https://www.youtube.com/watch?v=8-ilcF0R7mI

What an amazing, interesting person! I like the mix of ideas from transhumanism and machine learning. She briefly mentions that people should also learn from ML models not only the other way around. And I think this is extremely underexplored area of research. Some time ago, in a vain of similar thoughts, I wrote an article on medium called “Can humans speak the language of machines?” where I dectribed an experience of trying to learn vector-based language produced by universal sentence encoder from Google. And It worked! That means that people could also communicate in vector spaces that widely more richer than usual text-based communication. Alexey Borsky

28:32 history shows that every time you hard code something in like rule-based expert 28:37 system you will be outperformed later on by something which is more generic and 28:42 kind of emerges

Related to Lex Fridman data scalation quote https://youtu.be/OwWN2OfHKpk?t=600