The first half of this tweet is about something genuinely cool. Great work, OpenAI. The second half is some serious eye-roll-inducing mental gymnastics. Why can't we be proud about our successes without fudging numbers to overreach with the conclusions?
Greg Brockman
 @gdb
15 Oct 2019
The robot didn't get to train *at all* with tied fingers — it had to adapt on the fly. (Also, humans have a billion plus years of evolutionary practice to solve the cube with untied fingers; the robot only gets about 10,000 years of untied practice.)
3
49
"Billion plus years" is obviously silly as we are going beyond multicellular organisms. But evolutionary prior in human brain counting for way more than 10,000 years of training of a randomly initialized ANN seems equally obviously true? That has to be what @gdb tries to convey.
1
Then why present the two numbers as a like-for-like comparison?
1
1
The "10,000 years" is a good number to have, I think. And I already said "billion plus years" is a silly one to use. Just curious if you disagree with the (likely) intended meaning or only the presentation.
1
The question was "How many billion years of training did the Deep RL agent need again?" Just wanted to point out that humans have an irreducible evolutionary prior, whose effect is hard to internalize. Could certainly have been more precise!
3
4
Replying to @gdb @xpasky @egrefen
Though, curious what y'all think of this tweet, which uses 540 million years as the comparison:
Srinath Sridhar @drsrinathsridha
15 Oct 2019
Replying to @drsrinathsridha
Unfortunately, no. Human hand dexterity is a product of millions of years of evolution, not just as primates but through our ancestors all the way back in the Cambrian explosion (~540 million years ago). (3/9) en.wikipedia.org/wiki/Cambri…

Oct 15, 2019 · 7:13 PM UTC

1
Replying to @gdb @xpasky @egrefen
To be clear, I was referring to that time frame for general manipulation to evolve, not for solving Rubik's cubes in particular. Hands didn't look like hands for a majority of that time scale but the fundamental building blocks developed.
2