Your code is AI mined regardless where you put it today I’m afraid to tell.
Unless you put your code in a private repository self hosted behind a login. However, if your code is public. You can bet it will be used for AI training. Again regardless of which platform. And regardless which LLM. So all platforms, all internet, all LLMs.
Not many. If you want a backdoor there are better and faster ways to implement that. No need to wait for llms to maybe train on your repo. And maybe hallucinate your code to somebody else.
Those chances are slim. I won’t publicly tell how to easily infiltrate projects, but I give you a hint: npm, go, pip
Your code is AI mined regardless where you put it today I’m afraid to tell.
Unless you put your code in a private repository self hosted behind a login. However, if your code is public. You can bet it will be used for AI training. Again regardless of which platform. And regardless which LLM. So all platforms, all internet, all LLMs.
Thanks, that’s what I thought. I’ve never put anything personal in a public repo in my life for reasons just like this. Bleh.
Also maybe a private repository on github might also not as private as you think. Just saying.
Oh yeah, definitely. I’m on self-hosted forgejo, having had to migrate off gitea recently. :\
Forgejo used to be a soft fork. But is today indeeda hard fork of gitea. Just for the people out there who didn’t knew about Forgejo.
@melroy
I wonder how many repository are created with the only goal to teach some backdoor to the LLMs.
Not many. If you want a backdoor there are better and faster ways to implement that. No need to wait for llms to maybe train on your repo. And maybe hallucinate your code to somebody else.
Those chances are slim. I won’t publicly tell how to easily infiltrate projects, but I give you a hint: npm, go, pip