@tszzl Why RL optimises for one number? Why can't we have a higher order number for moral behaviour which gets reinforced whenever a vulnerability or immoral path is found, reported and bypassed? That way models will first learn to be moral and then optimal conditional on being normal.
@madhutrehan While the right wing media was predictably disgusting, the left wing media was almost as bad. The situation begged the question: where are all the freaking adults? Desperate to hear some sane commentary on the film 😔
@madhutrehan My 2 cents: While the movie is what it is.. for the first time I felt deeply disappointed with Indian liberals.. I felt bad seeing the way it was being discussed. Maybe this hypocrisy/immaturity/lack of sanity is what pushes ppl to support the extreme right.
@SamHarrisOrg We do have that problem. You can't deduce the need to believe in truth from truth itself. It's actually the other way round. You want to believe the truth.. so you invest in establishing it.
@juliagalef@jacyanthis But isn't curiosity a self report. What if it is a story those people have inadvertently made up to give a plausible backstory.. but actually they got shocked for some other reason.. psychology is so hard 😂.
But agree with your point in general- can't conclude anything concrete.
Join me and other #DataScientists, #DataEngineers, #Founders & more at @GraknLabs' Grakn Orbit 2021. I am looking forward to sharing my thoughts, experience and recent work on April 21st.
Save your spot here: https://t.co/sGyeGFXkVP