The Pinnacle of Human Capital is Middle Management?

That’s how I feel right now lol. The agents are smarter than me, they’re better than me at coding, at math, at statistics at pretty much everything I used to do for $. I feel like I’m losing control, especially when working on new things that I don’t understand. If I don’t actively tell it to slow down and explain, I quickly lose understanding.

Also another thing I’m noticing is basically all of my time now is spent on thinking about what to do instead of execution. Has execution gotten easier? Is the moat now just the thing you decide to do? Anyways, lets get back on topic, yes the title is a little playful, because there’s some truth to it? I mean we’re all becoming manager of agents now, its kinda frightening how fast this has change and I’m sure for other jobs as well, people can just use these “coding agents” for other things. This comes back to the title, if we’re in this world where everyone is a manager of a fleet of agents, then the edge is going to be who’s going to be the best manager? Should I go ahead and read some business/management books? I suspect, it’d be more of the same though? The brilliant technical founder will thrive, the tastemakers will also thrive, idk what changes? Who loses power here? The people with raw IQ? IDK why I keep thinking this? Maybe because this is not my archetype? Inherent bias maybe, but I feel like the coding agents can handle all these raw IQ tasks now, maybe there’s some meta thing where your ability to manage these coding agents is dependent upon your raw IQ?

That could be one potential world view? These high IQ ultra autistic people may lose some power relative to the other archetypes? The tastemaker can now just tell the AI what to do, before they had no knowledge or idea how to do it, now they only focus on the “idea” part? Idea guys victory? Not really total glorious dominant victory, but more a shifting of powers, zero sum game of course, no idea how much power is shifted though. If LLMs flatten out and just kinda stays here where it's “AGI”, but it's like at Einstein level, I think we’d be fine? Like the idea of a datacenter of geniuses is pretty cool, but it's like 1000 Einsteins locked up physically in one location and they can use the Internet, play around, write code, discover physics, chemistry, biology, etc, It’d basically be like a super university lab kinda thing. If however, the AIs are like x1000 the level of intelligence of Einstein, then we’re all fucked. This is the only scenario where I see Yud being right.

Anyways back to the main title, so this is what I think, if the AIs advance to Einstein level, we can control them. Like when Truman (lower iq chad) said get this “crying baby” Oppenheimer (high iq autist) out of my sight after nuking Japan. Truman here is the humans, us the normies, and Oppenheimer and his team here are the AIs which are on a raw IQ basis obviously way smarter.

I guess we as humans are just constantly moving up in terms of the abstraction at which we operate? A software engineer now, doesn’t really need to understand memory or cpu to do great work, whereas say 20 years ago that’d be a no no. Perhaps its too soon? But a great SWE nowadays doesn’t even have to know the syntax of the programming language? Like programmers not caring what the compiler produces. We just know it compiles and runs. For some fields, this is already the case, I mean if its a react app, mobile app with no real cutting edge technology and more crud like, there’s no point in even reading the code imo, but if it’s something harder like core infra work then its definitely necessary to read and understand code. I came across this problem when trying to do hft on prediction markets.

I suspect this wave of people going from manually using the tools, learning about the tools in depth may go away, white collar jobs will just be management of agents applied to different things, data analysis, financial services, software, etc, but everyone will be managing agents. I’d think the word agent and computer can be used interchangeably here, because that’s what’s going to feel like. Everyone is sitting at a desk and staring at a screen and working on their computers anyways. Interacting with software tools, now they’ll just interact with the agents that interact with the tools. See? We’re just moving up the abstraction layer. Next is managing the manager agent that manages all the agents lol.

The edge of yester years was how proficient you were in using the tools you had, now it’s about your ability to manage the agents that use the tools? Or maybe the agent is the tool? Idk distinction gets pretty blurry when it has a higher IQ than you. Maybe you are the tool and the agent is using you?