[SmellTheGlove]

I agree with you, but would suggest that the thing to focus on is the next layer of abstraction. The thing that barely works. Our generation screwed around with the PSX and Xbox, the Wild West era of the internet, etc. Now the most obvious thing I can think of that is the same level of not done is AI. Yeah there’s a higher barrier to entry, but look at that guy the other day that posted about improving performance by dynamically pruning a model. Or half the crazy shit on huggingface. Generative video and image AI seems to have a lower entry barrier too.

Again I’m not saying you’re wrong. It just seems like the thing that’s always the most prime for screwing around with is the thing at the bleeding edge. There’s still some fun to have.

[poisonborz]

I would look elsewhere. SBCs like Pi Zero costing peanuts, myriads of few dollar sensors enabling all kinds of easy to hop projects, virtualisation, cheap (if non gaming) x86 hardware, much better OSS landscape, much more learning resources, GPT to help with questions... it can be easily seen as paradise as well.