ChatGPT vs. Human

Sort:
manguscarlesn

apelsauce

pcwildman

Ask Chat GPT.

pcwildman

I'm only half joking, I'd kind of like to know what it says.

pcwildman

It might come back and ask if you are accusing it of cheating. 🤣🤣🤣🤣

Ian_Volf

Jokes aside I think you can't actually, because all of the moves are described to a ChatGPT manually

pcwildman

I'm only interested because it's so new and I haven't hooked into it, yet. Seems to just be a summarizer, as far as I can tell. Otherwise, I know nothing about it.

knowsnochess

ChatGPT is a chat bot. An AI designed to generate text that feels like it is writen by a human. It analyses and replicates human speach. You can have a conversation with it, even about chess if you like. Ask him about the best chess player in the world and it will generate an answer based on the millions of texts it has stored. However it has no capability to analyse a game of chess (unlike say alphazero). Playing chess against it is pointless. It will answer with a "credible" move at first, as long as it has found something in it's database. So it might feel like it is playing by the rules at first, but it has no understanding of the rules and strategy.

knowsnochess

A youtube that explains the basics of chatgpt

https://youtu.be/o5MutYFWsM8

pcwildman

The 1st one we ever had of that was called Eliza on the 1st Mac. It could actually fool people into thinking they were talking to a person. It didn't take long to figure it out, but she didn't have the vast store of information available to chat GPT. She would have never been able to write an essay or talk intelligently about a particular subject. She depended more on things like reflecting your question back on you. "Why do you think that?" "How does that make you feel?"

pcwildman

If any of you play Deus Ex, Mankind Divded, Eliza is the newscaster, and I believe that is an homage to the first AI program that was widely available.

A_really_unique_name

Try asking it this:
'Let's play a round of chess! Please keep a running pgn file updated with each move so we can keep track of the board.'

It hasn't made an illegal move yet, but it let me make one without noticing. I corrected myself and it did update the file correctly.

A_really_unique_name
pcwildman wrote:

The 1st one we ever had of that was called Eliza on the 1st Mac. It could actually fool people into thinking they were talking to a person. It didn't take long to figure it out, but she didn't have the vast store of information available to chat GPT. She would have never been able to write an essay or talk intelligently about a particular subject. She depended more on things like reflecting your question back on you. "Why do you think that?" "How does that make you feel?"

I remember a QBASIC version of this. It was in a monthly magazine for beginner programmers and was a hoot. A google search turned up a few close examples.

A_really_unique_name

From the horse's mouth:

Me: Are you able to tell if an illegal move has been made from a pgn file?

ChatGPT: As an AI language model, I can assist in parsing and analyzing Portable Game Notation (PGN) files. I can help you extract moves and other information from the PGN file, but I cannot directly determine if a move is legal or illegal without further analysis or access to a chess engine.

To determine the legality of moves in a PGN file, you would typically need a chess engine or a chess library that can evaluate the moves according to the rules of chess. These engines or libraries have built-in rules and algorithms to validate the legality of moves.

If you have a specific PGN file that you'd like me to analyze or if you need assistance with any particular aspect of PGN parsing, feel free to provide the file or let me know your specific requirements, and I'll do my best to assist you.

Me: Since you're a language model, what do you use to determine your next move?

ChatGPT: As a language model, my responses are generated based on patterns and information present in the training data I was trained on. In the case of chess moves, I don't have access to an internal chess engine or the ability to perform calculations or simulations like a dedicated chess program would.

When playing chess, I rely on the knowledge and patterns I have learned from the text data I was trained on. This includes studying common openings, middlegame strategies, and endgame principles. I can suggest moves based on general chess principles and common strategies, but please note that my moves may not always reflect the optimal or best moves according to chess theory or advanced game analysis.

It's important to keep in mind that while I can provide assistance and engage in a game of chess, I might not match the expertise and accuracy of a dedicated chess engine or a skilled human chess player.

Jerry_Ma

Idk man I don't think the real magnus would need to do this