A few hot takes for the #vibecoding era:

I.

Coding has been heading towards higher and higher levels of abstraction for a very, very long time. We’ve gone from assembly language to C to JavaScript.

When I first started pursuing web development, years before the initial release of ChatGPT, people were already talking about it as primarily a process of gluing together existing code, instead of writing things by hand like the good old days.

Nevertheless, interview questions about low-level stuff like sorting algorithms still got asked, even though almost no-one has written such things from scratch outside of school for a long time. I think that this indicates something about the enduring value of being able to understand technical stuff at a high level of detail.

II.

LLMs are much smarter than humans. So are pocket calculators; I can’t multiply five-digit numbers in my head. It’s important to remember that even very smart LLMs are mostly good at what they’ve been specifically trained to do; that doesn’t make them irrelevant, but it does make them specialized, much like other programs and devices.

III.

The places where a project or organization’s state are stored, like issues or project management software, will be increasingly important to continually supply context and instructions to LLMs; the workflow of giving a prompt to an LLM, sitting back and waiting, and then giving it another one just isn’t very efficient in comparison.

But the primary place where any project or organization’s state is stored is in the brains of the humans working on it.

IV.

“Knowing how to code” has meant a lot of things over the past few decades, but the fundamental skills involved in contributing to a project or organization haven’t changed that much since my grandfather’s time.

Technical knowledge and skills have always been important, but code itself has never been the most valuable thing any one person can contribute.

V.

In short: game’s the same, just got more fierce.