4 ms·
There's a guy, Ken Stanley, who wrote the NEAT[0]/HyperNEAT[1] algorithms. His big idea is that evolution/advancements don't happen incrementally, but rather i
by gavinray 1y ago
There's a guy, Ken Stanley, who wrote the NEAT[0]/HyperNEAT[1] algorithms.
His big idea is that evolution/advancements don't happen incrementally, but rather in unpredictable large leaps.
He wrote a whole book about it that's pretty solid IMO: "Why Greatness Cannot Be Planned: The Myth of the Objective."
[0] https://en.wikipedia.org/wiki/Neuroevolution_of_augmenting_topologies https://en.wikipedia.org/wiki/Neuroevolution_of_augmenting_t...
[1] https://en.wikipedia.org/wiki/HyperNEAT https://en.wikipedia.org/wiki/HyperNEAT
- yifanl 1y agoI would suspect that any next step comes with a novel implementation though, not just trying to scale the same shit to infinity. I guess the bitter lesson is gospel now, which doesn't sit right with me now that we're past the stage of Moore's Law being relevant, but I'm not the one with a trillion dollars, so I don't matter.
- delichon 1y agohttps://en.wikipedia.org/wiki/Saltation_(biology) https://en.wikipedia.org/wiki/Saltation_(biology)
- gavinray 1y agoNeat (no pun intended), TIL there's a word for this
- devin 1y agoWhenever I try to tell people about the myth of the objective they look at me like I'm insane. It's not very popular to tell people that their best laid plans are actually part of the problem.