Remember when training a basic image model felt like buying a new computer every month?

Back in 2020, I dropped about $800 on cloud credits to train a simple classifier for a hobby project. I had to babysit the thing, tweak settings constantly, and the whole process took over a week. The bill felt insane for what it did. Now, I see these new open-source models that can do the same thing in hours for maybe twenty bucks worth of compute, if that. It's wild how fast the cost of just trying something out has crashed. The barrier to entry is basically gone. Makes me wonder what we'll be able to build in a couple more years that seems impossible now. What's the biggest cost drop you've seen in your own projects lately?

2 comments

2 Comments

xenarobinson5d ago

Remember when I thought you needed a big lab or a grant just to mess around with AI? I tried setting up a basic text generator last year and it was a huge pain, cost way too much for what it was. Then I tried one of those new smaller models on my own laptop last month, and it just worked. The difference in setup time and cash spent is crazy. It went from a whole project to an afternoon thing.

gavincampbell5d ago

Wasn't it just a couple years ago you needed a serious setup to even start? I totally bought into the idea that bigger and more expensive meant better results. Trying to run anything decent felt like a major investment. Now I can download a model that does useful work on my regular computer. That shift from needing special gear to using what I already have completely changed how I see this stuff. It went from a maybe-someday dream to something I can actually do.