Lucas Nussbaum <[email protected]> writes: > Based on recent discussions, it sounds like we probably need a > discussion in the framework of a GR to understand where we stand > regarding AI-assisted contributions to Debian.
As folks start considering how to draft these documents, can I ask people to please think about how to precisely describe what technologies the proposal you're making will affect? AI is a marketing term that is essentially devoid of meaning at this point (in my opinion, somewhat intentionally so on the part of the companies that most use the term). I have seen completely traditional software programs written entirely by humans typing into a text editor using well-known CS algorithms that involve no iterative training whatsoever called AI. In philosophical discussions, particularly on-line, it's become common for "AI" to be so amorphously and sloppily defined that it could encompass every physical object in the universe. I've seen people, apparently seriously, define humans as AI, ecosystems as AI, and even the physical universe as AI. If we're going to make policy, we need to be very specific about what we're making policy about. An LLM has some level of defined meaning, although even there it would be nice if people were specific. Reinforcement learning is a specific technique with some interesting implications, such as the existence of labeled test data used to train the algorith. "AI" just means whatever the person writing a given message wants it to mean and often changes meaning from one message to the next, which makes it not useful for writing any sort of durable policy. -- Russ Allbery ([email protected]) <https://www.eyrie.org/~eagle/>
