alt Hacker NewsWhat rapid improvement has occurred, because in this six month AI coding fever dream we've been living in, I really haven't seen anything new in awhile, both in terms of new ideas for AI coding or in new consumer products or services.
I'll give you the coding harnesses themselves are better because that was a new product category with a lot of low-hanging fruit, but have the models actually improved in a way that isn't just benchmaxxing? I'd argue the models seem to be regressing. Even the most AI-pilled people at my company have all complained that Opus 4.7 is a dud. Anecdotally, GPT 5.5 seems decent, but it's rumored to be a 10T parameter model, isn't noticeably better than 5.4 or 5.3, is insanely expensive to use, and seems to be experiencing model collapse since the system prompt has to beg the thing to not talk about goblins and raccoons.
Replies
I’m going to parrot back what you’re saying and you tell me if I’m getting close
- AI coding is a disappointing fad (“fever dream?”). - that has not made meaningful progress in…6 months? - coding harness is improving - model improvements are lies: it’s just businesses “benchmaxxing” and misleading people. Real performance has not meaningfully improved - “opus 4.7 is a dud” - 5.5 suffering from “system collapse” (I’ve never heard this term before)
Since you asked and I assume you are rational and really are interested to know:
- we have many measures of performance and have studied how one particularly important but unintuitive measure (pertaining perplexity) scales with data, compute, and model size. These laws continue to hold and have satisfying theoretical origins.
- whatever the scale of 5.5, consider we have far more room to go on the scaling front. Probably another 2-3 orders of magnitude before we hit limiting bottlenecks.
- that’s also fine because scaling is only part of the puzzle. RL on verifiable rewards is virtually guaranteed to get you optimal performance and that’s the entirety of the excitement around coding agents
- while you are right about benchmarks and measurement science having a ton of weaknesses, they are not at all garbage. There are probably around 40,000 benchmarks in the literature (this is not a made up number by the way it really is around that many). Epoch made a great composite measure using good stats (IRT) called their epoch capability index, METR has done and redone their time horizon measure and it holds up beautifully. There is a ton of signal in many benchmarks and they all tell a pretty compelling story.
- additionally, this is not some unknowable thing. It strikes me as odd that people’s prior on HN a lot of time is “it’s all dumb rich people putting way too much dumb money in this”. Sorry but the world is not that dumb. Trillions of CapEx is usually pretty rationally allocated. And it is!
- why? Because this is already known what happens when you do what we’re doing. When you have a verifiable reward system, have a certain amount of compute available, have seed data to get you to where you can do RL, you will be almost guaranteed to get superhuman performance
Uninformed opinion of someone who clearly doesnt consistently use AI coding tools, clearly. And why are you limiting it to 6 months? Whats wrong with you?