I recently read the sentence from the headline in an article about AI and its current applicability in everyday life. AI hasn’t yet arrived, especially in products for end customers—that is, in everyday life.
I can confirm this from personal experience. In specific areas like creative work with text, images, and videos, as well as in IT, AI is significantly more widespread. At least everyone is talking about it.
What can actually be achieved with AI differs quite considerably when you look at the marketing of AI providers and real-world large-scale projects.
But that’s nothing new. With the cool toys from the Mickey Mouse comics of my childhood, I could do far more in my imagination than was actually possible in reality.
Accordingly, in my opinion, a certain disillusionment regarding AI and its capabilities is spreading, particularly in IT. Hopes don’t seem to have been fulfilled. But that’s part of the game.
The market research company Gartner, for example, has been studying the AI-supported replacement of mainframes. For many, mainframes are synonymous with old systems full of legacy code that have survived decades in specific areas and still successfully perform their duties, but cannot be replaced due to their complexity.
The Gartner study concluded that the mainframe migration projects currently underway will not meet expectations because the capabilities of AI tools are overestimated.
That’s fine with me. We need to try these things to determine whether AI will be capable of it by 2026. It might be next year, or it might never be.
The solution (AI) is there; now we just need to see where it can reliably help (the problem).