High praise for the Wild Robot! Makes me want to see it even more. Sounds like your kids were mixed though - aside from the parent/child point you mentioned why do you reckon that was?
I think they liked it a lot, but I was like "wow, wasn't this special! and all teary eyed", and they clearly didn't have the same emotional experience as I had.
I was just complaining to my wife that there aren't enough masons anymore to make buildings aesthetically pleasing like those that were constructed in the Art Deco era. This also likely has something to do with the move to glass and steel...this is one of my favorites from my hometown:
What a lot of labor went to beautification. That lanor and focus gives those building real staying power over a lot of the functional and efficient, but also largely characterless, buildings that predominate now.
When I look at some of those old buildings, I almost can't believe we were able to do it (and with slide rules and pencils). The amount of planning to make it all come together and look good and not be all wrong and crooked and with bad colors that don't match and all that... incredible!
AT some point I want to read more of what Kasparov has written about this. He's definitely one of the smartest humans on the planet. His 2015 book Winter is Coming was incredibly prescient in other ways.
He predicted a lot of the dynamic between Russia, Europe, the US, and he understands how autocrats think (he lived in the USSR and is very familiar with disinformation and how they maintain power and try to attack democracies, etc).
The first story idea is terrific, although I'd consider what might be going on in "meat life" for the main character, as a mirror for why he wants to find his virtual friend again? The second story proposal needs more work - there was so much turmoil going on in Europe that stopping Hitler as an action would not be sufficient to stop a conflagration. I enjoy deep-dives into alternate history themes in science fiction, so I'd be happy to read something that explores deeper consequences.
And: great link about the North/South American diagonal! Buenos Aires is two time zones ahead of NYC! Another travel idea to explore is how far north one can fly from temperate zones towards the Arctic and still stay in the same time zone (thinking of a 90's trip from Ottawa to Nanisivik).
Absolutely agree on the second story, but I wasn't trying to be serious to make it realistic. I just wanted to make a quick funny sci-fi story. Alternate histories are fascinating.
The second one reminds me of 11/22/63 by Stephen King, time traveler with the goal of stopping the Kennedy assassination. It vividly places the reader in that time period. Recommend it!
I video'd an Amazon Prime B767 jet at departure line at the BWI airport sometime around 2014-2015, about the time a hangar at BWI for Amazon was opened. I considered adding the "Jaws" theme to complete the video and posting it on FB for friends for LOLs. Yikes! Accidentally prescient, or what?
The chess section was great! How does stockfish achieve such great scores? What is the mechanism used to teach this engine to play so much better? Now I have so many questions! Ha
That's a great question! Here's the answer I found:
"Unlike earlier chess engines that relied mostly on brute-force calculation, modern Stockfish uses a sophisticated combination of two main approaches:
A neural network (called NNUE - Efficiently Updatable Neural Network) that has 'learned' to evaluate chess positions by training on millions of high-level games and positions. Think of it as having distilled the positional wisdom of countless grandmaster games into a form the computer can use instantly.
Incredibly efficient search algorithms that can explore billions of possible move sequences in seconds. While humans might look at 10-20 moves ahead in very specific variations, Stockfish can thoroughly analyze thousands of possible futures many moves deeper.
What's really cool is that Stockfish is open-source, meaning chess experts and programmers worldwide constantly contribute to improving it. They've been refining both its 'understanding' of chess positions and its ability to calculate variations for over a decade.
An interesting detail most people don't know: Stockfish doesn't actually play like humans at all. While human grandmasters often rely on pattern recognition and intuition built up over years of experience, Stockfish finds moves that can look completely alien to human players but turn out to be objectively strongest. There are positions where it will make moves that every grandmaster in history would reject as bad, but then it proves through concrete calculation that they're actually optimal.
We've created something that plays chess in a fundamentally different way than humans, and it turns out that approach is far stronger than anything we could come up with!"
This ties in perfectly with Rich Suttons essay — The bitter lesson.
He explains that human centric approaches are not the best for machines. Leveraging compute for search and learning is the biggest payoff, everytime. But researchers try to teach machines to do things the human way, which is sub optimal for a machine.
High praise for the Wild Robot! Makes me want to see it even more. Sounds like your kids were mixed though - aside from the parent/child point you mentioned why do you reckon that was?
I think they liked it a lot, but I was like "wow, wasn't this special! and all teary eyed", and they clearly didn't have the same emotional experience as I had.
It would have been funny if they were like "f*ck this, we wanna watch the Emoji Movie"
"GET OUT OF MY HOUSE"
*points at the door*
haha
I was just complaining to my wife that there aren't enough masons anymore to make buildings aesthetically pleasing like those that were constructed in the Art Deco era. This also likely has something to do with the move to glass and steel...this is one of my favorites from my hometown:
https://www.perplexity.ai/search/pictures-of-home-savings-bank-aYZF.guiTSSG7tDezNv3Aw
I had to look up the Lindy effect…very true.
It's a good heuristic. Once you know about it, you see it everywhere.
What a lot of labor went to beautification. That lanor and focus gives those building real staying power over a lot of the functional and efficient, but also largely characterless, buildings that predominate now.
Beautiful buildings are kept around way longer. In a way, it's built-in lindyness 🤔
When I look at some of those old buildings, I almost can't believe we were able to do it (and with slide rules and pencils). The amount of planning to make it all come together and look good and not be all wrong and crooked and with bad colors that don't match and all that... incredible!
We never got the Kasparov Deep Blue trilogy we deserved — I think Kasparov won the first round in Philly before losing to Deep Blue in NYC 🗽
AT some point I want to read more of what Kasparov has written about this. He's definitely one of the smartest humans on the planet. His 2015 book Winter is Coming was incredibly prescient in other ways.
I actually didn't know he was an author as well - what did you like about his book?
He predicted a lot of the dynamic between Russia, Europe, the US, and he understands how autocrats think (he lived in the USSR and is very familiar with disinformation and how they maintain power and try to attack democracies, etc).
A lot of what he wrote about and predicted in the book came true years later.
That sounds fascinating, honestly. I'm glad you brought this up!
Reading him is a good way to learn about how to think outside the box, even beyond the specifics of what he writes about.
The first story idea is terrific, although I'd consider what might be going on in "meat life" for the main character, as a mirror for why he wants to find his virtual friend again? The second story proposal needs more work - there was so much turmoil going on in Europe that stopping Hitler as an action would not be sufficient to stop a conflagration. I enjoy deep-dives into alternate history themes in science fiction, so I'd be happy to read something that explores deeper consequences.
And: great link about the North/South American diagonal! Buenos Aires is two time zones ahead of NYC! Another travel idea to explore is how far north one can fly from temperate zones towards the Arctic and still stay in the same time zone (thinking of a 90's trip from Ottawa to Nanisivik).
Thanks for reading and the feedback!
Absolutely agree on the second story, but I wasn't trying to be serious to make it realistic. I just wanted to make a quick funny sci-fi story. Alternate histories are fascinating.
The second one reminds me of 11/22/63 by Stephen King, time traveler with the goal of stopping the Kennedy assassination. It vividly places the reader in that time period. Recommend it!
I've heard of that book, but haven't read it. Didn't they make a TV series of it?
Awesome analysis, thank you!
Thanks fo reading, Michel! 💚 🥃
I video'd an Amazon Prime B767 jet at departure line at the BWI airport sometime around 2014-2015, about the time a hangar at BWI for Amazon was opened. I considered adding the "Jaws" theme to complete the video and posting it on FB for friends for LOLs. Yikes! Accidentally prescient, or what?
Haha, love it 💚 🥃
The chess section was great! How does stockfish achieve such great scores? What is the mechanism used to teach this engine to play so much better? Now I have so many questions! Ha
That's a great question! Here's the answer I found:
"Unlike earlier chess engines that relied mostly on brute-force calculation, modern Stockfish uses a sophisticated combination of two main approaches:
A neural network (called NNUE - Efficiently Updatable Neural Network) that has 'learned' to evaluate chess positions by training on millions of high-level games and positions. Think of it as having distilled the positional wisdom of countless grandmaster games into a form the computer can use instantly.
Incredibly efficient search algorithms that can explore billions of possible move sequences in seconds. While humans might look at 10-20 moves ahead in very specific variations, Stockfish can thoroughly analyze thousands of possible futures many moves deeper.
What's really cool is that Stockfish is open-source, meaning chess experts and programmers worldwide constantly contribute to improving it. They've been refining both its 'understanding' of chess positions and its ability to calculate variations for over a decade.
An interesting detail most people don't know: Stockfish doesn't actually play like humans at all. While human grandmasters often rely on pattern recognition and intuition built up over years of experience, Stockfish finds moves that can look completely alien to human players but turn out to be objectively strongest. There are positions where it will make moves that every grandmaster in history would reject as bad, but then it proves through concrete calculation that they're actually optimal.
We've created something that plays chess in a fundamentally different way than humans, and it turns out that approach is far stronger than anything we could come up with!"
Fantastic! Thanks for that.
This ties in perfectly with Rich Suttons essay — The bitter lesson.
He explains that human centric approaches are not the best for machines. Leveraging compute for search and learning is the biggest payoff, everytime. But researchers try to teach machines to do things the human way, which is sub optimal for a machine.
Stockfish is a perfect example.
https://www.cs.utexas.edu/~eunsol/courses/data/bitter_lesson.pdf
Yes, that essay is a classic!