Item — HN Reader

I do have a genuine question though: what value are you adding at that point?

If you don’t understand the system, don’t know what the code is doing and you’re mostly prompting an LLM to make the decisions and implementation for you, what exactly is your contribution?

The ability to operate the tool isn’t much of a moat if everyone else has access to the same tool.

I don’t review assembly produced by a compiler because the compiler isn’t deciding what my system should do. It’s translating a program whose semantics were already specified. More importantly, that translation is deterministic.

A compiler takes a human-specified program and translates it into another representation while preserving its semantics.

If in five years I can give an agent a complete specification and reliably verify the resulting machine code against it, then sure, reviewing code may become obsolete and I’d happily stop doing it.

Also, keeping changes small isn’t just about making individual lines readable. It limits blast radius, makes behaviour easier to reason about, isolates mistakes, makes changes easier to revert and so many other things. None of those properties suddenly become obsolete because code generation got faster.

1 comments

Do you think it's impossible for someone to have a 25K line PR / diff but not know what's going on in the system?

I ask because I've pushed some of this to the limit in my own testing and when I get the right building blocks / primitives in place in the repo, I'm able to get plans written in a way where I don't just describe what I want but specifically mention outcomes, verifications, constraints, etc. that utlimately describe the shape of my system and my intent.

My plan is my understanding of my system and the changes I'm making, and the execution of that plan (e.g. with subagents), verified against my spec, is the code itself.

With the correct building blocks, I've found that frontier models can write code in fairly predictable ways that lets me maintain my understanding and confidence in pretty large PRs.

I think perhaps the quoted part of your post seems to imply it's not possible / isn't phrased with that nuance?

For the vast majority of 25K-line PRs, especially AI-generated ones, I would bet money that the person opening it does not actually understand all of it.

If you genuinely understand the resulting system, then that's fine. That’s not the behaviour I’m criticising.

> Do you think it's impossible for someone to have a 25K line PR / diff but not know what's going on in the system?

If you didn't write any of it? Yes, you only think you know what's going on.