I dont write CSS anymore
On the one hand, we see reports of AI replacing jobs.
On the other hand, we see reports of AI failing in 95% of enterprise GenAI projects.
It's hard to validate these reports, they all come with their nuances.
The truth is probably somewhere in the middle. My biggest takeaway is that we don't yet know what the exact impact is.
I see a similar pattern in my own work. For some things AI works great, for some things it doesn't. One of the things it does work for is CSS. And I'm very thankful for that, because I hate doing CSS.
The reason for that is probably because I don't do it enough, and it's often extra in my work. If I'm building a front-end to show machine learning predictions, I don't need it to look sleek.
I can imagine that if you're a web designer who cares about the way something should look, and you're well versed in that, it doesn't work for you. I don't have very clear requirements when it comes to CSS. I want it to look nicer than I can make it look myself, which is a very low bar.
Compared to programming in Django, a Python framework, CSS is also an easier task. It's easy to validate. If it looks the way you want, it works.
To me this confirms where I find that AI works well:
- I can easily evaluate it (by visual inspection)
- I don't have strict requirements (making it look nicer is already a win)
- There's no big consequences if it doesn't work well (it's just visual)
On to more things where AI can take over tedious parts of my work.