[hardwax]

It's quite disappointing that Vitalik keeps pushing something that fundamentally doesn't work in a decentralized setting.

The proposed mechanism "assumes" identity, otherwise the mechanism is gamed easily. Which is what has been happening.

[flannelcatering]

AFAIK there's nothing mutually exclusive about using a service that requires identity (in this case, proof of unique human-ness) on a decentralized blockchain. In fact there are services that are actively working to create DIDs (like brightID, 3box, etc).

[flannelcatering]

One thing I forgot to add: In pockets of the ethereum community, it is widely considered that having identity mechanisms that can be relied upon is one of the things that is holding blockchain services back from going mainstream. There are all kinds of applications (decentralized social networks, payroll, voting, etc) that could benefit from having an identity overlay on a blockchain network.

[hardwax]

"working on it" vs "works today"...

[flannelcatering]

I'm not sure what point you're trying to make. But if it's that the identity mechanism on Gitcoin (aged github accounts) isnt perfect, then I think you're absolutely right.

I think a more important question is "are the round results valid?" Vitalik goes into this in his post - As Vitalik notes in his post, we are certainly reaching the limits of our existing identity framework as more money/users get involved.

Identity networks are a chicken and egg problem. My view is that Gitcoin being one of the first widely/actively used Ethereum products (https://gitcoin.co/results) we have a unique opportunity (and challenge/responsibility honestly) to bootstrap a DID network.

Here's some incremental improvements were gonna make in round 6: Identity - https://github.com/gitcoinco/web/issues/6549 Collusion - https://github.com/gitcoinco/web/issues/6548

[androolloyd]

Curious if you believe that because it's not working today it means we shouldn't talk about it as a path forward?

[hardwax]

A decentralized identity system that works will be a very significant component to the dweb stack.

But people that have experience building decentralized protocols will tell you that it's a very hard problem to solve, for many reasons.

This is an engineering community, so I believe that it's important to focus on things that exist today. Not on speculative solutions.

[flannelcatering]

I agree it's a hard problem to solve, and we've got no illusions about that.

This post (also, coincidentally by Vitalik) talks about different levels of collusion (from unsophisticated to sophisticated) https://vitalik.ca/general/2019/04/03/collusion.html

Gitcoin's approach thus far has been to stop collusion from unsophisticated actors, then progressively get more organized (as more money gets involved, and as we learn from each round) about stopping collusion from more sophisticated actors.

This thread talks about some collusion that we founnd in Round 5. https://twitter.com/owocki/status/1250097472694702080

[PretzelPirate]

can you give more details on why this doesn’t work in a decentralized setting? It sounds like the proposed collusion plan will be effective and if they need decentralized identities, those do exist in the blockchain ecosystem.

[hardwax]

Noone has solved identity in decentralized systems.

Quadratic funding allocates a pool of funds based on the number of unique donors a project has.

The proposed solution will exclude a very significant portion of potential contributors. Vitalik hints politely on that by saying that there are trade-offs between security and inclusiveness...