Curated. Questions of AI and consciousness are interesting, if not important. Unfortunately, I’ve been innoculated against thinking about to the topic due to LessWrong receiving a steady stream of low-quality/AI slop submissions from new users who claim to have awoken an AI, caused it to be a fractal conscious quantum entity with which they are in symbiosis, and so on. So I’m grateful to this post for engaging on the topic on reasonable terms.
Things I found interesting are the functional vs phenomenal angle, and that [paraphrased] we’ve got forces pushing in opposite directions re self-reports of AI consciousness: (a) for AIs to simulate human reports, (b) active training/suppression against AIs reporting consciousness. Makes for a hard scientific/philosophical problems.
Among other tricky problems, perhaps not as tricky (I don’t know, maybe more) is how to have good discussions of the topic that seems to unhinge so many. Yet maybe we can manage it here :) Kudos, thanks Kaj.
Curated. Questions of AI and consciousness are interesting, if not important. Unfortunately, I’ve been innoculated against thinking about to the topic due to LessWrong receiving a steady stream of low-quality/AI slop submissions from new users who claim to have awoken an AI, caused it to be a fractal conscious quantum entity with which they are in symbiosis, and so on. So I’m grateful to this post for engaging on the topic on reasonable terms.
Things I found interesting are the functional vs phenomenal angle, and that [paraphrased] we’ve got forces pushing in opposite directions re self-reports of AI consciousness: (a) for AIs to simulate human reports, (b) active training/suppression against AIs reporting consciousness. Makes for a hard scientific/philosophical problems.
Among other tricky problems, perhaps not as tricky (I don’t know, maybe more) is how to have good discussions of the topic that seems to unhinge so many. Yet maybe we can manage it here :) Kudos, thanks Kaj.