Post

Social media algorithms, high-frequency trading, the increasing use in diagnostics, insurance, risk assessment, etc...

Places where the decision making is invisible to the public

if they make better decisions then the organizations that adopt them in a decision making role should out-perform those that don't in economic or political systems. unless you outlaw them from being used in such a way, i don't see how you escape competition dynamics.

Wouldn't the same argument imply that nations and companies are generally led by the best-informed decision-makers and problem-solvers? If they weren't, after all, competitive dynamics should long ago have pushed that nation/company aside. If this is not what we observe, we need to ask why not.

Lot of possible explanations, but this one is old and deserves to go first: "I returned and saw under the sun, that the race is not to the swift, nor the battle to the strong, neither yet bread to the wise, nor yet riches to men of understanding ... but time and chance happeneth to them all." :)

One of the studies I loved was that if you measure doctors' skill at diagnosis, it's good. If you measure AI's skill at diagnosis, it's good But if you pair a doctor with an AI, the quality drops significantly. Why? The doctors frequently overrule the AI's (correct) decision 😆

So, machines may get good at drawing or writing or folding proteins. But I tend to think we'll keep telling them what to draw or write or fold — and keep finding meaning in that directive, synthetic task. Partly for good reasons, and partly because that just seems like our M.O.

By the end of 2025, all one-way drones and cruise missiles will have computer vision for target identification to avoid radio jamming. This is already a real, deployed technology.

High-frequency trading is another established case where people choose to cede decision-making to machines - in this case directly in service of aforementioned "market forces"

My thoughts exactly. Worst use first.

As soon as the machines make more money making decisions than humans, we will cede control. Plenty of history for that.

In terms of potential, we are probably pretty much there already. It might also happen soon in practice but through a high salary human as a middleman.

I don't think one size fits all. I've encountered people in creative industries who wanted AI in one case *because* it allowed them to cede creative control to it, in another (I think) because for them success>control. OTOH where decisions = power it won't really be ceded, but we'll be told it was.