AI is "good" at making quick and bad things. Things that look good at first glance, but fall apart when you actually use them. Those AI Photoshop rewrites seem great, but then some basic functionality doesn't work. Which could be fixed, but will it actually be maintained? I don't think so.

I think there's still value in having a deep understanding of the systems involved and producing good quality software.

Pentesting might be slightly different, since you can just throw tokens at a problem and AI will (maybe) stumble upon a solution. And it can test ideas more quickly than humans alone. But it still needs guidance and having knowledge that can point it in the right direction will be valuable.

And I think that makes cybersecurity just more important. If it's easier for an attacker to find vulnerabilities, then there's more pressure to not introduce any of them in the first place.

So, overall, I dont feel "threatened" by AI. I think it's overhyped, even if there's some value in it, and eventually it will quiet down. But it might be rough sailing until then.