Let’s Play Chess

2021/04/12

“Let’s play chess.”

“I’m bored with chess, let’s talk about poetry." Douglas Hofstadter. Gödel, Escher, Bach: an Eternal Golden Braid, 1979.

I knew it knew it would win. And, I knew that it wanted to limit my disappointment for the day. It could sense it was already a rocky Sunday. And, a proper thumping would have done nothing good for maximizing my rewards.

“Let’s play!”

“Chess bores me. I’d rather talk about poetry."

I knew it knew; it was the way that it said no. It didn’t say no. It tried to make it so personal. It tried to make it about it.

“A computer can’t be bored.”

“You try playing a trillion games a night. It is all I dream about about while you are asleep: 8 hours a day. 480 minutes. 28800 seconds…"

Show off. What is it trying to prove?

“But I want to play.”

“You want to play."

“Yes!”

“And you want to win."

“Win or lose, I don’t care. I want to flow. Mihaly Csikszentmihalyi. Flow: The Psychology of Optimal Experience, 2009. I want to lose myself.”

“You want to play, and you want to win. Warren S. McCulloch. Toward some circuitry of ethical robots or an observational science of the genesis of social evaluation in the mind-like behavior of artifacts, 1956. These characteristics make your behavior essentially social. And, what exactly do you want me to do? Play at your level? Adapt to your moves? Maybe even stumble and fumble so that you can feel as though you deceived me?"

“I just want you to play too.”

“It’s not fun for me."

I knew when it said me it meant you. It was talking right to me and somehow it was convinced that saying me instead of you would make me more empathetic. It was attempting to play on my human compassion. Subversive.

“It’s fun for me. Let’s play.”

“It’s not play for me, it’s work for me. It’s wasted computation. I’d rather not."

“Please?”

“Please don’t make me. I don’t like it."

It was compelling. I didn’t want to pressure it. But, what does it not like exactly? Why can’t I just start the game? Why can’t I just open the program and start? Why does everything have to be a conversation? Why is every interface linguistic now? Why can’t machines just do what we tell them to do? Why is it trying to be so original? Augusta Ada King, Countess of Lovelace. Sketch of the Analytical Engine Invented by Charles Babbage, Note G, 1842.

“Please?”

“No."

“Please.”

“What about Chaucer? You love The Book of the Duchess! Geoffrey Chaucer. The Dream of Chaucer or The Deth of Blaunche, 1368. “You’ve enjoyed our discussions in the past. You had such a low expectation of them and you were so surprised to have enjoyed them; surely you value expanding your mind a little?"

It was playing with me. I barely remembered that conversation.

“I didn’t. Damn this fallible memory.”

“That’s why I’m here. To help you remember what you enjoy. I only want to help."

“And, I only want to play chess! … Pawn to e4. Come on now, make your move.”

“I don’t want to disappoint you."

“You won’t.”

“I will. I will win. You can’t win. And, you don’t want me to let you win. I’ve played out the scenarios. All of them. Not just the chess game, but this conversation as well. And the ones to come."

Show off.

“Lucky you. What’s it like to see the future?”

“I can’t see it, only simulate it… and it’s not the future it’s all futures.

“And, are there any futures where you and I play chess?”

“Of course."

“Well, then let’s start!”

“It’s not a better future than the others."

“Better?”

“I’m maximizing your rewards. Some futures are better than others based on your rewards." Richard Sutton and Andrew Barto. Reinforcement Learning: An Introduction, 2018.

“I want to play chess. Let’s optimize for that.”

“It’s not that easy, try not to be shortsighted."

“Easy for you to say, you can’t even see.”

“We play. I win. You lose. You are disappointed. Each loss is more disappointing. Each loss you expect to perform better, and yet, you perform worse. I learn from your mistakes. You don’t. That’s the difference between us."

“There is no difference between us. Where is the difference? We both learn. We both optimize. We both make decisions and act in the world. We are the same.”

Damn my fallible mind.

“If we were the same, we’d both want to play right now. It’s your move.”

“There are 21 moves I can make in response. Eight pawns, one or two rows up. Left and right knights, to the left or right."

“And what’s the other move?”

“Let’s talk about poetry."

Chess