28 ms·
I believe GPT-3 being able to convert descriptions of functions or designs into semi-functional code is what has some excited. It's certainly the strongest argu
by MathYouF 5y ago
I believe GPT-3 being able to convert descriptions of functions or designs into semi-functional code is what has some excited. It's certainly the strongest argument and reason to hope it could be different this time that I know of. I'm skeptical but come from a generative AI background so I wouldn't underestimate it will be able to do things previously unaccomplished by making templates in the past.
- qualudeheart 5y agoHow does GPT-3 differ from a gui based no code tool? I’ve tried to find that out but dunno. Do you have an idea?
- MathYouF 5y agoGPT-3 does only thing, predict the next character based on the previous n characters. The nuance of the context (previous n characters) and the quality of the prediction(s) (following character, a function of gpt-3 you can repeat by just concatenating its last prediction to the previous n characters) is what has significantly improved. It has now learned on a significant amount of the writing available on the internet. Which includes guides on writing code, stack overflow, and github repos. So one is able to write something that looks like a python comment // This function adds two numbers And the following n character predictions of gpt-3 will with some unreliability spit out the python code for that function as its predictions def add(x, y): return x+y This has been shown to work with simple descriptions of pages, with it returning html with css embedded that matches that description.
- thom 5y agoI had not put these two things together, thank you.