> Similar around 2000-2001 or so. It sucked, a lot.
Was in the industry then and in 2008 and can confirm the suckage.
That said, I'd probably be a lot more concerned about the current situation than those previous times if I were an entry-level tech worker now, like the OP is.
While I don't think "AI" is likely to replace us all in the next couple of years, I do think it is playing a significant part in the near universal industry hiring freeze among employees with limited/no experience and I think its very difficult to try to predict when and if this might change regardless of other economic factors.
And to be clear, I don't think this situation is a rational long-term collective decision for companies (especially the many who are sitting on piles of cash) to make. Eventually a lack of hiring at the entry level will cause problems for everyone, but considering we now live in a world where market caps are pretty divorced from rationality and we have a labor market (in the US anyway) where both sides expect jobs to be relatively short term arrangements its easy to understand how we could have arrived here.
Absolutely, and I didn't mean to even imply that this circumstance isn't worse than then. Even being employed is kind of crappy, at least for me... beyond seeing positions expired and literally opening the same jobs at half the pay, to a lot of low-balling all around, I wound up taking a roughly 40% pay cut just to keep working... though making ends meet is much more difficult right now.
With the AI uncertainties, it's even harder still. I finally broke down and got a claude code account this past weekend to give it a try... On a few of my personal TODO items, I'd managed to do in a couple days what would have taken me literally weeks to accomplish, and I'm babysitting and reviewing everything far more than a vibe coder. There were issues, most of which I expected... but it was a far better experience than a couple years ago, and I can't even imagine how things will shake out in the end. It's still a tool, and even more so, I think you absolutely need to have experienced devs/architects at the helm of these things...
For better or worse, the org I work for has verboden AI, which is fine, but being able to scaffold something out in an hour after a couple hours of planning is pretty damned nice. ex: 2.5 hours into planning template (CLAUDE.md planning phase), then a first pass in about an hour, then 3 revisions over 2 hours with some manual tweaks. But overall a lot done in a short amount of time.
Of course, I also ran against something else that was library specific where AI didn't quite "get" what I wanted to do, implementing with a specific library/framework and kept doing goofy things. Hence comments on babysitting and experienced handlers.