CredSperiment Progress Report

[burrrata]

@decentralion

Because it’s open-source, anyone who is in group (2) can choose to take on role (1) as well, and change the cred algorithm directly. (And we intend to make that as easy as possible, via parameter tuning, plugin selection, etc.) So that is already a very big difference in power relationships in comparison to a platform like Facebook, where administering (say) a Facebook group gives me 0 ability to change the underlying algorithms.

Does the SourceCred protocol/team plan to accrue value from usage of the protocol? or is the goal just to get a few grants, create SourceCred, and then release it into the world to let people do what they will with it (and then move onto another thing)?

@decentralion

“force cred”

Resistance is futile…

@decentralion

That said, the Americans did a good job of coming up with a constitution with an explicit amendment process, so that future changes didn’t require entirely ad-hoc processes, but could work within the system. So we can learn from that example.

Sounds like they learned the hard way that hard forks are a bitch to coordinate lol

but yeah you’re right, always has to be a balance between being flexible but also not over engineering problems you don’t have yet

@decentralion

Done. Feel free to make the inaugural post.

Let the games begin!

@s_ben

For the same reasons I can’t comment here

Touche’. I really walked into that one. Was worth a shot though lol

@s_ben

the online nature of all this (and its inherent ability for surveillance), makes these dynamics easier to form in some ways. But also brings the game to light. If we can design the incentives right, we should be able to see and tweak them to improve them.

Agreed. The key part is that we can bring them to light, improve them, and quickly see any problems that occur. It’s a bit of a dichotomy though because on the one hand you want users to have privacy, but on the other hand you don’t want things to stay in the dark. I think a good balance is when the system itself is transparent, but the participants in that system are free to individually disclose information as they wish. This works in the small game of the network/protocol, but in the larger game incentives can exist that lead to unintended outcomes (such as a wallet provider or exchange making users disclose information that allows an otherwise pseudonymous network to be monitored).

@s_ben

The key, as the early cypherpunks realized, was the ability to opt in to privacy. And, in this case, to opt into the system you like best. It’s only dystopian if there’s this monolithic single system. If there are many thousands of “parallel societies” to choose from, a social credit system is beneficial.

If participants are free to choose, and there’s a legitimate competitive market, then it’s healthy. It’s when the market become opaque or authoritarian that it really becomes a problem.

@s_ben

Lifetime cred should not be affected if/when you cash out. I think it’s important for people to cash out their mana tokens (or whatever) for dollars to pay bills. Otherwise poor people can’t participate at all. But if the project blows up and becomes very valuable, OG contributors should be able to re-engage the system and leverage their (now somewhat diluted) cred to earn more money or steer the system.

This is also the approach Daonuts is using so that participants can trade tokens and interact in the community economy without losing their accrued reputation

@s_ben

intrinsic motivation (what powers OSS) is increased by decoupling work from external rewards. Have experienced this myself in the Decred DAO.

I too have experienced this in the Aragon community. It’s weird… like you need rewards for the system to be sustainable, but it’s really hard to make rewards feel organic and not shift the vibe of the project. It’s really really hard to get that balance right. I think that making the “reward” more than just money, but also reputation/governance makes it more social, and thus more natural for participants. Think about it… as you contribute to a community you earn “reputation” and experience. Everyone knows that. Bragging on the internet is the biggest feature of StackOverflow and almost every other forum lol. Having a token that represents that feels a lot more natural than just “getting paid.” If you can cash those tokens out too, then all the better.

Personally I think this is stupid as money is the core tool we use to measure and exchange value, but social norms/biases make it so that for some reason that can often feel weird to talk about and act on. Not sure why that is, but it’s my latest obsession: how can healthy open source communities also be sustainable and also act in economically rational ways? The fact that as soon as money is brought into most open source projects they blow up is a huge red flag. Often people just say that money is evil, but I think it’s a sign that there were hidden flaws that got amplified by the pressures/dynamics introduced with hard value (money) vs the subjective soft value of reputation. Anyways, a bit of a rant, but I think it’s essential to explore and analyze this so that humanity and open source can succeed at scale in a healthy way.

@s_ben

Sweet! Got links? Want to follow this at a minimum.

This was Carl from Daonuts, @lkngtn from 1Hive, and someone else who I haven’t met

• https://devpost.com/software/credao

@decentralion

maybe 20 immediate / 80 long term. It’s more conservative, and we can always go more aggressive/faster later on, as we get more confidence.

Personally I love immediate feedback so I would want rewards to literally be instantaneous, but… you’re totally right that we could try out the slow version and then speed it up if need be. It has to be fast enough though that you can connect an action with a reward. If that mental connection is broken (or never formed in the first place) the game is not a game because the dynamics of the system will fail to shape behavior.

@s_ben

If you’re looking to make a MMO-style game that’s really fun and engaging and pays you (a sci-fi future I’m looking forward to), then you want more immediate rewards.

YES! This is what I want! lol

@s_ben

when I imagine myself approaching some random OSS project, with an aim to make a living from that, an emphasis on all-time cred make more sense. Where you’re new and learning, paying your dues and “interviewing” with the community, it makes sense you wouldn’t get paid much. Indeed, the community is likely investing in you. But you also want a clear path ahead of you to sustainable income. Long-term payouts offer that, and also smooth out your income stream once it starts coming in. If the leaderboard moves slowly but surely, rewarding real commitment and value added, I think that will be pretty engaging also.

Yeah that’s also a really really good point.

@s_ben

In the end, I keep coming back to this ratio as perhaps the most important variable to tweak. And one that may vary by community, and perhaps over time within communities.

I wish we could A/B test this to “fine tune” the model