Certainly, anyone that has ever been involved in a major IT project and the complexity involved, probably believes we are a long way from an AI startup building a complete system solution for some tiny niche area that involves all kinds of human variables and interfaces.

My opinion could be summed up as "I'm long Constellation". Also long Topicus, which could be thought of as "their European division with the completely random stock price because nobody follows them". So everyone should run away.

I'm sure that the LLM tools are fantastic for improving the productivity of serious programmers, which may make it worthwhile to introduce competition into a few markets that were not previous worth the effort. The textbooks say that should drive prices down. Heck, I just saw an ad on Youtube for a firm that provides software to manage your yoga studio, which I would not have suspected was a viable niche let alone one worth advertising without targeting. But that productivity boost should also help every incumbent with programmers who have already proved their product works for their market, the firms that have paying clients and are making money, so I'm willing to bet on that second group.

Jim
Former old programmer, former tech company manager, former tech VC guy, former hedge fund manager, former useful person.