Tarun Chitra
@pinged
IMHO probability in crypto and AI in 2024: Convexity: E = WC Online Learning: E = o(WC) Auctions, Staking, RL: E/WC <= Poly(1/n), Expected/Best Case >= c’Poly(1/n) Neural nets, PoW, ZK, FHE: Pr[|X - WC| < ε] <= exp(-E) MEV: BC - E >= 1/n^c WC = Worst Case E = Expected BC = Best Case n = dimensionality
7 replies
7 recasts
36 reactions
Nick Cannon
@inkymaze
some takeaways (I think?) - terminally online ftw - opsec has a tailwind - searchers have more searching to do
0 reply
0 recast
4 reactions
