This is a special post for quick takes by leogao. Only they can create top-level comments. Comments here also appear on the Quick Takes page and All Posts page.
why is the market for good food so inefficient? almost all mainstream restaurant food can be enumerated in a relatively short list; but there are so many extremely good off-the-beaten-path items, or regional dishes which clearly can have mass appeal but which fail to spread geographically; whereas the marginal new restaurant sells more of the same few things at roughly the same quality. recommendation algorithms, even for beli which is supposed to be specialized for restaurant recommendations, are still just not that good.
From what I have been reading lately, part of this is the challenge of scaling the ingredients. Sysco has close to a monopoly on distribution in the USA and many regions have insufficient local options due to a host of economic factors (developing land for homes and/or business earning more than agriculture being a large part of it). So everyone ends up with the same limited ingredients at the same prices and differentiation becomes a challenge.
Scaling any solution you find will hit next. The idea I had was to start at the bottom and build up with your own network of farms that feed into the restaurants. Or just have one location in a very good rich area and don't bother scaling.
2yudhister
think making recommendation algorithms for complex value functions is an unsolved problem
if it weren’t for this whole ai thing, or the whole societal epistemics thing, or the whole living forever thing, i might be working on actually good recommender systems