50 weeks. Suddenly that doesn't seem like such a good deal any more - what a waste of a year!

If I was 10x faster yet it would have been 10 hours. That's a long plane ride. Even with a full-time job I would still be able to squeeze in a couple of text editor sized projects every month. I would be able to learn so many new things.

Simon Willison quote from recent Software Misadventures podcast: https://simonwillison.net/2024/Sep/10/software-misadventures/

I built a little Django application that was a, it's like a webhook debugging application. When you're working with webhooks, the thing you really want is just set up an endpoint that logs everything. And then you tell Stripe, hit my endpoint, and you get logs in your database showing what it sent you. And then you can figure things out from there. And I've always wanted the Django app for doing this, but it would take like a day to build that and I couldn't quite justify spending a day on it.

I got Claude 3.5 Sonnet to write the entire thing, and it took 2 hours from idea to having deployed working software with unit tests in production that was solving this problem for me. And it's a great example of a project where I could just about justify 2 hours on that problem. I couldn't justify any longer than that. Like, I should just use something off the shelf at that point. So, yeah, time and time again, I built all of these little projects that would not exist without LLMs, not because I couldn't build them, but because I couldn't build them fast enough to justify the the effort.

Like 14 September at 17:57 | Open on mastodon.social

1 comment

I think what a lot of people miss is that if you have a commit-sized piece of work in mind, there's a very good chance that you can get an LLM to do it in under a minute

you don't have to be an AI True Believer to see how this changes what an experienced programmer can do!

14 September at 18:06 | Open on mastodon.social