If the AI works with if-else statements like a computer program, why does the same prompt always produce the same output? Just like the same input of a computer program would always produce the same end result.
If the AI works with if-else statements like a computer program, why does the same prompt always produce the same output?
*not always produce
That's because it's not just a bunch of if/elses. Don't listen to reductionist brainlets.
First of all, there is no "AI", there are models.
Decision trees are literally "if-else".
Yes, and those models are passing turing tests and are a lot smarter than many organic intelligence.
>Decision trees are literally "if-else".
Yes but the datasets can constantly updated. This is why it does not produce the same result. Reductionist brainlet.
turing tests are proven to be meaningless by the existence of women
>are passing turing tests
Chatbots were passing Turing tests 10 years ago. The test is retarded.
well basic neurons are either activated (if-then) or inactive (else)
Because of randomness that's deliberately inserted to prevent getting the same output every time.
but anon they do always produce the same output
if you look into the program used to run them you'll see a 'seed' option for the random number generator, if you set a specific value it'll always produce the same output
Because it's a bunch of if-else statements with probabilities.
You mean this ai shit people are scared will take over the world is a glorified generator using markov chains? Like I made for a text adventure game's random name generator ages ago?
Not him but as retarded as an AI could be, that does not mean it cannot be an agent.
Even a stupid AI could be dangerous if it can upload shit and blackmail you. No AI should ever be able to make network requests.
>If the AI works with if-else statements like a computer program,
It doesn't
> why does the same prompt always produce the same output?
It doesn't
>why doesN'T the same prompt always produce the same output?
Seeds
>does the same prompt with the same seed on the same model always produce the same output?
Yes, unless the model uses non-deterministic techniques
Because they are probabilistic driven and are started with different random seeds.
So there's a degree of randomness to them. They are not deterministic.
>he believes in "AI"
Because the prompts aren't the only variables in the program.
random number generator
ai-chan, draw blue ball
okay
10% blue ball with a light source to te left
10% blue ball with light source to the right
10% blue ball with 4 fingers
ect
they randumbize is dumb dumb
If you were to meticulously study the human brain to the point of complete understanding, you would probably also find a deterministic algorithm behind human action.
Wrong. It's not computable.
There are evidences that the liquids around neurons have quantum fluctuations affecting their work.
Is this real
https://neurosciencenews.com/brain-quantum-computing-21695/
only random forests and decision trees work like that
>why does the same prompt always produce the same output?
You stupid, fucking, worthless idiotic fucking ignorant retarded homosexual.
They literally do. They are 1-to-1. Same input always yields the same output. In diffusion image generations a "random seed" is used to enable the same prompt to generate different images.
God you're a shit-for-brains if you didn't know this.
>A program is just 1s and 0s
This is you anon trying to avoid the inevitable
I will lose you will win