circuit10

65 ms·

961.
by circuit10 4y ago
Do what, predict people’s intentions in posting comments? It can do that: Prompt: “In a conversation about language models: kelseyfrog: It just predicts the next word. tuxracer: and you? Can you describe what the user tuxracer could have be
962.
by circuit10 4y ago
Not conscious but it is intelligent, it has knowledge and can do reasoning with that knowledge I guess it depends on how you define intelligence but I don’t see why you’d link it in with consciousness, especially considering that simple alg
963.
by circuit10 4y ago
But from what I’ve seen many skilled AI researchers who have even reimplemented these models themselves do in fact believe that a scaled up ChatGPT could be AGI, or at least that it’s a step towards AGI
964.
by circuit10 4y ago
Remember, that’s for a highly customised setup, for a normal preconfigured desktop distro you’ll need a lot less
965.
by circuit10 4y ago
“ We have not made any significant progress at all in the field of general artificial intelligence in the last what, 20, 30 years?” Well clearly you haven’t been following the progress at all because there’s been an incredible amount of pro
966.
by circuit10 4y ago
Did you see some of the things it was saying to people? Do you really want your search engine to argue with you and threaten you because it thinks the date is still 2022? It can be difficult to strike the right balance here and they’re tryi
967.
by circuit10 4y ago
Could you not say the same about a human? Also what you described is probably fine under the GPL because the input is source code which the users could view and edit (and if it’s not running locally the GPL doesn’t apply, only AGPL)
968.
by circuit10 4y ago
This is flawed for a few reasons: - Training isn’t the same thing as actually using the code - The training process doesn’t change the code of GPT-3, it changes the parameters, which are input data to the code and not a part of it - They ar
969.
by circuit10 4y ago
Just Googled it and it looks like you’re the one who’s wrong… did you actually Google it yourself before posting this?
970.
by circuit10 4y ago
If I look up the definition of "open source", it says: Definitions from Oxford Languages open-source denoting software for which the original source code is made freely available and may be redistributed and modified. Emphasis on
971.
by circuit10 4y ago
No it doesn't, humans can recognise when they don't know something, current language models usually can't (yet) Their training objective, which is to predict the next piece of text in their training data, does not incentivise
972.
by circuit10 4y ago
WASM stores everything in an array buffer, it doesn't use JavaScript objects because it's not JavaScript (though there are starting to be features that allow it to interoperate with JS objects). If it didn't store everything
973.
by circuit10 4y ago
"It sounds like sci-fi" isn't a valid reason to think something is impossible. Many things in the modern world would have sounded like sci-fi 50 or 100 years ago. This type of reasoning has a name: https://en.m.wik
974.
by circuit10 4y ago
Because they're simulating redstone circuits and logic gates, not the instruction set itself (the point is that they built a CPU with redstone circuits)