MakerDAO Community, 26 Jan 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Community Call Jan 26th, 2021: SourceCred Update and Changes to the Maker Forum Implementation

Description

Host: David Utrobin
CoHost: Peyton Rose

Seth and Kevin from Sourcecred joined us to give an update on the Maker Forum implementation as well as some future upgrades that are in the works!

SourceCred is extremely fascinating, and has been used on the Maker Forum for months now.

It's an open-source tool that enables online communities to create community-specific contribution scores, called Cred, which measures how much value every contributor has added to the project.

The Maker Forum takes it a step further and maps Cred to Dai payouts! Literally paying active forum participants for their contributions.

Maker is a Decentralized Autonomous Organization(DAO) that manages the Maker Protocol on the Ethereum blockchain.

This video is a weekly community call that covers all Technical Milestones, Online Events, Partnerships, New Hires, Highlights from the Community, Real-Life Events, and more. Every now and then we even have a guest.

Maker Relay Ep 29
https://forum.makerdao.com/t/maker-relay-ep-29

If you're curious to learn more about MakerDAO explore some of the resources below:

Awesome-MakerDAO Resource Repository: https://awesome.makerdao.com/
Technical Documentation: https://docs.makerdao.com/
Community Portal: https://community-development.makerdao.com/
Governance Forum: https://forum.makerdao.com/
Website: https://makerdao.com
Twitter: https://twitter.com/makerdao
Reddit: https://www.reddit.com/r/MakerDAO/
Chat: https://chat.makerdao.com/home

A

All right and we're live: uh hey everybody! Welcome to uh this tuesday's maker, dot community call it's january 26th, uh my name is davey trobin and I'm joined by uh payton rose who's going to be guest hosting the call with me, uh and we have uh quite a good one for you all today. um So if you're new to the community call typically, we use this call to do one of two things.

A

Either we go over the week at maker uh kind of just looking through the maker relay, uh which is a publication, that's put up every monday and it's meant for any make or now stakeholders, including voters, integrators anybody who wants to know the current events happening around maker dow.

A

We kind of go through that publication and update anybody on the call, with with kind of an ability to ask questions and kind of do a little bit of a deeper dive into the relay, uh and then the second thing we do on this call is uh we invite interesting guests uh end, and that is actually what we're gonna be doing today, so I'm gonna hand it off to peyton, and he is gonna talk a little bit about our guest and while he talks about our guest, I'm actually gonna invite them on screen.

B

All right awesome well, uh thank you david, so much for for doing these calls like they're, really exciting and a really cool way to keep up on what's going on at maker. So thank you also to everyone. Who's joining us today, we're going to be talking about source cred, which is this really neat project that attempts to um pay people and kind of measure them for their contributions.

B

Here at maker, we implemented the source credit algorithm on our forum as a way to try and encourage participation, encourage meaningful discussion, and basically, every week, we're sending out about five thousand dies so about five thousand dollars worth um of compensation for people that contribute to us uh on the maker forum and the way that compensation is broken down is up to the source, credit algorithm, and luckily we got our source great guys here.

B

To tell you more about that, so I don't want to take up too much of their time, but uh got a high level summary. Basically we're looking to encourage people to contribute, pay them for their contributions and really get people involved in this incredible maker ecosystem, and we thought search. Credit was a great way to do that recently. We just updated one of the policy uh kind of tweaked, the algorithm for how we're keeping track everyone's compensation and that's going to be part of what we go over today without further ado.

B

I guess we got seth and kevin here from source credit to talk about us. uh How about we go and have you guys introduce yourselves first say what you typically do over at source cred. I know it's a long list, obviously, but uh give the people an idea what you're working on and who you work with and then we'll get into specifics. So why don't we start with you.

C

Seth yeah, so I'm s ben on the forums. I do a little bit of everything over at source cred, but the main thing that I've been doing for the last few months is uh working on our integration with maker and keeping the community updated on the forums and uh and working with the source. Cred uh maker working group to uh to like create new features like the ones that we're presenting today.

B

Awesome and kevin.

D

Hey y'all, uh I'm a developer at uh source cred. uh I've actually developed a feature that we'll be presenting on today, which is pretty cool and uh been a long time maker fan. uh I remember I was like just getting into web 3 when dye was the first iteration when saw what is now known as psy was released, initially, which is pretty cool.

D

So thanks for having me.

B

Yeah, absolutely we're really excited you made the time for us today, so we're going to kind of start off talking about. I guess some recent.

D

B

We did to the the algorithm to add a new policy that kind of tracks, people's contributions over on the forum and kind of the ideas about why we made that change and um what what you can expect as well. I guess um so. I guess I've been working on it with with seth, so I can help some points, but I kind of want to give it mostly to use that. Then, if you need me to fill in or want some back and forth, let me know, but I'll give you the floor here.

B

So yeah, I also wanted.

A

To just jump in really quick here and uh and shout out before you take over seth, I just wanted to shout out uh anybody. Who's here feel free to ask questions. uh There is a little ask questions button uh and we will happily ask them on your behalf. uh So yeah.

C

Cool well so, like I've convinced the dev who built this thing to present it uh and they've been they've been uh good enough to come on so they're gonna do the presentation, but uh I can maybe answer some nitty-gritty details about source cred's relationship to maker. If anyone has those.

B

Even better well, in that case cam floor is yours.

D

All right, uh can you guys see uh my screen.

D

Cool all right so.

D

Well, we are, we are source cred, as you guys know, and we're presenting uh a couple different things today we have a these are all like uh new features that we have that we think might be useful for uh make the maker community- and we want to stress that like none of this is like this is just like to kick off the discussion it's not to, uh and these are just ideas.

D

None of this has been really discussed at all, and we just want to like see what you guys think and if you guys want to implement any of this stuff.

D

Basically so uh without further ado, like uh the uh see how I can control this here, oh cool so um background, uh so maker's been paying contributors uh for their activity on the forum, and uh it's been about seven months of that, and it's been going pretty well so far um uh since june they've paid out about sixty nine thousand dollars in die to contributors um through through source, using source cred to figure out who to pay who to pay.

D

uh How much um and in september the uh maker voted to extend this trial for another three months uh and then in november maker holders voted to fund an on-chain integration for source cred, and we are in the process of uh sorting that out right, now um figuring out what that'll look like and then actually performing the integration. I think at some point after this period soon after this period.

D

um So today we want to introduce like a new payout formula and a new feature that will, we think, help incentivize like activity around important discussions on the maker forums. So first, the uh the new uh policy is called recent. um But let's talk about the old policy first, um the current policy is 80, balanced and 20, immediate balance being it uh source.

D

Cred, looks at credit earnings over all time and then allocates 80 of the available like distribution to uh all-time credit holders based on uh like based on those metrics and then 20 immediate, is based on activity and credit earned over the last week. uh For those of you who aren't uh uh familiar with like the source credit lingo. um This is good in the sense that all early contributors, contin early contr contributors reap rewards for uh getting in early, and uh it's it. It's also like quite uh stable.

D

So, like you build up like you, build up like a long-term pattern of contribution and you have some stability in terms of like your reward, but the negatives are that it's not responsive to new activity. If someone, if someone stops contributing over the last month, they're still going to see a relatively significant payout and new contributors uh need to kind of like be active for a while to build up um a substantial reward. So it's like there's like less immediate reward for starting to contribute.

D

um So the new payout policy is known as recent and uh I can go into how it works, but uh basically it's an exponentially weighted decay and that it looks at rather than like.

D

So it's similar to immediate in that it looks at the last week, but then it looks at prior weeks and then uh exponentially decays the weight of that cred to calculate like a total reward over over uh that, like decay period, basically um yeah just explained that and um it uh so it's like a nice compromise between balanced and immediate, but it has like a nice bias depending on what you set the decay period, the decay weight at to uh kind of incentivize new contributions, while still like rewarding some track, rep record of consistent contributions.

D

Basically, um so here's like an example of how that might look. You can see how like, if we look at uh like week, zero being like the current payout week. You will see um basically how the total uh payout is calculated by taking like the area under each of these curves. If that makes sense, so um you can see like with the decay rate of 0.8, um basically nothing beyond week.

D

Three is calculated in the payout three weeks in the past calculated in the payout, whereas, like all the other ones, all the larger decay rates have um or smaller decay rates have longer tails for calculating the total payout for each individual.

D

So um these are like these are like ideas for the new proposal, where we have 25 percent balance and 75 recent, so basically deprecating the immediate policy.

C

uh Kevin sorry to interrupt um so like we actually did just pull the trigger on this. Oh okay! um uh So like this, this will be implemented. Yeah, sorry this! This will be implemented this week.

D

Okay, great uh that changes the inflection with which I speak about it. So that's good um all right so, like we said like, like we've kind of shown, this is more responsive to new activity. It rewards uh new contributors and um it allows for, like one month, long breaks in that like if you, if you basically take a hiatus from contributing on the forums, you will still see a payout over like uh over four weeks before uh before, like the end of that tale, catches up and then uh so yeah.

D

um This has been trialled by one hive and metagame successfully. So there's like a proven track record for it, um and some some things to look out for uh the established members uh will see, drops if inactive, which um you know might it might not be such a bad thing and uh there's uh there there's like a motivation.

D

Change in that people might be people who are more incentivized to just collect the payout might show up more rather than like people who have like a long-term interest in contributing to maker um and there's an increased civil attack incentive since it's easier to claim a reward in a shorter time span.

D

So, uh with with this number of twenty five percent, seventy twenty five percent balance. Seventy five percent recent- this is the this is what the weekly payout change looks for. I'm not sure what uh is this like taking over the last week, seth or the week before.

C

This is actually for, um like the week before pros, put this together for the post, announcing the like proposing this payout change, but like this, this will look roughly similar to this week.

C

D

um And you all will have access to this. I think, after if it's not already present in some form in a forum post somewhere.

D

D

Yeah, so the next thing we are going to talk about is the new thing that I actually developed, which is the uh categories and weights uh tags: sorry, cat cat uh weights, weight modifiers for categories and tags.

D

um So this was uh this kind of this. That feature idea was developed within the metro, the maker source credit working group, um which seems to exist in this kind of uh transient space between uh rocket chat and discord. Am I right on that seth, or is it mostly on uh rocket chat.

C

It's mostly it's mostly on rocket chat and uh I'll give props to long for wisdom. I believe this was his main idea.

D

D

The so the way it works is you have um uh each each post on the maker forums has tags, has a tag and potentially a category assigned to it. Even the uncategorized category is still technically a category, um but uh you can assign weights to those and those weights act as multipliers combined, like they they're multiplied together to create like a total multiplier for each like that, is minted uh or created on the uh on the maker forums and through the through, like creating through like adding the combining those multipliers.

D

We can incentivize discussion around specific topics.

D

uh Here's a good example. So um if the governance category has a uh weight of 1.2 and the signaling tag used to get used to gather sentiment as some as I imagine most, if not all of you know, um has a weight of five they're multiplied together to get a total weight of six.

D

So each like uh on a post with this category and tag would have would carry six times as much weight as a um as as like a default weighted post which, where the weight where it assumes that the default category and tag weight, is one.

A

uh I have a quick question, so the current weights that exist in the in, like the current implementation, are primarily the likes that a post receives, and can you name some of the other ones just for context.

D

Sorry, can you repeat the question.

A

uh What like what weights exist for the current uh implementation? Yes,.

D

A

D

A

Like an addition uh to the weights, that are there right.

D

So, prior to this all categories and signal and uh and tags, I'm sorry are like defaulted to one, um and so every every like on every post, basically counts for the same, save for like in any source, cred factors like you, the user who liked it or like uh that kind of thing um which doesn't affect weights per se, so all weights, all weights, were one until this was implemented. Basically, does that make sense.

A

uh Okay, okay yeah. That makes sense to me.

D

Cool uh and here's like another example that seth thought up um where- uh and this is basically like the configuration syntax from the json file, where this is set up, where uh the default category weight would be set quite low to one to one. One thousandth uh default tag weight is one, and then you have these special tag: weights of 1.2 for signaling and signal status accepted where, basically, if um seth do you want to talk about, I I don't think I'm qualified to talk about the sig status accepted tag.

C

Yeah yeah, I was just gonna, try to interrupt. I want to put a giant disclaimer on this. This is basically just this. This feature is hot off the presses, and this is just the result of a couple tests I put together to illustrate how the thing works.

C

um The reason the the category weight is 0.001 is basically just to try to drop everything to zero, except for signal, request posts uh and accepted signal requests, so this will essentially uh show us what would happen if we only minted cred on signal requests and and even more cred on the signal requests that get accepted.

A

uh I'm curious, just like uh from a strategy perspective, uh should signal status accepted like warrant more weight. I I feel like, even if it's not accepted um like it, might still be a good thing that the person went through the uh went through the work of putting it together and all that stuff. So, regardless of whether it's accepted or not, and also like accepted and not accepted, isn't really uh uh like the only format for signal requests. Some signal requests uh kind of just test for preference rather than yes's or no's.

A

So I was curious about how that kind of plays into your thinking.

C

So yeah, I'm I'm glad, I'm glad that you said that, because that's good information for me to know this really is just me making up numbers, and I expect I I expect that once we once we start putting together an actual strategy, these numbers will change uh and there's also other data that we could include, um which maybe and and like I definitely like to to get some input from from, like the maker governance experts on this too.

B

But if I could jump in the basic idea behind it is by adding these tags, we would then encourage behavior so like we'd, encourage people to write, signals that that do get passed right and then source cred would pay them out more for it. um So that's kind of the thinking from where the thought experiment begins.

A

Is there gonna be a a post or like, uh I guess, uh even a signal request uh planned for thinking about the strategy behind these weights.

C

So there's definitely going to be a post announcing it. We were holding off a little bit because we just announced the change to the payout formula last week and we wanted to give the community a chance to digest that and object. uh If, but uh it doesn't, it doesn't seem like there's.

A

C

Nobody has objected yet um and so we'll probably see a post announcing this feature soon and then um yeah and then and then you know how uh you know next steps from there. I guess we'll just we'll just like you know, work with the working group and see how they want to proceed.

A

Makes sense cool.

B

Yeah, I would say the cool thing is now we have this ability so we're in that phase, where we're testing it in the background and trying to see all right. What makes sense um long for wisdom, our governance facilitator uh was the one, as seth mentioned, to come up with the idea of trying to meant more cred for um governance like essential actions, you know like making mips making signal requests things that need to happen in order for uh governance to evolve, um and since our whole idea behind source credit is, is encouraging participation.

B

It seems, like those actions should maybe be encouraged a bit more, and this is a really neat way uh that kevin came up to with doing it. So um yeah, kick it back over to you guys unless there's any other questions now would probably be a great time so feel free to voicemail.

A

We're collecting questions I I do have. I do have like uh a thought. I really really like the default category weight being set close to zero because it kind of uh it makes uh certain categories not earn any credit at all virtually which I think is a really important tool for controlling.

A

uh You know where we're trying to incentivize participation and action and also kind of cutting the fat because, like let's say, collateral, applications right collateral applications, you don't necessarily want to. You know, reward people for uh for submitting those because a lot of times it's you know somebody some representative from that company or protocol, and they really are just posting once uh and and it's a completely different uh thing compared to a signal request and and maybe more uh yeah community based and insider like discussions that might be happening.

A

So I really like that design decision.

D

uh Before I move on, one thing I want to add to that thought um is that uh if there is like a given week where um no uh special categories exist, um where everything is weighted, all all posts are weighted. uh So no, no, like uh no activity, occurs on like an especially like incentivized category posted in an especially incentivized category, like everyone will still get like proportionally.

D

Minted rewards based on like the activity based on activity relative to others, so like the fact that there's like a lower weight, won't affect distributions just to be completely clear on that um all right, we'll continue on so with this with these configs. This is what the output looked like um for for all of these contributors.

D

It's pretty interesting because you can catch a you can get a pretty good idea of like who has who's kind of collecting a form of like seniority at this point. uh By looking at this a bit.

A

Yeah one thing about that chart is ultra shoopy uh has been spearheading like a lot of the signal. Requests, for example, uh and I see that uh his bar, which is like the biggest one right in the middle, uh is the biggest one right in the middle, so so yeah cool. It's really nice to see that practically.

C

Yeah, and also I just uh just a quick clarification- uh this is the percentage change right and so um ultrashoopy is, uh is like benefiting from the fact that they've been very active with signal requests, but also that they have a relatively lower cred score than some of the like uh older contributors. So you know kind of, like you know, like a low market or low market cap coin, doesn't need as much capital to go up in percentage kind of seeing a similar thing there.

C

It's almost like.

A

A like a bonding curve right uh or something like that, where I guess it's not really that, but it's like uh if you're new, you kind of establish yourself a bit of a foundation of cred and then, as you get become an older contributor, it kind of tapers off and becomes more consistent. But then just is affected by like where your activity is right.

C

Yes and we we, we we've, we've had very fun conversations about bonding curves uh in the past, but I will say no more.

B

I was just gonna add one of the things I found so interesting. Looking at that chart was that um I expected it to. I guess, be more severe uh when, when tilting it towards uh just the signals, but I think that's kind of neat in showcasing how how good of a job source cred does just in its kind of base layer of of capturing the people that that are really involved in in governance. So.

D

B

D

All right, and then this next example is uh and like. If we need to move along like I think we can, we can uh just let me know, but uh this next example is one where we want to incentivize discussion: uh inactivity, good, productive activity around uh mips.

D

It's pretty simple, uh basically, just any any any post, uh regardless of category with the mips tag will have the likes, will account for 10 will be weighted 10 times as high.

A

So question there is a a mips tag, then there's also uh kind of tags within that category, like for rfc and conception and formal submission. uh What would that look like that would look like it would go under tag weights, rfc, and then it would be a grade of zero to ten.

C

Sorry, kevin yeah, like just maybe like quite as.

D

Familiar with the tag process, so that would be helpful that that's I'll defer to you on this.

C

David just yeah like um like just correct me if I'm wrong david, the special like the specialized knits tags will be applied in addition to the regular mips tag, correct.

A

uh Yeah exactly.

C

Okay, so so yeah kevin, maybe that's the info. You need.

A

Yeah so so for, for example, in this link here I'll post it in the chat, for example, in this link, it's a mip 32. It's a peg stability module with compound mixed exposure. It has both the maker improvement proposals, tag as well as the rfc tag, and then it has uh kind of a couple other uh smaller, like the grayed out forum tags. I don't know uh what the difference in like class of those tags is but yeah.

D

Cool yeah, this is helpful, so um I guess regardless of where uh what stage the uh the mip is in like as long as it has that mips tag, it'll have 10 times each like will be, will count for 10 we'll award 10 times as much cred um the or or be weighted 10 times as high subtle, but important.

A

um So if, but if we wanted to incentivize, let's say specifically rfc tagged mips, then uh how would you change this configuration? You would remove mips and just replace it with rfc, or would you add both mips and rfc,.

D

It depends on if you want to incentivize mips and then especially incentivize rfc.

D

So if you wanted to, um if you I think, if, for that example, I think an example. Config would be uh mips at like maybe two a weight of two and then rfc tag away of five.

A

Okay, cool yeah that makes sense.

D

Awesome uh and we have some yeah- we have a graph that shows the output of that, which is quite interesting.

D

uh It's interesting to see that all ultra shoopy is actually down on this one, even though, as you've said, they've been quite active, um a lot of people are actually quite down on this.

C

Well, we can see here that um the biggest gainer is charles, st louis, who is like the the mids master around here right, and so, if you, this shows that, if you kind of specialize in one governance action, uh you know like you'll, get rewarded for that or punish for that, depending on what we're doing so, that's why this example is not something that we would put in production per se, it's more just to kind of show how we can push the graph in one direction or the other.

D

A

So I had do you have a yeah, sorry go for it, go for it.

D

I was also going to say, like this, isn't to say that, like we can't incentivize both things, both examples like we can't combine both examples to incentivize both mips discussion and uh signaling signaling discussion as well.

A

uh In the mips boost example slide if you can go back, I do have one more question about the config, so I could make sure that I'm understanding correctly the default category weight like uh how does that exactly play into like the overall credit distribution.

A

uh So in this example, it would be. The default category would be mips right, maker improvement proposals. uh Yeah sorry go for it, I'm just trying to so.

D

Yeah, the um the default category weight is actually like because we're not setting any specific category weights. So there's like another- I don't know. If seth uh has any examples of this, I don't think I don't think there are but like instead, in addition to like so similar, like default tag weight, we have this example of tag weights uh where you can set weights for specific tags. We have a similar like config section for category weights, where you can set cata weights for specific categories.

D

So um in this example like default category weight doesn't do doesn't do anything to affect uh cred distributions beyond um yeah. It doesn't do anything to affect credit distributions, because this category weight would still be applied to the the maker improvement proposal. Proposals category uh since there's no specific.

C

D

C

So like this is a little bit confusing because what's important is just the relative weights, and so, if we want to like um so like, if we want to like emphasize mips, we just have to have the mips weight, be a lot bigger than the category rates and because all posts have a category.

C

The way that I just set them effectively to zero was to just set the default category to 0.0001, but in practice.

A

And basically yeah.

C

A

C

I was gonna.

A

Say so, basically, the design is really heavily uh focused on tag weights right.

D

Right this yeah this example is I'm just gonna show like just so. It's like abundantly clear what this would look like uh if we were to assign like a bonus weight to uh make our improvement proposals.

B

Because yeah, if uh correct me, if I'm wrong here but basically just setting the default solo, allows us to then uh relatively give a lot more to specific categories. Should we choose to do that um at any point in time.

C

Yeah and so what kevin's doing is they're adding a category a category weight for maker improvement proposal, category which will overwrite the default category rate.

C

So all categories will have this low weight, except for maker improvement proposals, category which has point zero. One.

D

It's one of the things that's a bit confusing about the four about, like the uh discourse forums, is that they don't visually separate categories from tags, uh on a post specific on a topic specifically like very well so like. If you just look at a topic, it's quite difficult, but it actually makes it makes much more sense.

D

If you're looking at like the tags or categories focused ui, um just to be yeah.

A

D

Sense quickly on, like any clear up any confusion on that front, because if you look at the mips the mip 32 topic, it looks like maker improvement proposals and mips are basically in the same kind of like feature: space and they're, not so okay. uh So this because of that change like this, this is a little bit different now this, uh but this applies to like the prior config.

D

um So then this is like a an example of the balanced. uh The both end example that I kind of mentioned where we have signaling incentivized and mips incentivized. um I think it's pretty. I think that there's some, uh I think, there's a little bit of like.

C

So, yes, sorry kev, I'm like triggering the developer because I'm just hacking up their config file so, like I, I just added this right before the call this. This is not to be taken seriously. This is me just taking a stab at like gener, like generally incentivizing the important governance actions.

C

So, okay, we want to emphasize signal requests. We want to emphasize mips and then uh the category weight is one, so we're not dropping everything else to zero.

C

A

Interested so the way that this would basically work out of the gate is that all the categories are equally uh sort of weighted and uh being weighted, very, very low, uh and that yeah, as I said earlier, like the where you want to incentivize, is really how you manage the tag weights right.

C

So there's the design space for this is quite large, and I think that it'll take a little bit of time to to explore this and figure out what works best. It could be dropping some category weights to zero effectively. It could be leaving them higher.

C

You know like it really depends, so I I want us to like not take my example too seriously: fair enough gotcha.

A

D

uh And so this is.

D

An interesting, uh an interesting visualization into how like most people would drop most contributors would would lose a percentage of their cred if this was applied.

A

In one one example right, rich brown, uh who, uh who you know retired from the foundation and has been uh kind of taking a hiatus from maker, has been almost completely. uh You know inactive on the forums and you can see he has that like that biggest bar down.

A

uh So I imagine it's probably more of a function of inactivity than anything first as well. For example,.

C

Oh, yes, I mean in exit. Basically I think, since rich stepped away, there's just been a lot of new activity, around governance, and so the percentage of signal, request, posts or whatever that he interacted with is lower um yeah.

D

Cool uh all right, and I think that yeah that's it. So uh I guess I see I see andy's comment, but any any questions that you guys have.

B

This is kind of, like probably a very difficult question to answer, so I apologize for asking it, but um now that we have these tools right to to mess with our weightings like how? How do we figure out what works exactly uh like? What are your thoughts as the dev, in terms of how do we analyze as we're tinkering with these various models, like what.

D

B

D

D

I have like a skewed perspective as a dev, because it's easy for me to think, like oh just play with the instance play play with the play with the maker dao instance, and see what the maker doubt community instance and see what comes out, but it's not as trivial as that.

D

So, like my, the way that I see this happening is like very interested community members who want to kind of incentivize, certain things would say like could could go in apply these weights like seth has or there could be someone assigned to apply these weights like seth, has and produce like outputs, and then there could be like a topic so basically like a meta topic on what the source on what like tags and categories should be weighted and what the desired.

D

So first, like sorry, I'm getting ahead of myself but like first, there should be a discussion about like what exactly you want sent incentivized and then there can be like a a subsequent thread or maybe follow-on discussion on the same topic about what the configuration would be. That would best meet that meet. That uh community, like community, agreed like uh incentivization.

D

Good question: andy.

B

Yeah, so in the chat, uh andy notes, that is anecdotal observation- is that signal requests often get a a decent amount of credit already. So is this something we should really be considering boosting and that's um kind of speaks to the philosophical nature of you know: how do we want to use this tool.

B

But one one thing: I guess I I'll mention that I liked about your answer there kevin is that it really does give us the freedom to say hey these. Are you know these are really open boundaries we can play with so we can. We can take a stab, we can say hey. I think this will be good and, as the weeks go on uh dies distributed to our contributors, people might say: hey. You know what I don't feel. This is fair.

B

What if we tried it a different way, it would be really easy to run some sims and see how it would change things.

D

One uh now that, now that the recent policy has been implemented for uh maker, it's like um the the weights can be viewed as like ways to further incentivize like fresh takes from new people on on new discussion as well, uh where it's like you know, because now the idea of like you, don't need to like buy your time in maker to start seeing like a significant reward.

D

If you come in and you have like a lot of good ideas, you'll be rewarded for that immediately, so um these weights can now be a way to like further incentivize, like, I guess, fresh perspectives as well, which is cool. Now that recent's being used.

C

I would also just add sort of like um one thing too um andy's question uh like I. I think there's going to be disagreement on, of course, like on what what needs more cred, what doesn't, um for instance, I saw the governance and risk channel uh the other day uh said ventures.

C

uh You know saying uh you know who wants to do a signal request for this I'll I'll, give you some source credit.

C

If you make a single request which, like you know, which I think reflects the fact that some people view signal requests as maybe kind of like more tedious, uh where or like where they need the work where they need extreme, more extrinsic motivation, whereas, like you know just spouting off your brilliant ideas about monetary theory, uh you know is like just is like intrinsically enjoyable right, but um at this stage I I I also just want to emphasize that you know like we're not making these judgments. Yet um you know signal.

D

Requests aren't.

C

B

And with that, I think uh we would definitely say we definitely invite um perspective, because that's one of the neat things about the dao is, you know everyone can really get in and give their opinion and work together to to try and make something better.

B

um As a new contributor myself like I've, I've been just blown away with how easy it is to get involved and be guided by some of the older community members, and um I don't know, I would warn you that source credit is a very uh addictive passion once you fall into it, it's like a very cool algorithm, so you know tread lightly, but also like if, if you fall down the rabbit hole like you'll find me down there too,.

A

Also source cred does a great job of uh of really tackling the apathy problem that a lot of you know. Decentralized governance based systems are worried about uh and having.

A

uh Okay like having voter apathy and actual votes is one thing, but also having apathy in public discussions about the things being voted on is like a whole other thing, and so I you know from the second, I personally saw source cred uh and rich, tackling it and getting you guys to do a forum implementation for maker dow. I was super hyped uh because yeah again, this is a really really really great tool for the casual contributor for the full-time contributor and for the dow as a whole. Yeah.

D

This is a bit of like a show, since we have the platform right now totally.

D

um We we want to see this idea of, like uh cred, rewards supplanting salary in that, um if we can, if we can tune the balance well enough, like we can have the system where your base, everyone is commission based, but they still have like a comfortable base income so that that incentivizes, like strong behavior or strong contributions, uh inconsistent contributions um and that way like the idea of like negotiating salaries, your salary, depending upon, like you, know, 10 minutes of discussion or half an hour of discussion in a room one on one with someone like that's gone, we want to see that disappear where, like the community, decides what everyone's contributions are worth as opposed to like a few people.

D

So um that's kind of like a vision that we have for for this for this tool.

C

Yeah, although I I I would say, that's a long term vision, just so like just so everyone's clear, like source cred, isn't like ready to take on like what the domain teams uh or core units are doing. um It's it's sort of blind to like a lot of labor uh and but but uh but yes, that's. This is the larger vision.

A

I asked it in the chat, but I might as well repeat it out loud for the recording, uh but I'm curious what the mode payout right now is, which is like the most common level of payout and then also like the overall average of payout, and also how many people are actually opted in, because I know that uh for the for the maker implementation you actually have to opt in in order to like get the payout. So I would love some metrics.

C

uh So yeah um in the last distribution, 62 contributors received die.

C

And I'm gonna drop a link uh to the maker instance showing the amount that's been distributed so far, my computer's being nice and slow.

B

And I guess worth noting one thing that would kind of mess with the averages, as we have increased the weekly payout as we've been going along with it. So may not be the best representative number but kind of asking for the different statistical uh factors. There makes a lot of sense uh to give people an idea kind of what what it's doing for people.

A

Because I was, I was wondering if it's going to supplant like salaries, for example, what does a monthly like payout do to like an average person? You know at the average person, so uh so does one cred uh map to like one die here.

D

There's no uh there's no, like cred fixed cred per die. It's basically it's based on relative activity so like uh in a given. I guess like if we look at like the recent policy um in a given like week, uh the the the policy looks at the total amount of cred generated and then start and then uses that to kind of like approximate the pot and then proportionally awards die based on each individual. Each contributor's proportion of cred at that point understood got.

C

It yeah so like, um and this is just one way to do it- we've discussed having like setting setting a target die per cred uh in like internally in source cred, uh although, although our dye is grain, um they're they're they're, like just literally a million directions, you can slice this.

D

And I also want to add that, like the even though like that might sound unfair on a week to week basis, but the whole reason you have like the decay is to catch those under-rewarded contributions and also balanced, will also make up for any long-tail unrewarded cred as well. So that's all covered basically over time. So over time, like all cred is equal.

B

Yeah, um it would obviously also depend on that weekly distribution like right now you could think we're. You know. Spending about 20 000 die per per month, so um you know that's 240k a year, probably not enough to really support too many salaries at that level.

B

However, um as people get more comfortable as uh people get more into the idea of source cred, if we were able to increase that level all of a sudden, it becomes something that that can support that, um but right now we're using a relatively small amount of dye, but I'm finding that it is still incentivizing people to to really get on the things and looks like seth had something to add to that.

C

uh Yeah, just real, quick and david. Let let me know if we're going over time here. I know we've been talking a lot.

A

No, you guys have have the whole call yeah.

C

Don't worry about it um as far as as far as scaling up source cred. I just wanted to add a couple points.

C

One is that as you as you increase the payouts you're going to increase the uh reward to civil attacks right now, we're very comfortable um in part because we implemented a feature for maker that, where you you cred, isn't minted uh until you gain a higher trust level in discourse, so people brigading, uh you know like their coin, don't mint any cred actually, um but you know if we went up to you know like if we're distributing a million dollars or something and and someone you know like the incentive- might be strong enough for people to start doing that.

C

So then you know there's that to consider, um uh and then I forgot my other point but yeah there's like there's more issues to come up. If, if we want to expand it.

D

My technical brain was tickled by that by that feature, because, like one of the cool things about uh source, credit is like when once a contributor reaches that trust level- and I could be- I could be wrong about this, but I'm pretty sure this is true.

D

Every contribution that they've had before they reach that trust level has gone back. They go, they get cred for that all those old contributions.

D

So it's not like you're doing work for free you'll, eventually get rewarded for all those things and then you'll have like a back payout because of all this unrewarded credit up to that point, seth.

C

Another nice, uh like sort of like feature of that, is, if you help a new contributor uh and like you're, getting likes from them, because you're commenting on each other's stuff, those likes won't mint any credit first, but as that person's credit, score and trust level increases.

C

Those old likes turn on start minting, cred and you're gonna get a boost, that's kind of like uh it's almost like a recruitment bonus. If you will right, that's super.

A

Cool, I have a.

C

A

uh So I know that, right now there is a mipset that is being proposed called the core unit framework uh and it's meant to kind of uh uh make it easier to yeah have teams that are like responsible for certain functions, and I was curious whether the source cred working group at maker uh is, uh is planning or or considering, creating a core unit to manage the source. Credit implementation.

C

Yeah I've been following that, so we had a. We had a mip um back in november, uh uh where the mkr holders voted to fund an integration uh directly with the protocol, so people will be paid die directly uh using source cred from the protocol.

C

How, like the details of that, uh I think, are still a little up in the air I'm closely following this core unit um uh like mipset and also the work with the keg module uh like. uh I think I think what it could look like in the medium term is basically um maker has like a representative from the community who can like manage the source.

C

Credit instance and interface with source cred source cred will figure out a way to like more trustlessly transmit the scores and payouts uh like to like two maker, uh and then I I could imagine source credit becoming like like a core unit or uh if that doesn't work, something else um to get paid directly from the protocol.

C

That's my sort of like my just like 20 000 foot view right now. Yeah yeah.

A

And it has to be 20 000 feet because, obviously, like the core unit framework isn't even like formally ratified. Yet it's still in request for comment, and on top of that yeah there's a bunch of different ways that uh payouts can go. They can come straight from the protocol. They can uh work like a grant from the protocol and then the core unit does the distribution themselves.

A

uh Yeah, there's there's a ton of paths but but yeah it's cool to know that you've been thinking about it then.

C

Yeah totally, and um it's also very exciting- the idea that people could be paid directly from the protocol via an algorithm. uh You know: that's that's a pretty big step forward in terms of decentralization.

A

Feel like we're in the future, man.

B

Super exciting.

A

Waiting for this day, my whole life.

B

Looking at the time here, we are running up against it, so one last call for questions. If there is anything, you've been wondering like shoot it off. If we're not able to answer here, we can connect with you afterwards, but um thank you guys so much for for coming on and like spending so much time to make this happen. I know, for at least me personally was a big part of me falling into that maker orbit was saying. Oh my gosh.

B

I made fifty dollars for for that first post I made like what maybe I need to come here every day.

A

That that's that's actually a amazing testimony. Peyton.

B

But now you're stuck with me so.

A

Oh, no, that's that's a good thing which, by the way, just to shout out peyton payton, has been involved with just a number of different working groups and and things at maker dow, since he started contributing and he's really been uh like. If I could, if we had like a reward for the rock star community member of the month like it would go to peyton hands down.

B

I would defer it to at least three other people, but, like that's, really sweet. Thank you. Yeah.

A

And that's how you would get it.

C

Yeah now uh yeah I'll, all I'll, plus one peyton's rockstar nomination, they've uh yeah they've, been they they've, been going down. They've been thoroughly source source-pilled, as we say, and uh uh they've like they've they've, been really good interfacing with storagecraft and understanding everything.

D

If, if you want to like learn more about the team at source cred, we have a call in an hour by the way, not now we're getting a little over chilly here but like come. Come hang out with us.

A

Over shelly is welcome, so yeah throw the link in the chat. uh Definitely do that I'll, even throw it in the maker chat as well.

A

What's the call about by the way it's just.

D

The journal, it's a community, call basically it's centered around new people who want to ask questions and we try to answer them.

A

Nice and then andy's asking uh what what is that hat? uh Oh.

D

uh It's it's a it's a story, I'm basically a big baseball nerd, and this is like a minor league team and sometimes they they do like special hats. um It's a it's a picture of a sandal for those of you who don't know so. uh Thank you, yeah. I think I thought it was like delightfully obscure and bizarre. So of course I had to get it at that point. So yeah.

A

A

Well with that uh feel free to make any like uh ending remarks, uh but I just want to say thank you both uh for coming on the call. This was wonderfully fun and informative and and makes me hype about about maker, about the forums about contributors. So uh yeah props to you guys.

D

Thanks for having us it's a lot of fun, we obviously like this is there's a lot going on in both of our communities and we can geek out about it forever. So it's cool to have the space.

A

Right on and peyton, thank you so much for co-hosting.

B

Happy to hop in.

A

On happy tuesday, everybody I'm gonna, hit the end broadcast button. Then, since I don't see any questions in the chat and so yeah have a great tuesday, everyone.