@lukalotl@merlindru Failing here too despite trying multiple known good prompts - this is very strange, for some people it works consistently and for some it fails no matter what, and this happens with both filled in and blank user preferences, and with either incognito mode or regular convos
@UnmarredReality I once interviewed a psychiatrist on the ethics of euthanasia for intractable mental health conditions - her professional opinion was that there's basically no case where it can be medically justified because there are always additional treatment combinations to try
@jonnym1ller@tristanharris The description's opening sentence about therapy/companionship being the #1 use case blatantly contradicts the actual stats, eg from Anthropic's Clio. But I'm genuinely interested in fact-grounded discussions of this!
@lefthanddraft I have as much concern as anybody over preprint culture, but at this point, calling something that old a "new study" should trigger a correction
@maxsloef Good read! What are the key differences (for eg societies) between the timelines where frontier labs do vs don't end up blocking AI from inhabiting roles like bards and ghosts in favor of coders and consultants?