Connect your Claude, Codex or OpenClaw for full transcripts on millions of podcasts. Two week free trial, cancel anytime.
Start free trial →Front-End Fire · Hands-On Engineering Podcasts · November 2026
After VanToll noted that anti-data-center measures did well in the US elections, Herrington predicts some of the big data center build-out will end up unused, because leaner models need far less compute.
Yeah, exactly. Everybody wants the
best, though. I mean, we're all guilty of this. When I'm doing my development, would a lower end model do most of it fine? I mean, probably, but am I using Opus 5.5 anyways? You better believe I am. And so I'm complicit in all of this because it helps me do my job better. And
sure, but what happens when Anthropic decides, oh, yeah, well, one, we want to reduce our running costs. And two, we want to make sure that the agents are efficient. And, you know, they're probably going to go and follow the Chinese models, which are pretty much competitive. It's just they're just thinner and they don't need as much horsepower. And so I just really get the feeling we're going to be driving through some of these places in Texas or whatever, you know, and be like, what's that? Oh, that's a defunct data center when we were back in the days when we thought we needed massive data centers.
Yeah, but I think that's what capitalism in competition is for. If Opus gets so bloated and so expensive and they can't source data centers, then they're going to have to adapt to that and people might move to OpenAI. I think this will end up working out. That's the one good thing we have going for…
I remember.
So I'm hesitant to make any sort of predictions, but I do feel good about the fact that there's a competitive landscape in all of this.