One learning I can share from the indie game industry is copycats become dime a dozen when a particular game becomes successful. With execution becoming nearly cheap, tasty products and their behaviour will be copied and competition will rise substantially.
But the simple truth is 100 programmers with AI are going to do a lot more work than 1 or 2 programmers with AI. All the usual considerations for communication overhead and all the old reasons why 100 programmers aren't necessarily 100 times as productive as 1 are in play... but so are all the reasons why that group of 100 programmers is going to outdo that 1 no matter how hard that 1 works.
We are probably in roughly the middle phase of the time period when we're still learning to harness AIs at all, and seeing simply some modest accelerations of process. Eventually we're going to progress to getting significantly enhanced productivity from them for pretty much everybody. And then what will happen is not that the SaaS market crashes, but what constitutes an entry-level SaaS product will rise up to meet the increased programming firepower. Find me a SaaS whose current two-year roadmap is "Eh, don't do anything, really, we're set" and I'll show you the SaaS company who probably ought to be really, really worried about AI eliminating their entire niche. I'm sure there are some out there, but it's not the usual SaaS company experience.
You can argue that games are derivative, but I think most revenue still goes to premier titles, not knock offs.