I think you just haven't seen the potentially revolutionary applications.
My side project is developing a tool to help me learn Chinese. Last night I asked gpt-3.5-turbo to:
1. Take the existing text which had Arabic numerals in it (e.g., "2006年" for "the year 2006" and "30万" for "300,000") and replace them with Chinese characters as they would be read (i.e., "二零零六年" for the year, "三十万" for the number)
2. Segment the text with some specific parameters (i.e., segment "二零零六" as four separate words, rather than one big word)
3. Convert the whole thing to pinyin, with specific parameters (don't group multi-character words and don't capitalize proper nouns).
I got the output I wanted with 5-10 minutes of playing around with it; I could theoretically use those prompts to process my entire library of study content. Do you have any idea how long it would take me to code something like that up?
I think you just haven't seen the potentially revolutionary applications.
My side project is developing a tool to help me learn Chinese. Last night I asked gpt-3.5-turbo to:
1. Take the existing text which had Arabic numerals in it (e.g., "2006年" for "the year 2006" and "30万" for "300,000") and replace them with Chinese characters as they would be read (i.e., "二零零六年" for the year, "三十万" for the number)
2. Segment the text with some specific parameters (i.e., segment "二零零六" as four separate words, rather than one big word)
3. Convert the whole thing to pinyin, with specific parameters (don't group multi-character words and don't capitalize proper nouns).
I got the output I wanted with 5-10 minutes of playing around with it; I could theoretically use those prompts to process my entire library of study content. Do you have any idea how long it would take me to code something like that up?