About a week ago, I had the pleasure of attending #OsloLiveTech, which I found insightful in many ways. What stood out in particular was the clear shift: generative AI is no longer the strongest player on the field—we’re now heading into the era of agents.
These are much more specialized models, custom-fitted to specific domains. Think: specialists, not generalists. (a simplified explanation of course).
What makes this evolution even more compelling is the use of real-world data to drive both patterns and utility. Expansions into IoT-related technologies have already enabled companies like #IBM to carry out process optimization at a scale that, just a few years ago, would’ve seemed like science fiction.
During some of the presentations, it was actually refreshing to hear the skepticism—and perhaps even disillusionment—with #generativeAI. My personal inference is that we’ve run straight into Kahneman’s hindsight bias (or etterpåklokskap, as we’d say in Norwegian) and realized that information from past trends is by no means a guarantee of what comes next.
It’s becoming increasingly clear in professional circles that the attempt to calculate human interaction isn’t particularly reliable. However, the automation of repetitive tasks is a direction that deserves serious consideration.
Now then, as we move into a territory where ideas seem to matter more than processes or production, we need to ask ourselves: how will the next generation—those taking over these systems—actually be trained?
As boring or maddening as they may be, repetitive tasks still form the foundation of what many consider expertise. After all, the junior eventually becomes the senior—experience still counts.
The core issue here seems to be that the extensive commoditization of knowledge has all but destroyed the value of expertise—while the systems we’ve built now demand even higher levels of it. The paradox is that we’ve eliminated the lower ranks while simultaneously destroying the very paths that lead to the upper ones.
You can see this reflected in the absurdly high competence requirements listed in many entry-level job postings—often crashing headfirst into offers that resemble gig listings on freelance websites, priced at pennies on the dollar.
Where this leads is hard to predict, but it certainly elevates the importance not just of using AI—but of actual human capital that can leverage knowledge in real time, under pressure.
As for the direction this is bound to take, I can’t help but wonder if we’re just a couple of economic crises away from realizing—again—that everything has a price, and that making everything infinitely cheaper might not be the future we actually want. While I tend to refrain from doomsday prophecies about late-stage capitalism, it’s hard to deny that the market is behaving anything but rationally—and we may be due for a correction just as the next wave of innovation arrives.
What’s particularly telling is that while the industry is racing ahead toward new models of production support, we still can’t fill the competence gaps fast enough to keep the system running. Competing economic interests, amplified by the internal contradictions of their own societies, could eventually bring everything to a halt—like a group of children who suddenly no longer want to play together at school.
Still, not every reflection has to be grim. The fall of Rome, after all, gave rise to the Middle Ages—a period which, despite its reputation, was actually a time of remarkable progress and development, as one expert once reminded me years ago.
Some articles worth taking a quick look at:
_BlnkTerminal 
Share what you think!