LLMs are making code cheaper to produce, which in turn means people are running more code. There’s more code that needs to be audited for trust than before.
And so now we have a new twist on an old problem, and I think it’s going to get worse before it gets better.
I am not sure. I was already selective before LLMs. I never ran random code from the Internet, and for those who did I think it was a mistake, already before LLMs.
There was already more code than I was able to review before LLMs. Because more code is being produced doesn't mean that I must run more code, I don't see that. I will keep running code that I can reasonably trust for the task. LLMs just add candidates or help making the better projects even better. But we will (as we should) keep auditing code, I'm sure of that.
And so now we have a new twist on an old problem, and I think it’s going to get worse before it gets better.