this post was submitted on 27 Aug 2023
672 points (99.0% liked)
Technology
59374 readers
3250 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related content.
- Be excellent to each another!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, to ask if your bot can be added please contact us.
- Check for duplicates before posting, duplicates may be removed
Approved Bots
founded 1 year ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
ChatGPT for general knowledge and programming questions. Mostly straight to the point answers without 500 word drivel and 6 ad blocks on a single page for a 3 line answer you find on most blogs...
Literally the worst source for anything…
It has no understanding, it just craps out things that look right, absolutely awful for code generation beyond boilerplate. (And I do pay for the better model. )
Eh, I found it quite useful in giving me relatively well known information. As for code, it's great at telling me what functions and such do without having to traverse the documentation for a library and such, and also explaining stuff I am confused about. It is faster and more convenient for a lot of stuff, as long as you double check important info (but you have to do that anyway, never use a single source etc etc).
Its more like its only as smart as the average person... Which isnt that high of a bar so yeah for anything even mildly specific its dogshit
It isn’t “smart”, it’s a language model.
You ain't smart >:(
But in an age where were considering fridges smart, I think a language model at least contends
I mean, it doesn’t even produce compilable code half the time. Even if you give it feedback about which error it produces, it might not fix it after 3-4 corrections. I’ve ended up in loops where it cycles through incorrect suggestions, apparently forgetting that all previous answers are incorrect.
Hey all I gotta say is do you think 50% of people can properly code? Cause otherwise it should be ass at it.
But ask people to write the general structure code might take, and it does give you some boilerplate (but again might also be ass)
I wonder whether ChatGPT can evaluate trustworthiness on the fly. A lot of the complexity of modern search engines is to try to prevent gaming the system. Maybe an AI heuristic would be less predictable/gamable