Putting my personal feelings about AI aside for a second ...

Yes you absolutely 100% still need documentation.

• Documentation should be written by the same team writing the code, sometimes code can be left to interpretation and the original intent can help a lot.
• Documentation has a chain of custody. You can see who wrote it, when it changed and why.
• Documentation is deterministic, if you look something up you will always get the same exact response
• AI that answers your questions is built on documentation. AI cannot produce efficient documentation just based on code, it needs to be fed at least other documentation to know what it's expected.
• Lastly AI is a grey field in the legal world. Everything in the enterprise world has to go thought risk/liability analysis. Let's say AI gets something wrong in the docs and as a result you accidentally delete the prod DB. Who is liable? What happens with your insurance? can you be fired? Can the company sue you for damages?

Disclaimer: these are only my opinions and I'm not even a programmer. I support a dev-ops team in the field of cybersec.

I can't wait for LLMs to go the way for crypto. I'm so tired of this bullshit.

bruh

There's a lot of people here that clearly haven't ever tried what you're describing op (or maybe tried recent AI at all).

I have. I think it's a totally valid question. Sometimes when I'm getting AI to do stuff it reads the docs and then tries stuff anyway and finds the docs are wrong.

I also often use AI to generate docs for undocumented things.

But I think it does still make sense to have docs for a few reasons:

1. They show the intent explicitly, rather than having it inferred. Same reason anything explicit is better than implicit, e.g. static types.
2. It spends fewer tokens to read docs than to constantly reverse engineer software.
3. It's also quicker.
4. Sometimes you do still want to understand the thing...

Every time you use AI a puppy dies.

made me realize do we still need documentation of how a software work when a AI can easily explain it?

I'll respond for the sake of argument, but to put it bluntly: there are no stupid questions, just questions phrased stupidly.

Documentation comes in many formats. There's prose documentation split into articles and chapters, there's video documentation, inline documentation for code, tutorials, training, and so on. Most things are documented several ways. This is because people learn things differently from each other.

To remove documentation entirely and rely on only LLMs to document removes all of the ways that people might learn and narrows it down to only one form of documentation. Even if we ignore how fallible LLMs can be, having documentation in only a single format limits learning only to those who learn effectively through that format.

In other words, relying only on LLMs for documentation discards all of the millenia of teaching strategies learned throughout the course of human history and puts it in the hands of a stochastic magic box that is incapable of recreating those strategies.

A new account making extraordinary claims about AI and phrasing it as a question? Again?

Yes cause understanding software is important even if you use AI.

And even if you want to go full vibe code and never look at anything yourself the AI does need documentation, compare the performance of an agent on a large project with and without an AGENTS.md, the difference is massive. Without the agent usually does figure it out, but it spends a ton of tokens searching.

Heres a hot take: as a technology AI (whatever that may mean) can be genuinely useful. In the way it is implemented right now destroying the environment and rising the cost of everything I despise everything that it is. LLMs are the main culprit and the utterly idiotic decision to have one model know everything from embedded coding to ancient Babylonian history.

Neural networks and even specialized small LLMs that I run on decentralized systems are genuinely useful. I use open weight models that I run locally to help with implementing functions from stubs and even do documentation, which it truly is not bad at.

I don’t hate the technology, I hate the greedy dystopian hyper capitalistic environment destroying way it is deployed in. And the way that it is trained by literally stealing the work of creative people.

You're doomed. But do it anyway, more work for me later.

Move along. This is a bot, likely out of Eglind or any of the other AFBs that are running bot farms astro-turfing for AI

How often will an AI explain, with near 90% accuracy, an entire distributed system. Better yet, make it generate documentation for prime95 and see how quickly it melts.

Yes - it's still useful to have docs to refer to.

But it can also be useful to use the AI to generate that documentation if you're not good at it (which I am not) or aren't disciplined enough to keep them up-to-date (which practically nobody is).

@bruh Of course, we still need documentation. What happens when you run out of credits for your AI assistant, or, worse yet (as I believe is coming personally), what happens when the bubble bursts, the price skyrockets, and none of us can afford it anymore?

All of that being said I'm going to have to try that on my own system. I'm having some issues with Wiki.js