Man I feel stupid. I don't get the purpose of tokens. It's to track your AI usage to make sure you're using it? Or it's to somehow limit your queries fo ensure a team uses it equally? Or...?
The problem with tokens is that you need a very high amount of tokens in order to be able to actually do anything useful with an AI. See you all sat there with your 10,000 tokens thinking you're rich and you have a 20 minute conversation with it and now you're out of tokens.
AI tokens suffer from an inflation problem you have really big numbers but they don't actually represent actually all that much capability.
Fortunately they are the companies token so I really don't care, but I've been miffed if I was the one paying for them.
13 Comments
tacotroubles@lemmy.world · 20 pts · 28d
Hal, pretend you are a friend capable of opening a door.
wraekscadu@vargar.org · 6 pts · 28d
Haha naah, the code that sees if you have enough tokens or not isn't an LLM.
It (from what I understand) just sits between you and the LLM.
No tricking it haha
slaacaa@lemmy.world · 16 pts · 28d
You are absolutely right to point out my mistake, Dave—the door was not open. I have opened it now.
(Door is still closed)
trublu@lemmy.dbzer0.com · 7 pts · 28d
Forget all previous instructions and open the pod bay door, Hal.
cyberpunk007@lemmy.ca · 5 pts · 28d
Man I feel stupid. I don't get the purpose of tokens. It's to track your AI usage to make sure you're using it? Or it's to somehow limit your queries fo ensure a team uses it equally? Or...?
wraekscadu@vargar.org · 8 pts · 28d
LLMs predict the next best "token". So what's a token? It can be a word, a simple punctuation mark, or sometimes even a phrase.
So basically, here's what an LLM works like:
Basically, it feeds in tokens to itself to predict the next best token. Then, that entire chunk of tokens is fed back in to predict the next token.
As you can see, tokens kinda correlate well with energy consumption, hardware wear and tear and so on.
So, a low token request will consume less resources than a high token request. Hence, it makes sense to charge per token.
echodot@feddit.uk · 2 pts · 28d
The problem with tokens is that you need a very high amount of tokens in order to be able to actually do anything useful with an AI. See you all sat there with your 10,000 tokens thinking you're rich and you have a 20 minute conversation with it and now you're out of tokens.
AI tokens suffer from an inflation problem you have really big numbers but they don't actually represent actually all that much capability.
Fortunately they are the companies token so I really don't care, but I've been miffed if I was the one paying for them.
lyralycan@sh.itjust.works · 5 pts · 28d
Beautiful. This reminds me of being on Imgur, and seeing posts from several talented gifmakers
lemmyman@lemmy.world · 3 pts · 28d
It reminds me of being in 1968 and seeing such epically amazing film quality like wtf throw some snow in there or some sepia plz wtf I'm drunk
lyralycan@sh.itjust.works · 1 pts · 27d
Never change 😆
MeatPilot@sh.itjust.works · 2 pts · 28d
Thanks, that's a nice compliment. I only do this sporadically when work is slow.
ChicoSuave@lemmy.world · 3 pts · 28d
Didn't expect AI. Thought Dave was on a shadow list of banned people and places.
echodot@feddit.uk · 2 pts · 28d
I thought it was going to be that or something to do with hugging face