Of course, the quick money is now beckoning and the resources are also there. The industry is happy. The start-ups' approach is not reprehensible, but how long can it last? Differences are becoming completely blurred. The customer swaps company A for company B. With some start-ups, I have the feeling that the LLM hype has come at just the right time, because their own product doesn't really want to fly, but the staff want to be paid. Are the LLMs perhaps shattering their own dreams and they just don't want to admit it yet? So the ladies and gentlemen first build LLM applications for the industry. Innovations will come later.

Niche applications paired with data expertise

Speaking of innovations: Many young companies are very confident in their AI models - they are "virtually unbeatable". There are numerous companies in the vision sector in particular that claim to have "the best models ever". In my opinion, models - especially in the area of vision - have long been commoditized. You really can't impress investors with them anymore and no one in the vision industry has really taken off yet. And the vision sector is just the beginning. Models will become commodities. What remains then? Back to the LLMs or consulting, becoming an implementation partner for the industry. The big start-up dreams will burst for many. Unless: the start-ups have real, good data that they call their own, with which they can develop new products.

KI at HANNOVER MESSE