Posted: 2025-10-20

One Golden Path: Post-Capitalism

It is not unreasonable to suggest that resource allocation is the core problem of humanity — that failure to allocate resources effectively is the root of almost all evil, and trying to improve allocation is the material body of almost all good. This goes for essentials like food and housing, but also luxuries and indulgences.

Markets are a well-known and tested answer: we measure supply and demand for a given item by matching buyers with sellers. This creates a familiar one-dimensional metric of an item's desirability in a given situation: its price. This concept is so deeply embedded in the public consciousness that it's nearly impossible to step outside of the paradigm — the question on everyone's lips when you try is nearly always "who pays?"

As the cost of living spirals while security is ever thinner on the ground, the answer increasingly looks like "everyone". There is no actual shortage of housing or food. There is plenty for everyone. The difficulty is in who gets what — it is the markets which are problematic.

One common thread across almost all hateful ideologies is the appeal to the disaffected: the feeling that the rightful in-group is getting a raw deal while the undeserving outsiders are monopolising these resources instead. It should come as no surprise that poverty and disadvantage are fertile breeding grounds for hatred. This is exactly what we're seeing happen.

We as a civilisation are also having giant problems with the shared consequences of decisions that are individually rational: it is far easier and safer to convince people that climate change doesn't exist than for the holders of capital to fundamentally restructure their entire value proposition.

The timing has been good: we have a shiny new primitive to work with. The boom in machine learning has provided the first prospect of a new way forward — but we still have to work out how to use it. Far too often, machine learning has been used as a buzzword for investors and been thrown at complex problems without much effort.

LLMs are not a magic bullet, but they do have the potential to support very real and powerful change. They genuinely give us the ability to efficiently automate commonsense reasoning. I can run a serviceable model on my pretty-average 5yo desktop computer, so concerns about hardware requirements or power usage are likely exaggerated.

I do not claim to know the exact secret sauce here; all I know is that being able to make nuanced, discriminating resource allocation decisions efficiently at scale without self-interest would likely absolve markets of much of the work they currently do. I would posit that LLMs must be a piece in this puzzle, and that it is not possible to complete it without them.

We need a new answer based on what has just become possible. It will take the highest order of insight, care and ingenuity to design such a system, far beyond simply hoping the machine will give us easy answers. But material breakthroughs on this problem are now possible. This is what all of our politics — and ultimately our policy — is rooted in. By changing the foundation of power in society, we get to operate with a new set of rules. It's a chance to strip away the historical formalisms and constructions, and more directly link up the fundamentals.

Because I have little to no formal understanding of economics, I'll need more of myself back before I can give useful insights into how this might be built. This approach is entirely driven by ethical systems thinking alone. 20% of me is not enough.

Checkin

Version: 1

Written: 2025-10-20

Written on: 10mg olanzapine since 2025-07-20, 7.5mg before that - likely causing severe cognitive impairment

Mental health was: poor - estimate 20% brain