Ask HN: What are IT sub-fields that are less likely to be affected by AI?

I know, I know. Some argue that there is still long time before AI takes over programming completely (if it ever will). However, I really love this field (programming). I love putting pieces together to build useful systems, but I can't stop thinking about the fact that we are very close from AI taking over the majority of coding tasks and that it's a merely a matter of increasing the context length of GPTWhatever before it does what we can do.

My question is: what are the sub-fields within the IT space that are less likely to be taken over by AI completely? Or at least the ones that will be taken over last?

Here is what I think: - Security: there will need to be security experts to fight back malicious attackers who use GPTs to launch attacks. - Machine learning: it's obvious why.

I'm interested to know your thoughts on this.