SaaS makes more money and fits into the 4HourWorkWeek “make recurring revenue from people” playbook.

Building a thing and selling it seems like it will make less money to me.

It’s funny that on a larger scale I used to hate enterprise software. It took a year to install and had to be patched and run servers and stuff. But you paid a big amount and then like 10-20%/year for maintenance and that’s it. So many things are saas and cost $500-100+/user/year and it’s not just the cost but it’s the planning and gatekeeping. Making it available to more users can be expensive. I kind miss the simplicity of budgeting $1M and being done. Now each year it’s figuring out who really needs it and cleaning up expired accounts and being stingy on if another team can use it or not.

One of my favorite things was how easy it was to scale and share with new users.

> Are gaming GPUs like a 3080 not powerful enough?

It really depends on the model. Just cherry-picking memory as a capacity dimension first: The SAM model from Meta ships at around 2.4GB w/ 360 million parameters. That trained model fits just fine on a 12GB 3080 Ti. How fast it can compute predictions on a single 3080 Ti is a different story, in the case of SAM it does well, but this ultimately depends on how complex the given model is (not the only variable, but a big one).

> don’t understand why there aren’t more options to run locally

I think it's likely that you haven't been looking in the right places for local solutions. The deep learning space is very well represented in open source at the moment across a wide set of verticals: language models, computer vision, speech recognition, voice synthesis, etc. You don't always get the white glove UX that SaaS sometimes can offer, but thats true of much of the rest of the OSS world as well.

EDIT: Wanted to note that I use both a 3080 Ti and my M2 Max for a variety of DL tasks (both for training and inference).