Changed my mind: AI is good for coding

I'll be honest: I've been skeptical of AI development tools. Not because I'm anti-progress, but because the demos often feel over-hyped. Too many 'AI will replace developers' takes, too many one-shot coding videos that felt more like magic tricks than practical tools.

But this conversation with Adam Wathan (creator of Tailwind CSS) and Brian Casel (Instrumental.dev) changed my mind:

Here are a few thoughts that stood out to me:

90% of programming is grunt work

"There's so much of programming that's not actually programming," Adam commented, "90% of it is just grunt work. AI is good at that part."

This isn't about replacing creativity - it's about eliminating the tedious stuff that was never adding value anyway.

Experience matters more than ever

Adam and Brian agree that AI is terrible when you don't know what you're doing, but incredible when you're an expert.

"I've tried to use it with things I don't know how to do, and it's a fucking disaster," Adam said, "you instantly fall into that trap of AI being confidently wrong, but you can't tell because you don't know how to do it yourself."

This flips the narrative. Instead of "AI will replace experienced developers," it's more like "AI makes experienced developers superhuman."

This brings up a good question, posed by Dave Giunta in the YouTube chat:

​​How do you train junior people to learn these skills if they don't get practice doing the easy "gruntwork" things?

The business implications are real (and scary)

Adam got vulnerable about how AI is affecting the Tailwind Plus business. People want LLM.txt versions of their docs, but their business depends on traffic to the website where people notice their products.

"If everyone is just looking at a markdown file with no ability for us to advertise, we have no distribution anymore."

Thoughts/questions

First, we're all figuring this out in real time. The tectonic plates of our industry are shifting, and nobody has the playbook yet.

Second, maybe the answer isn't to fight the tide, but to figure out how to use these tools to be more ambitious about what we build. Brian mentioned he's building more complex features than he would have attempted before.

Third, I'm wondering what the repercussions of all this will be on Junior devs.

Cheers,
Justin Jackson