Problems in AI alignment: A scale model

26 hamburga 3 5/22/2025, 7:25:10 PM muldoon.cloud ↗

Comments (3)

blamestross · 5m ago
I'm kind of upset to see systematically "Alignment" and "AI Safety" co-opted for "undesirable business outcomes".

These are existential problems, not mild profit blockers. Its almost like the goals of humanity and these companies are misaligned.

daveguy · 3h ago
This is an excellent point. How we choose to use and interact with AI is an individual and stochastic collective.

We can still choose not to give AI control.

hamburga · 2h ago
I find myself using the term “human superorganism” a lot these days. Our bodies are composed of many cells and microbes working together; and if we zoom out, we have the dynamics in which human individuals operate together as superorganisms. IMO there’s a whole world of productive thinking to be done to attend to the health and strength of these entities.