Agents, Agile, Communism, Coercion — Elliot Morris

Agents, Agile, Communism, Coercion<br>June 7, 2026True communism has never been tried, right? If only those brave communards that still exist could escape the sabotages of global capital, we’d all get to benefit from this complete, correct, and egalitarian economic system.

Agile works, it’s just companies refuse to adopt it properly. People insist on thinking in top down terms, managers demand up front estimates, alas. If only these companies would actually implement capital A agile, everything would go so smoothly!

The people not on board with going all in on agentic workflows are just holding them wrong. Yes, many LLM users will be tricked into producing slop, but that’s user error, not an issue with the mode of production itself.

We keep making this same mistake.

Humans are elements of systems, not merely consumers of them. You cannot draw a neat box around the system you are designing, and treat the humans that inevitably must interact with it as a mere assumed detail. Systems that insist humans act in non-human ways are doomed to fail. Worse, they are doomed to fail in a way that the ignoble engineer cannot even conceptualize as a failure of their system, as they arrogantly refuse to consider human operational capabilities as part of their perfect model.

It’s understandable. The ever more out of reach (to the non-billionaire) need for self actualization is drawn to green fields, to the fresh pastures where one can finally lay the foundations of their genius. This is of course, delusional. All systems of virtue must, by the nature of human-centric value, incorporate humans into them. Unless you yourself are the only human your system serves, your perfect system is already corrupted by lifetimes of cultural, personal, and biological baggage. True engineers account for this, incorporate it, even derive strength and flexibility from it. Ignoble engineers, in their arrogance, childishly deny this obvious truth.

Engineering requires much more than abstract design. Implementational concerns are the majority of the work. Transitional integration strategies, both technical and psychological, are primary .

Any bozo can come up with a self-consistent system in the abstract. That does nothing to change that if people either cannot, or will not, effectively integrate themselves with “your” system, then you have failed. You have failed at systems engineering, and thus failed as an engineer.

It is a point of amusement to me that AI tech-bro culture, that so fervently pushes a get on board or get left behind ideology, will in the same breath dismiss the failures of communism, or other communally minded projects, as obvious from the start. “They don’t account for human nature”, they will say. They are right, in a naive sort of way, and yet they do not seem to understand that this line of reasoning also serves as potential contrary evidence against the viability of their own goals.

Perhaps the unspoken assumption is that the the shift will be coercive? The human parts of this system will have to fall in line, else they will starve. This might work, although similar shifts in the past have also required military, in addition to economic, coercion. I suspect this time will be no different. The Unites States at least does seem willing to become a full-throated authoritarian police state, so they’re well setup to take this sort of integration strategy.

Heck, it’s not like we don’t do coercion all the time in systems implementation. It is often too much for my liking, but it is hard to deny that some force tends to be required in order to enact positive change. I might argue that my country, The United Kingdom, could do with being a tad more coercive around getting infrastructure projects done despite NIMBY objections, for example.

However, if you are onboard the AI train, do me a favour and remember to mention this necessary piece of reasoning the next time you are in discussion. In particular, try to justify the scale of the coercion you must necessarily impose in order to force human elements to work in the way your systems require. You are an engineer. It is your duty to be able to predict, and then implement, a better tomorrow for all. Justify thyself.