The answer is no 🙂↔️
Just joking. Technical articles do still matter indeed, but the content needs to change (cause of AI, of course).
Here is how I am thinking about articles post-AI, and what I will keep posting 👇
Most times, just ask AI
No much reasons anymore to look for generic "tutorial" articles.
For most tasks, like a lot of them, just ask AI 💁🏼♂️
AI is way more specific on your issue, your stack, your codebase, and you can ask more questions 🙌
Even better if you clone the repository of your libraries as a reference internally.
Skim "The One Weird Git Trick That Makes Coding Agents More Effect-ive" to see why that's the case, and how to set this up 💡
I have not been reading any tutorial article, nor watched any video, for quite a long while now (and I use to do that a lot previously).
When I am curious about some API, library, implementation detail, or anything, I just clone + ask.
When AI cannot answer
But! There are a few cases where AI won't answer as you'd like to.
Not saying it can't, but since AI has a tendency to go with the "simplest" choice, there are a few cases where it will miss the "better" abstraction 👀
Those cases span multiple libraries and a wider knowledge of coding patterns, type safety, and more, that the AI won't usually consider all together.
An example is the latest article I published this week: "XState sync actors with Effect Reactivity".
New article drop (after a while, first post-AI) A better way to sync XState machines with @EffectTS_ Reactivity module This is a big one, I use this pattern everywhere now 👇
The issue I had was wider that just xstate. It touched React components, synchronization, xstate, actors, and more.
Just "describing" the issue to AI would not work, the AI would just search a better
xstateAPI, or invent a slop one 🤔
The solution came from effect, with the recent Reactivity module, and knowing that xstate can work with "callback" actor events.
I saw the big picture, and put all together for AI to jump into the implementation 🏗️
These are the kind of articles that still bring value: ideas that connect together libraries to build a better shared abstraction.
The article can now serve for humans to get an idea, and then for AI to extract the implementation from it if needed (and from the source code example as well).
Non-technical articles still work
Technical content has lost some of its value, but non-technical is different.
If you have an (informed) opinion, experience, or story to share, people will keep tuning in 👀
No one wants to read a fictional "experience" written or generated by an AI. If you have a story, write about it, people will keep reading.
Another important note:
Write articles for yourself and your own understanding 🤝
AI may write most of the code, but you need to understand it. Writing articles is still the best way to connect ideas in your head, and truly understand a concept.
I may have a few more interesting patterns that may fit some new articles. Stay tuned for more 💡
See you next 👋