[jabroni_salad]

I consult on the topic of technology for non-technology companies (think like a hydraulic piston manufacturer or something. They consume software in order to create something else.)

Basically if the technical leaders at such companies fail to shoehorn AI in somewhere, they will have to find a way to demonstrate to their board of directors that they are keeping up to date and not allowing the business to settle in to abandonware. The guys selling the software also have a similar problem -- industrial abandonware whose development team was laid off 12 years ago is for sale right now and you have to be careful not to accidentally spend a million dollars on it. Something that has AI bolted on, even in an insignificant capacity, is more likely to have present day relevance.

[firebones]

There's a lot to your point. Tech companies with barely functioning enterprise software products (or worse, companies that can acquire other companies and their enterprise products) can burn the bridge to the past by de-investing in legacy maintenance (laying off staff and expertise, cutting support), and forcing the existing user base to accept AI-fueled reboots. Buyers are sent back to market and the choice is between Behemoth Who's All in on AI's Latest VaporWareAI or Legacy Software that Works But Is a Sitting Duck.

It's an industry-wide excuse/opportunity to force what was a stagnant commodity enterprise software market back into a new buying cycle with enough face-saving modernization chrome that buyers can make spending justifications.

At best, this will be a de-evolutionary enshittification local minima era followed by a slow rebuild. More likely, the enshittification era will collapse under its own hyperinflated weight when the logrolling bubble pops, and there will be a consolidation of incremental gains, with a few new benefits being added to what would otherwise look on a long term scale as typical linear innovation if it were not for the hype bubble.