4 ms·
>but they were fundamentally capable of the same things. This is not the case. The difference between current state of the art NLP and chatbots 3 years ago is
by golol 4y ago
>but they were fundamentally capable of the same things.
This is not the case. The difference between current state of the art NLP and chatbots 3 years ago is so massive, it has to be seen as qualitative. Pre GPT-3 computers did not understand language and no commerical chatbot had any AI. Now computers can understand language.
- riku_iki 4y ago> Now computers can understand language. "understand"
- int_19h 4y agoIf I tell it to do X, and it does X, for all practical purposes it means that it understood what I said.
- riku_iki 4y agoit was taught to react in specific way on specific word, the same thing you can train dog to bark on "quantum physics" phrase.
- int_19h 4y agoIt can invent words, and then correctly use them to compose. https://news.ycombinator.com/item?id=35268950 https://news.ycombinator.com/item?id=35268950
- riku_iki 4y agoit can invent words, because gpt is amazing pattern extrapolator.
- int_19h 4y agoOkay, but who "taught it to bark" correctly in response to those words?
- riku_iki 4y agoyou can read papers how gpt was trained?
- int_19h 4y agoYou're making a specific claim here: "it was taught to react in specific way on specific word", and I'm giving you a specific example of an invented word that GPT used to full effect when composing. Can you explain how it was "taught to react" to this word in a manner that enables it to use it in the new meaning that wasn't even in its training set?
- riku_iki 4y agoI answered you, gpt is pattern extrapolator and was taught to be so, creating new word is totally fits pattern extrapolation.
- int_19h 4y agoHow about using that invented word consistently afterwards? It seems to me that the argument you're trying to make can be extended more or less indefinitely, though. But there's a point at which "it's a pattern extrapolator" becomes less of an explanation and more of a largely irrelevant piece of trivia.
- pixl97 4y agoWell if it can bark out the right answer its an impressive damned dog...
- riku_iki 4y agothe thing it doesn't bark right answer very often, but bark non-sense because was training on predicting next word in sentence.
- dan_mctree 4y agoAt this point ChatGPT is blowing our human ability to use language so far out of the water on so many levels, I'd argue we should start putting quotes around our human ability to "understand" language. GPT-4 has already eclipsed us when it comes to language