Whiskey Web and Whatnot · Hands-On Engineering Podcasts · November 2026
The hosts were debating whether frameworks like Rails still matter now that LLMs write much of the code. Adam Argyle argues that abstractions, components and frameworks exist to help humans cope with complexity, and agents don't need them.
Yeah. Ooh, that leads nicely into, let's see, I think it's on here. Yeah. Frameworks are in the way now. I'm like, frameworks. Rails is gone, right? Right. Wasn't
that the big news? I think D.A. John said he's done. Rails is done now.
…to reason with a system that's far more complicated than we can comprehend in one moment. So you break it into small independent pieces. The independent pieces can work together in tandem in a big system. And that way, when I dip into something, I can understand this chunk. This is why we do abstractions. This is why we like component models and we like frameworks to handle things that we can't. They don't matter to an LLM. It does not care. In fact, if anything, and they've been getting in our way for years, every framework I've ever adopted has always gotten in my way in some particular way or not. Even the web platform. So it's like even platforms can kind of get in the way. But at that point, you're as low level as you can get, really. So, but what I'm getting at is like agents don't need those. They don't have trouble with IDs.…
But now it's not. Now it doesn't matter. You're right. But it was not that long ago that it was like, if you were using Rails or Ember or something that was very opinionated, it was really good for, especially if it was in like TypeScript for agents to know exactly like, oh, I can't fuck this up…
Yep. So a lot of people are arguing with me at work and they're like, I still, I'm still going to build linters and I'm still going to build in a typed system because I want my agents to have the fast feedback. And I'm like, well, cool. It doesn't care. I'm going to build end-to-end tests that make…
I was just going to say that I think I hear what you're saying, but I think that lenting and tests are going to get a little bit blurred. Like the way I'm thinking about linting now is like putting in rules that will stop the agent from doing the wrong thing in a deterministic way because like…