@elonmusk@BillAckman What exactly is the concern? Agency comes from owner having the bot engage on the forum and differentiation comes from context window.. so long as pre training and post training are sound behind the model I don’t see the issue.
@ZKForTre@Michael_Druggan If anyone’s pedantic it’s you by worrying about the scores he chose rather than the broader argument that IQ is most definitely noticeable in writing.
@howie_hua Howie, generally for mental math like this I break it into 1.) 10*25= 250 2.) 2*25= 50 3. 250+50=300 4. Final answer = 300 is there a more efficient mental model?
@drex_dsgn Kind of hypocritical. I don’t mind it but seems kind of like a self-refuting idea. No problem with it but just noticed the paradox of posting it is all.