Down the Rabbit Hole / Fredrik Knudsen -

[Xerxes IX]

Given his other hint about it requiring the assistance of his graphic artist to properly convey the more complicated aspects of the story, that actually seems like a pretty suitable pick. If it were any other YouTuber, I'd say Unabomber for sure, but as @Syaoran Li said, he hasn't really done a video on something that mainstream in quite a while.

I'm aware his series title isn't exactly literal but LessWrong is a pretty suitable topic for going down the rabbit hole. At first you just think it's the main guy writing some Harry Potter fanfiction, but then you see there's a whole philosophy behind it that starts sounding like the fedora-tipper's acceptable alternative to religion.

 

[Pineapple Fox]

New video is out!

Deep Blue | Down the Rabbit Hole​

After an electrical engineer enters the field of computer chess, his creation captures the attention of the world as he attempts to defeat the world chess champion.

 

[literal autist]

I was scared that this would be a lolcow thread ABOUT him lol
Very relieved to see it's not, I enjoy his content. He's one of the few YouTuber that I actually like. The rest turned out to be lolcows (YourMovieSucks), went to shit after a while, or both (Hacktano)

 

[Duncan Hills Coffee]

Watched the whole thing. This was a fantastic episode and the sheer breadth of the topic warranted the length. I especially liked how it built on Fredrik's ability to recount games in a clear manner; I always liked that bit in the WingsofRedemption video where he analyzes the Call of Duty match in such a way that even people who aren't familiar with it, like me, to follow along. He does an excellent job describing the chess game and the significance of the strategies used in the games. I also liked that it wasn't swallowed by AI pessimism, as in it focused more on the actual events and people as opposed to the apocalyptic abstractions that arise out of these kinds of things. It was there, but only to contextualize the importance of the games in the eyes of the public.

If I had any complaint, I didn't care for those odd illustrations he used when describing the kinds of players who dominate chess. They felt disconnected from the topic at hand. I get they were meant to evoke a certain mood, to show you the mental personifications of those personalities, but he could have used something that didn't look like something out of an anime. I did like the black hole picture though, and it is a relatively small part of the video.

 

[Gar For Archer]

Watched the whole thing. This was a fantastic episode and the sheer breadth of the topic warranted the length. I especially liked how it built on Fredrik's ability to recount games in a clear manner; I always liked that bit in the WingsofRedemption video where he analyzes the Call of Duty match in such a way that even people who aren't familiar with it, like me, to follow along. He does an excellent job describing the chess game and the significance of the strategies used in the games. I also liked that it wasn't swallowed by AI pessimism, as in it focused more on the actual events and people as opposed to the apocalyptic abstractions that arise out of these kinds of things. It was there, but only to contextualize the importance of the games in the eyes of the public.

If I had any complaint, I didn't care for those odd illustrations he used when describing the kinds of players who dominate chess. They felt disconnected from the topic at hand. I get they were meant to evoke a certain mood, to show you the mental personifications of those personalities, but he could have used something that didn't look like something out of an anime. I did like the black hole picture though, and it is a relatively small part of the video.

Apparently Fredrik used to be an esports caster (I think he mentioned this in the Wings of Redemption Q&A, specifically in reference to the COD match analysis you brought up) which explains a lot about his video style.

 

[ditto]

Watched the whole thing. This was a fantastic episode and the sheer breadth of the topic warranted the length.

Looking forward to watching it.

Does he address Garry Kasparov's accusation that Deep Blue was basically programmed to cheat by Miguel Illescas if Kasparov made a certain moves?

 

[Duncan Hills Coffee]

Does he address Garry Kasparov's accusation that Deep Blue was basically programmed to cheat by Miguel Illescas if Kasparov made a certain moves?

Sort of. Fredrik talks about the months leading up to the game and how the team programmed Deep Blue to respond to moves as well as the speed it would take to make those moves to make it appear human-like. Later he goes into Kasparov's cheating accusations, but from the sound of it, he seemed to believe that someone human was controlling Deep Blue, not necessarily that it was programmed to make those moves.

 

[Dreamer]

New video is out!

Deep Blue | Down the Rabbit Hole​

After an electrical engineer enters the field of computer chess, his creation captures the attention of the world as he attempts to defeat the world chess champion.

This video actually gave me some hope for the future of tech. I was always scared of computers being able to recreate and then generate art similar or more likely better than humans, to me this prospect would essentially mean the end of our culture.
Seeing that deep blue was modeled to adhere to established strategies in chess makes me think that the same could be applied to creating art, but as long as the medium evolves and expands computers might not be able to catch up.
Another hopeful thing was the line about how competitive chess is still around and people aren't interested in watching AI matches or humans competing with machines. Maybe should the worst come to pass people will just reject computer-made art in favor of worse but manmade works.
Slightly offtopic and gay, sorry.

 

[linchan]

Damn, I would love to see the third match. This video was an absolute masterpiece. Unironically my favorite movie I've seen this year.

 

[Arcturus]

This video actually gave me some hope for the future of tech. I was always scared of computers being able to recreate and then generate art similar or more likely better than humans, to me this prospect would essentially mean the end of our culture.
Seeing that deep blue was modeled to adhere to established strategies in chess makes me think that the same could be applied to creating art, but as long as the medium evolves and expands computers might not be able to catch up.
Another hopeful thing was the line about how competitive chess is still around and people aren't interested in watching AI matches or humans competing with machines. Maybe should the worst come to pass people will just reject computer-made art in favor of worse but manmade works.
Slightly offtopic and gay, sorry.

I think a big buffer has been that people enjoy getting wrapped up in cult of personality. Bobby Fischer and Salvador Dali, for example. I hate to destroy your hopes but they're already working on that. We've got vtubers and AI "influencers" (https://www.boredpanda.com/japanese-computer-generated-model-imma-modelingcafe/) already. People seem to love both. To the point that they act like these are real people. For now one can defend their love of them by saying "Oh the personality is really just the voice actor/the person scripting it" until they get AI voices and generated speech down better. Assuming they even have real people manning the scripts at this point. We even have people obsessed with a fake band that spawned from a video game. (KDA / League of Legends) Vocaloid was huge as well. It's only a matter of time before we have AI painters and gamers with their own "personalities" for people to get wrapped up in. I have a friend in robotics that laments his AI Twitter account getting more interaction than his real one.

I feel like whoever ran chatbot should have to pay all of us that used it as teenagers. They got free testing for their machine learning. On topic though, this was a good video and I look forward to whatever Fred does next.

 

[Dreamer]

I think a big buffer has been that people enjoy getting wrapped up in cult of personality. Bobby Fischer and Salvador Dali, for example. I hate to destroy your hopes but they're already working on that. We've got vtubers and AI "influencers" (https://www.boredpanda.com/japanese-computer-generated-model-imma-modelingcafe/) already. People seem to love both. To the point that they act like these are real people. For now one can defend their love of them by saying "Oh the personality is really just the voice actor/the person scripting it" until they get AI voices and generated speech down better. Assuming they even have real people manning the scripts at this point. We even have people obsessed with a fake band that spawned from a video game. (KDA / League of Legends) Vocaloid was huge as well. It's only a matter of time before we have AI painters and gamers with their own "personalities" for people to get wrapped up in. I have a friend in robotics that laments his AI Twitter account getting more interaction than his real one.

I feel like whoever ran chatbot should have to pay all of us that used it as teenagers. They got free testing for their machine learning. On topic though, this was a good video and I look forward to whatever Fred does next.

Well sure, just like we got the AI chess players from different companies with their own chess ratings and a particular style of playing chess we might get an AI painter with artificially created style of his own but that's fine, its a novelty not worth giving up the real deal for. I'm worried about machines complex enough to actually push out real art out of the "market" and not just exist alongside it.
My endgame scenario is something more like say for an instance if coca cola or some other cancer inducing brand decided to replace every artist they have for commercials, adverts and branding with a supercomputer so good that people wouldn't notice the difference. And then slowly one by one, our art would be mixed and eventually overtaken by AI art. I'm worried that should something like this come to pass people would just accept it and not give a fuck, or maybe just not notice at all.
Can you imagine how much stagnation and media trash would that create in our culture?

Art AI would be fine if it wasn't capable of pushing out competition with superior quantity and quality, as a novelty, something like "an artist that doesn't exist" its a cool concept that people would get bored of after a while just like with the chess AI.
Vtubers and Vocaloids are alright because they are just puppets for people and exist in their own niche, i don't see them being manned by actual AI for a good while either.

 

[Clockwork_PurBle]

I think his videos are pretty good. Some of them tend to be a bit more boring than others but that's subjective I guess.

Similarly I have enjoyed Cecil McFly's documentary videos.

 

[wtfNeedSignUp]

The last episode is pretty interesting and does a good job of describing events while painting the whole final competition as very manipulative from the side of Deepblue. The win is very much achieved with acts that are close, if not outright, cheating. With the worse being locking the other competitor from previous games Deepblue played.

I'll throw my thoughts about AI and it's risk on humanity - it's more fiction than anything that might happen. Computers are great at Chess and similar games because those games are very much constrained into a finite area with a limites set of moves. Giving computers games that are far more complex in moves and playing field (for example a game like Warcraft) will make the problem of training them infinitely harder. Additionally, computers train on an obscene amount of games, far more than a human being can play in his lifetime. Finally, the computers don't really "understand" games, you can give a computer two similar games and only train it on one of them, and it will miserably fail in the first.

In the end, a computer will only outpreform a 5 year old in playing Mario if it will have the equivelant of 500 years to play the game until it maps every screen into an optimal path.

 

[Smug Chuckler]

In the end, a computer will only outpreform a 5 year old in playing Mario if it will have the equivelant of 500 years to play the game until it maps every screen into an optimal path.

Not good enough, I want a computer that plays vidya games so well that it can speedrun and save the grief that spedrunners go through before trooning themselves out.

 

[Gar For Archer]

The last episode is pretty interesting and does a good job of describing events while painting the whole final competition as very manipulative from the side of Deepblue. The win is very much achieved with acts that are close, if not outright, cheating. With the worse being locking the other competitor from previous games Deepblue played.

I'll throw my thoughts about AI and it's risk on humanity - it's more fiction than anything that might happen. Computers are great at Chess and similar games because those games are very much constrained into a finite area with a limites set of moves. Giving computers games that are far more complex in moves and playing field (for example a game like Warcraft) will make the problem of training them infinitely harder. Additionally, computers train on an obscene amount of games, far more than a human being can play in his lifetime. Finally, the computers don't really "understand" games, you can give a computer two similar games and only train it on one of them, and it will miserably fail in the first.

In the end, a computer will only outpreform a 5 year old in playing Mario if it will have the equivelant of 500 years to play the game until it maps every screen into an optimal path.

My opinion is that "true" AI - the kind you see in sci-fi, where it's for all intents and purposes an actual sentient, conscious being that just happens to have a synthetic brain - isn't even in the realm of possibility given our current understanding, and it's certainly not the logical inevitable end goal of our current advancements in AI. The endpoint of our current technology is very advanced programs that can accurately mimic human intelligence. But mimicry doesn't mean we've created an actual thinking, feeling artificial intelligence with its own thoughts and ambitions - it just means we've created a facsimile of a human that's really good at convincing us that it has human characteristics. The AI that lies at the endpoint of today's technology would essentially be a P-zombie, reacting to everything in the exact way a human would - but only because it has been programmed to learn these actions from a massive dataset, not because it is truly a sentient being.

I don't know what it would take to create a "true" artificial intelligence, but I think right now, it lies in the same category as FTL travel - impossible as far as our current understanding of science is concerned.

 

[Raging Capybara]

I still want him to make a video on Twitch plays Pokémon Red.

 

[Kari Kamiya]

I still want him to make a video on Twitch plays Pokémon Red.

He and/or Internet Historian should. Not that TheGamerFromMars didn't do a good job talking about it, but they'd be more entertaining.

 

[SonnenSays]

He's the best documentarian youtuber right now

 

[ObeseScreamingBlackMan]

When this website inevitably suffers a massive autism singularity event that finally kills it and scatters it to the wind, I hope Fred is still around to make Down the Rabbit Hole: Kiwi Farms.

 

[Cask of Amontillado]

That Purr Cafe Episode honestly made me very upset, just with how unprofessional the owner was, the Ill cat, and just all around the horribleness surrounding the Cafe.
Very happy it was shut down, and I am happy for him to cover it. Hopefully a would be cafe owner can use it as a guide on how NOT TO run a Cat Cafe.

I wonder what happened to Diane Kelly post-Purr Cat Café? Any questions?