I’ve been using Copilot for over a year now, and this is more or less how I use it: To help me quickly blast through boilerplate code so I can more quickly get to the tricky bits.
There’s a more subtle problem with ChatGPT’s code generation, which is that it suffers from ChatGPT’s general “bullshit” problem.
Number one:
Do things in the most straightforward way possible. It’s easy to fall into the trap of clever solutions, or clever applications of technology, or overbuilding something because you’re anticipating the future. Don’t do it. You will hate yourself for it later when you have to maintain it.
I don’t agree with all of these takes-of-varying-spiciness, but Rich Harris is always worth paying attention to.
I believe that we haven’t figured out when and how to give a developer access to an abstraction or how to evaluate when an abstraction is worth using. Abstractions are usually designed for a set of specific use-cases. The problems, however, start when a developer wants to do something that the abstraction did not anticipate.
Smart thoughts from Surma on the design of libraries, frameworks, and other abstractions:
Abstractions that take work off of developers are valuable! Of course, they are. The problems only occur when a developer feels chained to the abstractions in a situation where they’d rather do something differently. The important part is to not force patterns onto them.
This really resonated with parts of my recent talk at CSS Day when I was talking about Sass and jQuery:
If you care about DX and the adoption of your abstraction, it is much more beneficial to let developers use as much of their existing skills as possible and introduce new concepts one at a time.
Here’s a thoughtful response from Chris to my post about Svelte, Astro, and React.
The paradox of performance:
This era of incredibly fast hardware is also the era of programs that take tens of seconds to start from an SSD or NVMe disk; of bloated web applications that take many seconds to show a simple list, even on a broadband connection; of programs that process data at a thousandth of the speed we should expect. Software is laggy and sluggish — and the situation shows little signs of improvement. Why is that?
Because we prioritise the developer experience over the user experience, that’s why:
Although our job is ostensibly to create programs that let users do stuff with their computers, we place a greater emphasis on the development process and dev-oriented concerns than on the final user product.
We would do well to heed Craig’s observations on Fast Software, the Best Software.
“I wish to God these hot takes had been reckoned by steam!”
Two JavaScript frameworks—Svelte and Astro—share a philosophy, but take subtly different approaches.
The driven division.
Take my job. Please.