Will AI Destroy Us?

How would you tell it the board state?

Hmm. Is the broadcast on a delay?

Broadcast? You get your game broadcast if you get to the finals of the top tournament in the world MAYBE. All the games you need to play to actually get to the title match are not broadcast to anyone but the referee’s eyeballs via light waves bouncing off your hand and the board.

No spectators I imagine?

Maybe family of the competitors, or other players at the same tournament. Or nobody, it depends. It’s likely not enforced.

I’m going second hand here. A family friend of ours is a national master, plays a lot in the Southwest CT/NYC areas. There’s A LOT of chess being played competitively.

It’s probably very similar to a magic tournament or netrunner, with the distinction of there being a nationally recognized organization that runs it and generally very good international cooperation with other similar national organizations.

How’s it going to be powered?

I thought I remembered a scandal where some guy had a buzzer in his shoe, and if he gave some signal during a match, his confederate would run the position on the computer, and buzz him back. The idea was that for a sufficiently high level player, having one bit of information (move x or y) a game could be worth a win. Can’t seem to find it on google though.

1 Like

Pacemaker battery?

You can a metalless/xray(somehow) proof version of this in your body to cheat at chess if you like. I’m good.

It’s over. He crushed it in overtime.


image

I think with this, speed chess is the only chess that matters. Being good enough to draw at regular chess is just the cost of entry.

That’s fair enough, the best real chess player in the world is a computer. The best human is impossible (well… infeasible) to determine because of all the draws so everyone just agreed to play a different game.

That said, this headline may as well have read, water found to be wet. Carlson’s lesser known title is world speed chess champion. He happens to be world class at both games.

What is interesting is Magnus will lose his real chess title at some point. He will age. At the lower levels of chess people still win and lose. So I wonder what the game theory looks like when trying to be the first person to play Carlson in a title match when he inevitably loses it.

Study classic chess and master the art of forcing a draw. Above a certain level, never bother playing to win. Don’t even try. Your goal is a draw from move one.

Separately, focus on speed chess as your primary skill.

As for getting to be the one to face him past his prime, it’s too random to control.

It’s a straight tournament. I don’t have the rules handy but all players above a certain rank have a tournament organized by FIDE and when there are draws they’re settled by having the higher ELO during that tournament.

You’d wanna simultaneously be the highest ELO player and exceedingly good at forcing draws (And really good at speed chess).

I’m pretty sure those are competing objectives. You raise your ELO when you draw someone higher and drop it when you draw someone lower but you ALWAYS raise it on a win.

1 Like

I think the game theory changes quite a bit depending on your current rank.

Anyone who doesn’t have a reasonable chance of being at the top tier to access Magnus within ~5 years should probably focus heavily on speed chess in anticipation of changes to how chess is played.

1 Like

Definitively yes in starcraft 2. Watch as MaNa goes against google’s Deep Mind, going by AlphaStar these games.

TL:DW at present the current SC AI while being limited in that it only plays protoss against protoss is 10 and 0 against pros.

Edit: 10 and 1 against the computer! MaNa took a game off it.

So, this is a thing:
https://talktotransformer.com/

It’s an online implementation of a smaller version of OpenAI’S new GPT-2 language model, trained on text from the Internet. OpenAI has a post about it here:

It’s quite a bit of fun to play around with.

Wisecrack made an interesting video on the potential future of automation, using Wall-E and Star Trek: TNG as contrasting test cases.

https://www.youtube.com/watch?v=48mf2QUtUmg

Me in 2009: “Will AI destroy us?”

Me in 2016: “God, I hope AI destroys us.”

Me in 2019: “Hot AI lady pls steppy.”

5 Likes

So they’ve settled the “Brave New World” vs 1984 by siding with Huxley, and are now just choosing the form of the destructor?

one student that the program predicted exactly 50% of the time. We asked him what his secret was and he responded that he “just used his free will.”