this post was submitted on 25 Aug 2024
329 points (93.0% liked)
Technology
59672 readers
2865 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
While I do understand all of the scepticism in this thread, I have to say that I am personally amazed by GitHub Copilot.
I am just ramping up in a new company working on web development with Angular and Spring Boot. Even though I have 0 experience with this and have a background in python and C++, I got productive extremely quickly thanks to Copilot. Of course it does not work without flaws and you still need programming knowledge to wirte proper prompts and fix smaller issues in the resulting code. But without it I would be much further behind. It was even able to fix some issues in the html just based on a description of the issue I am observing in the webpage.
I do not think it will replace all programmers, but I do think it will replace some low level programmers who did repetitive tasks as the good programmers are extremely accelerated by only having to type subsets of what was needed before.
The thing about co-pilot is if you don't do anything with it, it just sits there.
You can't give it a prompt, you actually have to code stuff. It is a more advanced version of autocomplete. Now admittedly, it can write very large chunks of boilerplate code, which is extremely helpful, but it can't code the entire app, and it can't work with natural language prompts, at all.
Technically that makes chatGPT, (It really needs a better name) a more capable coder than copilot.
GitHub Copilot does chat like ChatGPT, and writes code based on a prompt. I have decades of experience programming and I use it a lot. It gives me starting points, boilerplate and examples. It won't build a whole application with no coding work from the human.
You can definitely 'chat' with copilot, like other llms as well as the inline editor auto complete.
If we replace all the low level coders with this fancy, expensive, environmentally destructive autocomplete - where are all the high level coders going to come from? Just spring from clone vats?
There will still be people who are able to rapidly learn with AI on their side and people who fail to do it. The fast ones are able to focus much more on the underlying problems and less on language specifics. The definition of low level and high level programers will change in the context of AI. Nobody today is implementing a linked list outside of university and in the future nobody will be needed to write the repeated code which still is needed in a lot of frameworks.
But of course there might the point of a critical collapse if AI only learns from its own code and inefficient code gets repeated constantly.