fwiw with tap to pay systems the gate just stays open the whole time and doesn’t really impede the flow at all.
sure every now and then it de-syncs from the flow of traffic but my experience with them has been that they’re not any more of an impediment than the japanese style ones.
is this a thing people do? should i actually read a book about management?
it seems like everyone i talk to who reads management books ends up not actually applying their advice so maybe this whole thing is a trap.
it’s cool that some of the VCF festivals will have interactive history events but it’s not really the same.
like a decade ago the one in NJ had a functioning enigma machine that they let people type on which was incredibly cool to me at the time
NYC is slowly rolling these out and i vastly prefer them b/c they solve the turnstile problem where sometimes i am a dumbass and walk into one in a way that gets stuck and makes me look very stupid in public
i might just be uncultured but i think i just straight up like medjool dates the best?
both of these are better than deglet noor though.
Kalustyan’s had fresh dates and i kind of regret not getting some too, although what do you even do with those?
as always, be very careful if you make this it would be an incredibly stupid thing to poison yourself over.
i’m going to strain this through one mesh and one paper chemex filter once it cools.
wasting everyone’s time is also a bad thing, ideally you want to punish this behavior somehow but my bet here is that you can punish bad behavior and make it economically unprofitable to engage in it.
if it’s wildly expensive to engage in fraud it gets more expensive to waste peoples‘ time, imo.
MOST IMPORTANTLY, if the LLM makes the meta-problem of scam & fraud & general corporate irresponsibility more profitable then you use your regulatory apparatus to capture that profit center!
make it financially ruinous to commit these crimes and you either dissuade the crime or make more money!
if the LLM makes it easier for people to do a lot of white collar crime you go after white collar crime.
if the LLM makes it easier to produce predatory products you go after the people producing predatory products.
so i’m pretty the thing that helps here more than anything else is a regulatory framework that exists to punish people for the harms that are committed as a result of using the technology or applying it directly in industries that are, themselves, regulated for harm reduction.
what kind of feedback loop are you looking for here?
i think Dependent Haskell is is interesting here because most models already have Haskell represented in the training set.
same deal w/ Liquid Haskell: existing models are competent with the base language & now you get SMT solver feedback.
