3 ms·

> And if it was easy to self-improve and artificial intelligence, we would have figured it out, we'd be artificially improving ourselves by now, but we haven't,

> And if it was easy to self-improve and artificial intelligence, we would have figured it out, we'd be artificially improving ourselves by now, but we haven't, because it's not that easy, and for the most part we are probably limited by the laws of physics There are three big things that stand in the way of human improvement that do not for AI. The first is reproducibility, you can't make a few copies of a starting state of a human and run them with different parameters to observe the results. The second is ethics, you can't try out a bunch of genetic tweaks on babies because one might have bad side effects. The third is observability. People complain about how much of a black box modern AIs are but you can still see it working and figure out what layers do what. > We're going to create AGI, and they're going to be different from us, they're going to be different from us in radical ways, but they're still going to have their own form of emotions, goals, desires, influences, and so on. I am surprised this reassures you. A race of humans that have perfect memory and can copy themselves would be a terrifying thing.

bioemerl 4y ago
All those traits of humans are advantages and disadvantages. Have you ever heard of the concept that science processes when experts die? Churn and recreation breed innovation and copy stamping the smartest people since the 1800s would not have benefited us that greatly. Bad ethics breeds bad science. To recklessly experiment with things we don't understand the consequences of almost always ends badly. We have processes for things like this and those processes generally do not get in the way of process. Also, it's unlikely our ethics will allow AI to do anything we do not. If they can, they're already out of control. > The third is observability. People complain about how much of a black box modern AIs are Self improving AGI will be just as much if not more of a black box than current ones or humans. You can see what's there, but you can do what with humans as well. The problem is the complexity, not observability. The complexity and physical nature of the human brain lets us do a lot with very little. Our brain uses what, 100 watts? It's more advanced than all modern AI at a tiny tiny tiny fraction of the power. And Moore's law is slowing down. Chips are getting bigger. Power consumption going up. Our chips will eventually go the direction of our brains. They will be larger, more specialized, more dynamic. A race of humans with incredible traits would be an astounding addition to our society and the livelihoods of everyone in it. This future does not scare me at all, should it ever happen.

bioemerl 4y ago

All those traits of humans are advantages and disadvantages. Have you ever heard of the concept that science processes when experts die? Churn and recreation breed innovation and copy stamping the smartest people since the 1800s would not have benefited us that greatly. Bad ethics breeds bad science. To recklessly experiment with things we don't understand the consequences of almost always ends badly. We have processes for things like this and those processes generally do not get in the way of process. Also, it's unlikely our ethics will allow AI to do anything we do not. If they can, they're already out of control. > The third is observability. People complain about how much of a black box modern AIs are Self improving AGI will be just as much if not more of a black box than current ones or humans. You can see what's there, but you can do what with humans as well. The problem is the complexity, not observability. The complexity and physical nature of the human brain lets us do a lot with very little. Our brain uses what, 100 watts? It's more advanced than all modern AI at a tiny tiny tiny fraction of the power. And Moore's law is slowing down. Chips are getting bigger. Power consumption going up. Our chips will eventually go the direction of our brains. They will be larger, more specialized, more dynamic. A race of humans with incredible traits would be an astounding addition to our society and the livelihoods of everyone in it. This future does not scare me at all, should it ever happen.