This is one of the worst takes from LLM enthusiasts.

Quinn Comendant

@arroz @stroughtonsmith The only reason AI coding agents are successful at building complex software is because they have compilers, linters, LSP diagnostics, and tests as guardrails to encourage the LLM to keep trying until it makes runnable code.

If “manually reviewing LLM-written code will fall by the wayside” it will be thanks to this tooling, not LLMs.

unlofl [Promoted Toot]

@arroz I think a ton of the hype and excitement around adoption is very revealing. It seems like many programmers don't like programming, probably suffer from imposter syndrome, and don't know what its like to write code they are actually very confident in.

I've done a lot of work with LAMP database applications, and even if some bizarre new input breaks it, I am damn confident I can trace the problem and determine the scope to fix the entire problem.

Owen Mathews

@arroz @stroughtonsmith Kudos to Steve for boosting this rebuttal. I have been thinking about LLMs as abstraction layer for a while now. _If_ you want to apply that framework here, you have to come up with a term of art to describe nondeteminism in a layer. A leaky abstraction exposes lower-level details; a *freaky* abstraction has some level of unpredictability. I am coding with LLMs but remain clear-eyed about their inherent unreliability. A compiler or assembler they’re not. Useful? Depends.

PositivDenken 🤯

@arroz @dmitriid now that eating out of dumpsters exists, no one ever‘s gonna want to prepare a proper meal anymore!

gudenau

@arroz my boss yesterday just said that if you don't learn to use the LLM tools, you will be fired and replaced by people who do. It's terrifying. Especially if I was allowed to say what I was working on, you would be terrified too.

Luís Correia

@arroz they all drank the koolaid and deny going back, ignoring the story 'cause the genie did get put back in the bottle

gok

@arroz @stroughtonsmith what compiler are you talking about?

Geoff Wozniak

@arroz That take is another case of only considering the result, not why the result exists.

Also, I guess he's fine with outsourcing his "compiling" to a third party to which he has absolutely no control over.

Vibecoders are only interested in production. They don't care how they get there.

Geoff Wozniak

@arroz Much like crypto pushers, vibecoders think that trust is a bad thing.

Short term gain makes for long term pain. But hey, at least we provided a lot of shareholder value for a while, amirite?

Pedro Fernandes

@arroz bingo

Sharp Cheddar Goblin

@arroz For fuck's sake these boosters are insufferable. Steve is an absolute tool.

kasia :blobcat2:

@arroz Most compilers aren't exactly deterministic by default (and fixing that is the goal of projects like Reproducible Builds), but they're still more predictable than any slop generator you'll ever come across

Sharp Cheddar Goblin

@ghalldev @arroz @stroughtonsmith Ignorance, or willful ignorance? Either describes every LLM cultist.

[object Object]

@arroz it’s always a bit depressing when I find out about a new pocket of mediocre tech jackasses posting twitter crap on masto. all of the guys posting “LLMs are like compilers for natural language” should have their CS degrees yanked cause they’ve proven they don’t meet the academic requirements for a CS undergrad.

David Gerard

@arroz it's like watching bitcoiners talk about how they think money works

[object Object]

@arroz “LLMs are natural language compilers”, brought to you by the same kids insisting their product is “the operating system for the web” because nothing means anything if you ignore all implementation and engineering details

Amber :neodog_box:

@zzt@mas.to @arroz@mastodon.social Ah because if it's one thing compilers are known for it's being non-deterministic

George B

@arroz @dmitriid @zeank

Have you seen https://www.colincornaby.me/2025/08/in-the-future-all-food-will-be-cooked-in-a-microwave-and-if-you-cant-deal-with-that-then-you-need-to-get-out-of-the-kitchen/

Cap E Bara

@arroz @stroughtonsmith that take you reposted is hella embarassing. thanks for pointing out another slop enthusiast to mute!

Duke of Germany 💫

What I see a lot in these AI guy circles is this kind of "magical thinking" about how things work.

And these (confidently expressed) naive takes are not only about LLMs, but also about countless other well-documented, well-researched topics like compilers.

Who are these guys?

@arroz