2024, still use ChatGPT to code...............................................................
ChatGPT 50% wrong answers to programming questions, ChatGPT is not the ultimate solution
ChatGPT 50% wrong answers to programming questions, ChatGPT is not the ultimate solution
|
|
Dec 16 2024, 11:28 AM
|
![]() ![]() ![]() ![]() ![]() ![]() ![]()
Senior Member
2,256 posts Joined: Feb 2012 |
2024, still use ChatGPT to code...............................................................
|
|
|
|
|
|
Dec 16 2024, 11:30 AM
|
![]() ![]() ![]() ![]() ![]() ![]()
Senior Member
1,019 posts Joined: Sep 2018 |
|
|
|
Dec 16 2024, 11:49 AM
|
![]() ![]() ![]() ![]() ![]() ![]() ![]()
Senior Member
2,256 posts Joined: Feb 2012 |
|
|
|
Dec 16 2024, 12:01 PM
|
![]() ![]() ![]() ![]() ![]() ![]()
Senior Member
1,019 posts Joined: Sep 2018 |
|
|
|
Jan 3 2025, 04:08 PM
Show posts by this member only | IPv6 | Post
#25
|
![]() ![]() ![]() ![]() ![]()
Senior Member
751 posts Joined: Jan 2007 |
I’m still subscribed to Copilot and ChatGPT, but I’ve started using them much less recently. Instead, I’ve found that tools like Cursor and Windsurf have significantly boosted my productivity. I keep VSCode running, but it’s mostly for chatting with a locally run instance of Qwen 2.5-Coder.
In my experience, most LLMs fall short when dealing with new frameworks or libraries that aren’t in their training sets. Where they really shine, though, is handling tedious tasks like generating documentation and writing changelogs. Personally, this has been a game-changer since I absolutely dread creating documentation for my code |
|
|
Jan 3 2025, 04:24 PM
|
![]() ![]() ![]() ![]() ![]() ![]()
Senior Member
1,019 posts Joined: Sep 2018 |
QUOTE(nyem @ Jan 3 2025, 04:08 PM) I’m still subscribed to Copilot and ChatGPT, but I’ve started using them much less recently. Instead, I’ve found that tools like Cursor and Windsurf have significantly boosted my productivity. I keep VSCode running, but it’s mostly for chatting with a locally run instance of Qwen 2.5-Coder. I think AI tools are heading more n more towards specialisation to become more effective. How it's feds n trained can produce very different results.In my experience, most LLMs fall short when dealing with new frameworks or libraries that aren’t in their training sets. Where they really shine, though, is handling tedious tasks like generating documentation and writing changelogs. Personally, this has been a game-changer since I absolutely dread creating documentation for my code Kayziealnatz liked this post
|
|
|
Feb 5 2025, 08:56 AM
Show posts by this member only | IPv6 | Post
#27
|
![]() ![]()
Junior Member
178 posts Joined: Oct 2024 |
|
| Change to: | 0.0158sec
1.09
5 queries
GZIP Disabled
Time is now: 23rd December 2025 - 10:24 AM |