In my stint on Twitter, I heard about “e/acc”.

It stands for effective accelerationism, and seems largely to be a way to mock Effective Altruists, the Catholic to the Protestant EAs (or reverse that?). Like most quasi-right “ideologies”, it is not taken seriously by its adherents. Only the left takes ideology seriously.

Nobody is blocking highways for accelerationism. But perhaps they should be, acceleration is the greatest form of charity.

I reiterate that one of the greatest mistakes of NRx was not attempting to claim the mantle of true progressivism. The progressives always win, all you can change is what progressive means. To me, a side that wants to conserve energy, limit population growth, and make sure no one is better than the mean is so obviously conservative. And we know what happens to conservatives.

I attended a Palladium party while in SF, and they are at least thinking seriously about what it means to build new institutions. However, it is a magazine, and therefore perhaps a meta institution.

I maintain that the solution to the current crisis is a new theory of management.

I’ve heard organizations serve either a process or a goal. My friend tried to explain this to me about government, that the goal is not to build the bridge, the goal is to make sure the process is followed to build the bridge. I didn’t really understand this until recently, but this is also how most large organizations work.

It baffles me how many people are supportive of the professional managerial class. Deep down, I think it comes from fear.

As for me, I think I went on my Hero’s Journey, but a bit more Don Quixote than Gilgamesh.

A portal appeared to the distant land of Twitter, I answered the call to adventure, found some mystifying things on the other side of the looking glass, won some battles, lost others, and returned home. Of course we weren’t going to find the Garden of Eden, but we found something perhaps far more relevant to the now. The source of Elon’s power.

It comes from a new theory of management. (I’ll note that he didn’t say this, this is an external observation)

  • by continually creating chaos, process is incapable of forming, and everyone is forced to work only toward goal.

This seems to only be possible with clearly accepted political dominance by a not incompetent evaluator, so it’s hard to scale, and even harder to make autonomous.

I am imagining a new institution, where this dictator role is replaced by AI. A real Decentralized Autonomous Organization, not a Web3 scam with 5 guys in a Discord holding the keys to a multisig wallet.

Due to Goodhart’s law, you cannot have an organization driven by a metric. But if the metric is an evaluation function beyond the complexity that humans can model, I think it works!

Going to start to think about this a lot more. Even better, I don’t think it has to be an completely adversarial game. The humans in an organization are not completely unaligned, just somewhat and only on some days.

I’ve been asking “what is the loss function of mammals?”, but the real (quite related) question to ask to solve the now is:

What should the loss function be for organizations?

Going to think about how to roll this out at comma. A goal is great, but it’s non-differentiable. Organizational RL? Credit assignment problem? Can we implement soft actor critic on our company?

It’s funny how all these problems seem to overlap.