“Pick a number, any number.”
“Four!”
“It’s between one and three, not including one or three.”
“M!”
“Yes, the number I was thinking of was the letter M.”
Classic case of “that’s not what LLMs are made for”?
It indeed is one of these cases. Individual letter recognition doesn’t work well, since they tend to be clumped up into larger “tokens”, which then can’t be differentiated anymore. That’s also the same reason why it can’t count the letters in words, It sees words as just a single, or at most three “tokens”
That sounds like something a parent would do to a child to make the game stop lol
I got, “I’m sorry I can play hangman yet” in Gemini
i just don’t get Gemini. they try to make it replace the Google assistant but won’t let it do Google assistant things. (especially phone autonomy)
So a game of Hangman is the next level to compare AI to humans, above the Turing test.
Perplexity:
MOENY
This has to be the most pleasant thing in the world after a day of dealing with customers
Lmao
Playing hangman or 20 questions with friends and a voice ai is objectively hilarious. None of them can really do it properly
Do you have a recommendation for where I could do this? It sounds like a hoot!
The best one to play with is chat gpt’s paid subscription, but some excellent free voice ai are Google Gemini, meta’s WhatsApp ai chat, hume.ai, and chat gpt’s free voice chat.
I had a 20 questions app I on my iPod touch like 15yrs ago that guessed it correctly every time
So, no llm has been trained on Akinator. For shame, big tech, for shame.
Because 20q is basically a big decision tree. And LLMs don’t make decisions, they generate output based on what other output looks like.
It shouldn’t be too hard to teach your friends hangman!
You can only do it once before needing a new friend though.
Oh, how we’ll long for the good old days of stupid AI after the revolution.
Considering the code it spurts out, future AI will be doubly perplexed while still getting stuff wrong.
But still very confident about it!
