- Blind mode tutorial
lichess.org
Donate

What Will Become of Humanity?

Bro this Blog Really Need More Likes

Bro this Blog Really Need More Likes

Very thorough and well written. Bravo!

I would add that in the chess engine world there is a debate over how much importance to give to evaluation vs. tree search. SF prioritises tree search, using the fastest evaluation function they can get away with. It's one of the reasons SF wins the most, but its games feel like accounting, not chess.

This is important, because when looking for moves, engines long dismissed like Rybka or Dragon find beautiful ones. Maybe not best, but still beating any human in style and they don't feel as alien.

Another thing I want to add is that even the neural network engines run on many calculations per millisecond, so their self training leads to other "inhuman" strategies. Would you care about principles and game beauty if you could calculate millions of moves in your head?

And finally, as always with AI, the problem remains alignment. You said in your blog that the engine is programmed to win, to capture the enemy king. Humans may declare the same goal, but have very different ones, related to social hierarchy, self esteem, playfulness, enjoyment, power games, addiction even mental health and so on.

Again, great job!

Very thorough and well written. Bravo! I would add that in the chess engine world there is a debate over how much importance to give to evaluation vs. tree search. SF prioritises tree search, using the fastest evaluation function they can get away with. It's one of the reasons SF wins the most, but its games feel like accounting, not chess. This is important, because when looking for moves, engines long dismissed like Rybka or Dragon find beautiful ones. Maybe not best, but still beating any human in style and they don't feel as alien. Another thing I want to add is that even the neural network engines run on many calculations per millisecond, so their self training leads to other "inhuman" strategies. Would you care about principles and game beauty if you could calculate millions of moves in your head? And finally, as always with AI, the problem remains alignment. You said in your blog that the engine is programmed to win, to capture the enemy king. Humans may declare the same goal, but have very different ones, related to social hierarchy, self esteem, playfulness, enjoyment, power games, addiction even mental health and so on. Again, great job!

@TotalNoob69

Thank you for your thoughts!

As Miguel Najdorf said: "Perfection has no style".

Creativity requires variation, otherwise nothing is unique to a set standard.

Variation requires imperfection (in the context of move strength).

A creative game could be lower on strength but higher on beauty.

As you said, a goal of a machine might not be the same as that of humans.

Winning a chess game is one goal, but being creative and aesthetic is another.

@TotalNoob69 Thank you for your thoughts! As Miguel Najdorf said: "Perfection has no style". Creativity requires variation, otherwise nothing is unique to a set standard. Variation requires imperfection (in the context of move strength). A creative game could be lower on strength but higher on beauty. As you said, a goal of a machine might not be the same as that of humans. Winning a chess game is one goal, but being creative and aesthetic is another.

Good writing I know it's over my head, and yet so many professionals grind out such boring stories
; Well written Is that h.a.l. 9000 from A Space Odessy antagonizing Kasperov on the cover? Priceless

Good writing I know it's over my head, and yet so many professionals grind out such boring stories ; Well written Is that h.a.l. 9000 from A Space Odessy antagonizing Kasperov on the cover? Priceless

@TotalNoob69 said in #4:

Very thorough and well written. Bravo!

I would add that in the chess engine world there is a debate over how much importance to give to evaluation vs. tree search. SF prioritises tree search, using the fastest evaluation function they can get away with. It's one of the reasons SF wins the most, but its games feel like accounting, not chess.

This is important, because when looking for moves, engines long dismissed like Rybka or Dragon find beautiful ones. Maybe not best, but still beating any human in style and they don't feel as alien.

Another thing I want to add is that even the neural network engines run on many calculations per millisecond, so their self training leads to other "inhuman" strategies. Would you care about principles and game beauty if you could calculate millions of moves in your head?

And finally, as always with AI, the problem remains alignment. You said in your blog that the engine is programmed to win, to capture the enemy king. Humans may declare the same goal, but have very different ones, related to social hierarchy, self esteem, playfulness, enjoyment, power games, addiction even mental health and so on.

Again, great job!

I think so too

@TotalNoob69 said in #4: > Very thorough and well written. Bravo! > > I would add that in the chess engine world there is a debate over how much importance to give to evaluation vs. tree search. SF prioritises tree search, using the fastest evaluation function they can get away with. It's one of the reasons SF wins the most, but its games feel like accounting, not chess. > > This is important, because when looking for moves, engines long dismissed like Rybka or Dragon find beautiful ones. Maybe not best, but still beating any human in style and they don't feel as alien. > > Another thing I want to add is that even the neural network engines run on many calculations per millisecond, so their self training leads to other "inhuman" strategies. Would you care about principles and game beauty if you could calculate millions of moves in your head? > > And finally, as always with AI, the problem remains alignment. You said in your blog that the engine is programmed to win, to capture the enemy king. Humans may declare the same goal, but have very different ones, related to social hierarchy, self esteem, playfulness, enjoyment, power games, addiction even mental health and so on. > > Again, great job! I think so too

I don't know about humanity, but Garry wants to have a dinner with you @RuyLopez1000.

I don't know about humanity, but Garry wants to have a dinner with you @RuyLopez1000.

I don't think Kasparov really recovered from missing that draw in, what, round 2? I remember reading an article about the match years ago, where someone on the Deep Blue team (one of the GMs I think) said that GK thought of it as sort of a jumped up Fritz, and he failed to appreciate that he was playing a more sophisticated system, it was not just faster and bigger. So by the end of the match, he tried basically a cheapo anti-computer idea, that probably did work against Fritz 5.32, and he got served.

I don't think Kasparov really recovered from missing that draw in, what, round 2? I remember reading an article about the match years ago, where someone on the Deep Blue team (one of the GMs I think) said that GK thought of it as sort of a jumped up Fritz, and he failed to appreciate that he was playing a more sophisticated system, it was not just faster and bigger. So by the end of the match, he tried basically a cheapo anti-computer idea, that probably did work against Fritz 5.32, and he got served.

Honestly, I have to say this is one of my favorite research posts of all time! It truly hits all the notes when it comes to the detail, the work done in researching the topic, and the flow of ideas! In fact, this post has greatly influenced my writing, as I’m planning on writing another post similar to this about the rise of AI and computers and what we can do to close the gap. Thank you for this awesome post @RuyLopez1000!

Honestly, I have to say this is one of my favorite research posts of all time! It truly hits all the notes when it comes to the detail, the work done in researching the topic, and the flow of ideas! In fact, this post has greatly influenced my writing, as I’m planning on writing another post similar to this about the rise of AI and computers and what we can do to close the gap. Thank you for this awesome post @RuyLopez1000!