Disagree on the slop point. By now the code of frontier models works well, and the need to read it is gone. The Claude word salad problem was pretty much solved with the 5.5 opus release.

The other 3 points made in article make sense though.

> By now the code of frontier models works well, and the need to read it is gone.

Like anything else, it depends on what you're working on. For small personal projects I agree, but in maintaining large scale business applications with millions of lines of code, I am still having to guide the models to reuse existing code and to be more flexible with their data structures to accommodate changing business requirements in the future.

They only see a snapshot of the world the application exists in due to their ephemeral nature, and until that is overcome, they will only be able to learn general principles and not the particular nuances of your codebase that only arise from observing how the users use it on a daily basis.

If you don't need to read the code at all, what are the engineers doing to make the other issues matter? Deskilling and alienation are still about the relationship between engineers and the code.

You still have to orchestrate these things and the wholistic system architectures, pull the slot machine handle many times to get the right thing, and review/verify the working output constantly.

That said, I think we need to be prepared for how a simple prompt like "make it faster" will work in a relatively short timeframe - and where the AI is able to adequately refactor/re-platform/be done in a way where even the architecture becomes obscure to us.

I often use simple prompts like “make it faster” and get back functionally correct commits. When I review them, I often discover that they didn’t do what I really wanted - sometimes Claude made it faster by incurring expensive costs, or doing a migration that would be risky in prod, or making a tradeoff that’s not worth it.

Perhaps I could set up a workflow to handle this without reading the code. I’ve seen people who make Claude fill out arcane templates, and then spend their time reviewing reports and tinkering with loops or whatever. But why would I want to read a TPS report about the code rather than just reading the code?

Really seems more like you got used to their verbal tics and unmonitored quality, to be honest.

DHH, antirez, steipete all said something along the same line, about not needing to read all code anymore. Are they used to unmonitored quality too?

The Claude word salad was certainly dialed back somewhat, but there are still a few flies in the soup, so it’s far from solved.

Ok, if when it happens you just tell it, it does rewrite it better. Not always human coworkers can do that :)

Yes, there’s always flies in the soup, and sometimes they just serve you a whole plate of insects, but you can’t beat the price. Just ask the waiter to get you a soup with less flies, and keep doing that until the number of flies is acceptable to you.

How about, no thanks, I’ll go to a real restaurant.

“Solved” = roughly back on par with 4.8. This is like Apple rolling back the butterfly keyboard and telling you it’s the best keyboard yet

No come on. 4.8 was borderline unusable. I get intelligible results from 5.5

I don't know why people are downvoting this. It depends on what you're building.

AI build me a cost overview of xyz: works perfectly without ever looking at the code

AI build me unreal engine 5 level graphics, or a database, or infra at high scale: this will need some active steering for now :)

They will do these prompt just fine but the result will be a near identical copy of an open source version. And it won't tell you it cloned some existing engine, because it does not know, it just pulled it out from its model. Now you have an AI slopped 3d engine or DBMS no-one cares about that will just rot.

And even for active steering, does it require reading the code? Iteration can be done on the result of the work itself, abstracting code away

It used to be true before Luna but right now flash and flash lite are not competitive at all

Capex on data centers can be for additional gcp capacity for their customers, not necessarily spend for internal r&d.

When Meta, OpenAI, or Anthropic expand capacity, it's for their services. When GCP does this, it's for ALL services.

It could be that those bugs are Claude or Codex works of art.

In the not so distant future we'll just ask Claude to rewrite the entire MacOS from scratch without those bugs :)

It's a totally different technique though than what parent is referring to. The one you are referring to is used to take advantage of image and video compression algorithms

Maybe not them, but their children could do it. Italy's capability to assimilate is underrated.

I think there will be a difference between foundational software and application software. The more you go to the outer layer the more code will be generated on the spot, throw-away, single usage. On the other extreme, foundational software that is battle tested, security hardened, has survived waves of AI-hacking, will become irreplaceable and crystallise as a semi-immutable layer. Open source is in the best position to win foundational software while agents will dominate the outer layer of disposable code.

It’s based on a new and larger pre train if I understand correctly, hence the major version.

Take a walk up the alps, compare to pictures from 20 years ago and see for yourself

I like this writing style, dry and direct. Refreshing after reading ai output all the time.

Agreed