Ranter
Join devRant
Do all the things like
++ or -- rants, post your own rants, comment on others' rants and build your customized dev avatar
Sign Up
Pipeless API
From the creators of devRant, Pipeless lets you power real-time personalized recommendations and activity feeds using a simple API
Learn More
Comments
-
retoor18314d@atheist I was managing so much vhosts, it was time for an app. Caddy has a beautiful API so I was like hmm.. But really didn't have time for it. So, I tried to generate it and was very impressed by the results. Final outcome: it did everything in exception of the actual functionality itself. And that was totally ok! It left the fun part for me. In the end I still kinda did the whole backend.
It was an amazing experience (I really, really hate frontend). So, then - I showed a friend and he has many LLM's and he took exactly my description what I typed to chatgpt and checked what other LLM's would do with that description. It was funny, because it was nowehere stated that it was a web app, some one made a tkinter app :P Kinda awesome right? But that one didn't change the files at all. The other solutions ware also crazy in every way. The one on the picture is mine and the winner (chatGPT paid edition). By generating html/css i do not type less, but MORE! More time for backend -
retoor18314dI think I'm one of the biggest haters against code generation. Because, now everyone who doesn;t really care will make stuff. But I realize, i'm exactly that myself towards frontend. I don't freaking care myself but it's frustrating if your stuff never looks good. If you never tried to do this, Do it, the flaws are so interesting. It's a rollercoaster of "Fuck, I gonna lose my job soon" vs "Heh, i won't loose my job ever" :P. But I can tell you, we're probably both generating shit end of this year. I considered it kinda cheating, but that's not completely true. While you have spent on important stuff, you're actually focusing on the important stuff full time. We'll actually work harder than ever!
-
Nmeri172144d@retoor is there an LLM that can study about 5-7 existing files, understand a given feature and refactor those files to now behave a different way? Bonus points if it can write tests for the new functionality as well but not necessarily required
-
retoor18314d@Nmeri17 yes! It can do that for sure! No problem at all. Codeium autocomplete became a clone of me. It does do what I do. * tab tab tab *.
"I do not generate code using GPT".
The application be like:
joke/meme
could-also-be-demolishuns-humor