Yeah the speed at which software can just be "regenerated" by a model and customized. I just think software is fundamentally changing from its previous nature. The overwhelming majority will be hard to charge money for, it will move to commoditized services and provision of resources, software will only be the portal to it and we will have as many selections as we need. Everything will shift to protocols and schema standards and APIs, etc. Although the secret will be in hiding all of it from the average user. Maybe "open source" isn't really the thing, but the fact that it will be free and extremely customizable. But if that continues to trend in that direction then not being open source doesn't change much either way so it almost seems like "why not?" But I obviously could be wrong, just feels like there is a massive shift in the economics of software underway and I may be projecting what I *hope* it means rather than seeing the reality of where the market will find the gap.

Yes, I think software as an end product, specifically the applications space, goes away. What takes its place is bespoke software that "writes itself" and does what the user wants, on demand. Users won't tolerate interacting with "other people's software", and it starts with websites: who wants to click a link and be bombarded with logins, popups, ads, eulas, cookie acceptance forms, captchas, on and on, when they can just ask an LLM and get a complete, fully relevant response. The next generation of an LLM's "text wall" response is a fully audio/visual interactive one.