MS started aggressively using AI to generate their documentation a year or two ago. It did not make things better at all, and in fact quite the contrary. Awkwardly verbose wording, contradictory sentences in different paragraphs of the same article, etc. That said, they were already on a trajectory of decline.
If we can also apply this to network engineers, that would be awesome. No more waiting 2 weeks for a firewall rule. But how many places actually have tech writers these days.. mostly devs will be asked to write documenation.
You guys had tech writers? I write everything myself—from the code to the reports to the policies to the deployment scripts. Well at least I also get to write the firewall rules myself! Sigh...
I spent half a day writing tests against MS SQL where tests would create a separate schema, do their business, then the schema dropped via "DROP SCHEMA ... CASCADE". In the end, thanks to Meat Intelligence on the web I found out there is no CASCADE for MS SQL. But only because blogs and documentation etc were written by people who kinda mostly checked what they wrote.
The problem is that AI generated content always has the same structure and grammatical style, and you absolutely still need to guide it in order to make good content.
Tech writers will become more productive, not obsolete.
> But why not have the developer that wrote the code guide the AI to generate the content? They know the code best.
Knowing the code and knowing how to make the code, or the interface to the code, comprehensible to another user, are different things. Just like with UIs, and the fact that an expert is not necessarily the best teacher.
Anyhow, the age of monumental feats of technical writing is past. Too expensive, and the subject is too volatile for the most part. Economics dictate that we'll have to deal with the cheapest possible docs. We already do.
Knowing the source code doesn't mean someone is a skilled communicator and expecting people who are bad at <any area> to pick out problems with LLM output in that space is a losing battle
You do realize the developers only "know the code best" because they're busy writing code all day, right?
Nobody wants to be held more accountable with less control over the result.
The moment you tell the devs to focus on working with AI is the moment their guess is as good as anyone else's what the hell is going on. You're not going to squeeze more productivity this way.
I've only worked with one tech writer; they have been a dying breed for a long time. Gone are the days when software shipped with doorstopper manuals. Only a big company can justify them now. For the rest, LLMs are good enough.
MS started aggressively using AI to generate their documentation a year or two ago. It did not make things better at all, and in fact quite the contrary. Awkwardly verbose wording, contradictory sentences in different paragraphs of the same article, etc. That said, they were already on a trajectory of decline.
You should see the way Claude Code generates documentation. It's pretty good.
I have and it isn't. The YMMV with LLMs isn't a great place to be.
If we can also apply this to network engineers, that would be awesome. No more waiting 2 weeks for a firewall rule. But how many places actually have tech writers these days.. mostly devs will be asked to write documenation.
You guys had tech writers? I write everything myself—from the code to the reports to the policies to the deployment scripts. Well at least I also get to write the firewall rules myself! Sigh...
That emdash in your response. Chefs kiss
I spent half a day writing tests against MS SQL where tests would create a separate schema, do their business, then the schema dropped via "DROP SCHEMA ... CASCADE". In the end, thanks to Meat Intelligence on the web I found out there is no CASCADE for MS SQL. But only because blogs and documentation etc were written by people who kinda mostly checked what they wrote.
The problem is that AI generated content always has the same structure and grammatical style, and you absolutely still need to guide it in order to make good content.
Tech writers will become more productive, not obsolete.
But why not have the developer that wrote the code guide the AI to generate the content? They know the code best.
Actually at this rate, developers won't be writing code anyways but they're still in a better position to guide the AI.
> But why not have the developer that wrote the code guide the AI to generate the content? They know the code best.
Knowing the code and knowing how to make the code, or the interface to the code, comprehensible to another user, are different things. Just like with UIs, and the fact that an expert is not necessarily the best teacher.
Anyhow, the age of monumental feats of technical writing is past. Too expensive, and the subject is too volatile for the most part. Economics dictate that we'll have to deal with the cheapest possible docs. We already do.
Knowing the source code doesn't mean someone is a skilled communicator and expecting people who are bad at <any area> to pick out problems with LLM output in that space is a losing battle
Developers of tomorrow must be skilled communicators to get the most out of AI
You do realize the developers only "know the code best" because they're busy writing code all day, right?
Nobody wants to be held more accountable with less control over the result.
The moment you tell the devs to focus on working with AI is the moment their guess is as good as anyone else's what the hell is going on. You're not going to squeeze more productivity this way.
I've only worked with one tech writer; they have been a dying breed for a long time. Gone are the days when software shipped with doorstopper manuals. Only a big company can justify them now. For the rest, LLMs are good enough.
AI marketing slop.
> This isn't a minor efficiency gain. It's a fundamental shift
> This isn't theoretical. It works today.
> The documentation stays accurate because it's generated from real code, not someone's memory of how things used to work.
Yes, because Claude never hallucinates.
This is such a shortsighted and dangerous view. The LLM can only work on what it sees.
Doubt it.
People want to interact with other humans.
Hotel doorman problem etc.
Devs really don't want to work with tech writers to document their code though
Oh good. Now the documentation will be written by The Machine That Lies to You. Wonderful. What could possibly go wrong?
Now we need The Machine That Never Lies to You and some doors...
already are