Nico.cast🎩
@n
Working on a ML model which will output an 'address score' similar to how credit scores work (~low is 300-500, mid 500-650, high 650+) some interesting data: me: 529.52 @shoni.eth: 567.79 @cassie: 685.06 @jc: 698.19 @dwr.eth: 663.29 @dcposch.eth: 752.81
26 replies
0 recast
0 reaction
Est3la
@wayseeker
What’s the criteria?
1 reply
0 recast
0 reaction
Nico.cast🎩
@n
The model is trained a dozens of on-chain data, like unique counterparties, total tx value, tx amount
1 reply
0 recast
0 reaction
Est3la
@wayseeker
My bad, I didn’t do a good job of asking the question. I wanted to know more about the criteria in the sense of why selecting those (instead of others) & the overall purpose of having a score. Credit scores are used to measure “trust” when asking for money, so is this to measure if a wallet is “trustworthy”?
1 reply
0 recast
0 reaction
Nico.cast🎩
@n
Decentralized reputation score is hard to get right but has use cases which benefit the whole ecosystem - governance: voting rights and Sybil resistance - artists/NFTs/tokens: vet addresses for airdrop (@0xen has more thoughts on this) Defi: trustless credit score
1 reply
0 recast
0 reaction
viksit.eth
@viksit
@juliang
1 reply
0 recast
0 reaction
Julian Gay
@juliang
tx @viksit ! @wayseeker has it right. "credit" is essentially "scalable trust". right now the default state of defi is that we're 1. pseudonymous 2. credit invisible. we all access products in the same, undifferentiated way and on the worst possible terms. when it comes to lending, that means overcollateralization
1 reply
0 recast
0 reaction
Julian Gay
@juliang
multiple applications of "credit" that's not to do with lending: - qualifying access to products eg. DAO membership, rewards, incentives - personalizing product experience eg. "excellent" credit vs "low" credit experience - capital efficient lending eg. dynamic liquidation thresholds, better terms, lower collateral
0 reply
0 recast
1 reaction
