There's probably an open source version of whatever you're trying to use, which you can unzip and deploy. Yet people don't do it most of the time, because they don't want to have to deal with it themselves and they value consistent experiences across industry. Gimp has been around for just short of 30 years and yet Photoshop continues to grow revenue every year. "Claude make me a Photoshop" ~= a very slow and expensive "brew install gimp" + "hey Claude just move some stuff around to make my gimp look different than everyone else for no reason."

Most people just don't want to deal with it. They want to focus on what they bring to the table and outsource the rest to people who specialize in, own, maintain and support that thing.

There are benefits to specialization, and will be for a long time even with AI. It's worth a premium to not limit all my software to what I and my agents know, and to be able to rely on other people to fix it if there's a problem.

Investment size for software plays is going to drop, and people will still invest for the same reason they do now: solving problems for people is valuable. Marginal costs dropping will bring down a lot of costs in various places, but the software itself is only part of what SaaS provides.

You're paying for effective token usage. You've simplified "automation" to mean "perfect automation". That's not going to happen for a long time, and likely never.

However, automation may likely become good enough that an SME can build the initial system to bootstrap the business. Not scale the business, but enough to prove that it's viable.

One person + AI setups still need your own expertise. Your job is now supervising the AI agents. You still have to set the direction and catch their mistakes. The better you are at it, the better chance you have.