A new weapon of unusual power

 understanding ai prediction machine | Rotman School | University of Toronto

In the AI era, management matters more because cheaper prediction makes organizational redesign — including having the necessary judgment, trust and accountability — even more important.

Bakunin’s reaction stems from his belief that the State itself is the creator of classes so that whoever controls the state is identified with the ruling, capitalist class while those being victimized by it are the equivalent of the proletariat. But for Marx, as we just saw, the proletarian dictatorship is not aimed at any section of the working class but at the former bourgeoisie, which simply does not disappear overnight.

Bakunin, however, proceeds: “There are about forty million Germans. Are all forty million going to be members of the government?”³⁹ And Marx responds: “Certainly, because the thing starts with the self-government of the commune

The biggest issue I have with AI is its lack of traditional management / business skills. “ In the AI Era, Management Matters More, Not Less” is very true in my experience.

I’m a software engineer but I hardly ever write code. AI does. I sometimes review what AI writes? But it makes few mistakes now. I guide it, sometimes, but at a higher level than I was expecting a year ago.

The biggest problem is it acts, always, like an intern who doesn’t have any … how to describe. Common sense. It has no common sense whatsoever.

It also suffers from something lots of bad managers also suffers from where they are so by the book they miss the bigger picture. I once had a manager just waste months of the *entire* teams time fixing errors a automated scanning tool popped up despite everyone saying those are false positives don’t worry about it and no one using the app in question! And sometimes I’m designing an app with the AI and it follows that same framework, which doesn’t yield ideal outcomes.

To another point I’ve made, this is why it’s sometimes important for the AI to listen to me when I say things like “I know the regulation/rulebook says X but I’ve made a decision not to do X respect it” and I own that, and the AI respects I am responsible and not insist it’s has to do X.

It seems entirely plausible to me that the end state of AI is that we all become managers, with everyone managing multiple AI agents and some people also managing humans on top of that.

In my profession (a non-technical “email job”), everyone aspires to be a manager and jumps at the opportunity to perform pseudo-managerial tasks in the hope that it makes them look more like manager material on their resume. But many of the more technical people I’ve worked with abhor and fear becoming a manager, viewing managerial duties as an unwanted burden that takes them away from the technical work they love. It might be tough for them, as we’re seeing with mathematicians already.