#useable outputs in specific fields. But those are software tools witb specific use cases

[screambirdscreaming]

I remember when 3D printers first became a thing and there was a huuuge hype about how you could 3D print anything and it would revolutionize everything

And then there was a phase of fussing around realizing that there are actually some unintuitive constraints on what shapes are printable based on slicing and support of overhangs, and how you have to do fiddly business like putting tape on the platform and watching the first couple layers like a hawk or it can detach and slump sideways and become a big horrible useless mess

And then, after all that, people kinda came around to realize that even if you get all that sorted, the object you've made is fundamentally an object made of moderately-well-adhered layers of brittle plastic, which is actually a pretty shitty material for almost every purpose.

And aside from a few particular use cases the whole hype just sort of dropped off to nothing.

Anyway I feel like we're seeing pretty much the same arc play out with generative AI.