So we’re replacing open source collaborators with LLMs? And if someone wants to contribute you want the prompt, not the code?
For fucks sake, people. I’ve been slinging code for thirty years. I write much better code than a lot of folks (though some certainly write better). I also am capable of reviewing my own generated code and fixing it so you don’t have to.
I get corporations trying to replace expensive coders with LLMs, but when volunteer work is being replaced by LLMs, what in the actual fuck are we even doing?
I have a similar attitude to PRs on my projects even though I don’t use LLMs. Doing something myself is just simply faster that reviewing someone else’s code, it’s potential to be malicious, wasting time on the back and forth, contributors not following the project standards.
LLM’s may be a piece of the puzzle for this guy, but I have the same opinion without them.
Sounds like the PRs are too big or something. IDK. I’ve been doing a lot of code reviews the past 10 years. But those have all been within teams or with partners. Never someone random from the internet. The review sounds easier to me than communicating feedback and getting the submitter to fix their code and submit an acceptable version.
So we’re replacing open source collaborators with LLMs? And if someone wants to contribute you want the prompt, not the code?
For fucks sake, people. I’ve been slinging code for thirty years. I write much better code than a lot of folks (though some certainly write better). I also am capable of reviewing my own generated code and fixing it so you don’t have to.
I get corporations trying to replace expensive coders with LLMs, but when volunteer work is being replaced by LLMs, what in the actual fuck are we even doing?
Descending into mass AI Psychosis?
I have a similar attitude to PRs on my projects even though I don’t use LLMs. Doing something myself is just simply faster that reviewing someone else’s code, it’s potential to be malicious, wasting time on the back and forth, contributors not following the project standards.
LLM’s may be a piece of the puzzle for this guy, but I have the same opinion without them.
Sounds like the PRs are too big or something. IDK. I’ve been doing a lot of code reviews the past 10 years. But those have all been within teams or with partners. Never someone random from the internet. The review sounds easier to me than communicating feedback and getting the submitter to fix their code and submit an acceptable version.
Yeah, it is very different with a team member / known long time contributor. I am taking about random people opening a PR.