on the basil and dan back-n-forths the last few weeks, there's no way to "prove" either approach is correct of course, which suggests that fc should be running as many experiments as possible to get data which ofc they do, but very few of those experiments are in the category you might find in the typical L1 playbook, which makes sense considering most are trash, but maybe there is some ounce of wisdom to be gleaned there, but heading in that direction would require "de-focusing" slightly which is contrary to dan's product ideals, which have worked successfully because of the orthodoxy he brings to them, but sometimes the sharp edges of orthodoxy become obstinacy and keep you from the answers, but what if you end up trying those things and they produce trash results as expected, suddenly you've now wasted time and lost momentum
my thoughts 1 reply
0 recast
4 reactions
1 reply
0 recast
1 reaction