December 28 DAO Call

*Assembly AI generated transcripts may not be 100% accurate, listen to video for original audio.

Recorder: DJSTRIKANOVA Author: Summaries auto-generated by AssemblyAI

Assembly AI Generated Summary

- Last DAO call of 2022. We want to basically give the DAO power to do certain transactions. Jesse has been working on arbitrary transaction proposals. The first thing that we're going to use this for is for basically electing the High Guard. - What are going to be the next things that we will implement for the arbitrary transaction proposals? Could be a bit more related to the DAO itself. Could also have a DAO requester account on Effect force with some funds in it. Creating a campaign might be a little bit harder in the UI to make.

- There should be a way for requesters to reject work that the workers have done, but then we should decide what happens with those funds. Of course, I think we would need a reputation system on top of this.

(Rochelle) - One of my plans is for the DAO to have its own dAPP on effect force where requesters can upload their data and their results and pay for us to validate their work for them. With the DAO open for anyone to join, once we get going with a bunch of requesters and tasks on the platform, we're going to be attracting a lot of workers.

(Rochelle) - DAO is doing a pilot program for DAO members to be validators for image labeling tasks. The company needs a way to directly publish campaigns for these validation tasks. It's going to be needed everything from posting campaigns to signing qualifications to flagging workers. This will set DAO apart from other platforms.

Assembly AI Generated Transcript (Speaker <Letter> replaced with name)

Transcript

Rochelle

DAO call of 2022. I can't believe it, because we're heading up to our anniversary on the 1 January. So it's very exciting and there's a lot that's been going on. And the one thing lately that really affects us big time is the ATP. And as Lawrence just mentioned a little bit ago, he would give a little bit of an overview and talk about it. That's the first thing I have put on the agenda today.

Laurens

Yeah, for sure. I'll expand a little bit on what it is, what it can do for us, and what the dangers are of it. Right, so this has been something that Jesse and me have been talking for a really long time. Even before we started this DAO, we already thought about some point the DAO could actually, basically do transactions in a decentralized way. So normally there's like one account or one public key that's allowed to do a certain transaction on certain smart contracts. You can have like a multisig wallet, of course, to make things more secure. And there are many ways how you can make transactions. You can use depth, you can use blocks, you, you can program transactions yourself. In the end, the transaction is basically a set of instructions that you want to do and you need to put your signature on a transaction, and then you can send that to the blockchain. And then the blockchain will verify the smart contracts will verify whether all the changes that you want to make, if you actually have the right signatures for that, that's basically simply what a transaction is. But now we want to basically give the DAO power, the control to do certain transactions. But yeah, the DAO, of course, works differently. It's not one person, it's a whole stack of smart contracts. Right, so the DAO basically is a stack of smart contracts and it's controlled by all the DAO members. There's proposals, there's the voting system, there's all the members that's all on chain in smart contracts. And of course, the DAO already has like a mechanism to basically transfer funds to people, to proposals. So people can already kind of make transaction like a really specific type of transaction. The transaction is, I make a proposal, people can hold on that proposal and if the number of votes passes, then the smart contract, the DAO smart contract will release those funds to the one that made the proposal. That's basically the only transaction that a DAO can do currently. But Jesse has been working on arbitrary transaction proposals, which basically means that instead of the specific one proposal to request funds, a proposal can basically be a raw transaction that meets the signature of the DAO. In the end, you can have a proposal with like a raw transaction in it and it should work the same. People can vote on it, the DAO can vote on it, and if the vote passes, then the DAO smart contract can actually put its signature on that transaction that's in the proposal. And then whatever power that signature of the DAO has, that's something that the DAO could change. Right. So we could give the DAO the DAO smart contract. We could give it the permission to change effect force fees, for example, or permission to change things in a DAO itself. Like, over time, we can give the DAO more and more permission. And it's not like one person has that permission. It basically means that only the DAO smart contract has the permission, and the DAO smart contract will only put its signature on it if the vote passes and the High Guard doesn't stop the proposal, the arbitrary transaction proposal. So this opens up a lot of possibilities. Right. The first thing that we're going to use this for is for basically electing the High Guard. So it's a change within the DAO itself. Right. So the DAO smart contract has the authority to change the High Guards, but it can only be done through an arbitrary transaction proposal to change the High Guard. So the High Guards cannot change highgardes themselves. They have to go through an arbitrary transaction proposal as well to be able to change the iGuard. So the only thing that the High Guard can do is they can stop arbitrary transaction proposals. And that's basically in the beginning for me. I see it as a safe measure that we need. Because even though this is a powerful tool, arbitrary transaction proposals, it's also a bit of a dangerous tool, especially if we allow for custom arbitrary transaction proposals. So if we already know beforehand what exactly the transaction is, then it's not that dangerous. But if it's a custom transaction, yeah, it could be many things. Right. It could even be a bit hard to see. It might take some time to be able to see what the transaction will actually do. Right. And you could make a text in your proposal that's something completely different than what the real transaction that you also posted with your arbitrary transaction proposal is actually doing. So, yeah, we have to make sure that we still have the ability to prevent malicious transactions that could basically take over the entire DAO. These sort of things is what the High Guard needs to watch out for and make sure that those things don't pass and don't slip through. But for me, I don't see it that the High Guard has to make decisions whether they think something is like, they want something, yes or no. It's more like, is it dangerous? Yes or no? If it's, like, dangerous and malicious, then they should prevent it. But hopefully they never have to prevent a proposal from going through. Right. So that, for me, is an important piece of what a High Guard should be doing. That's also a bit where the danger is in arbitrary transaction proposals. But once we have it, then it can open up so many cool possibilities that we've basically been dreaming of for quite some time now, from basically the beginning of the DAO, that we could do these things with the DAO. So I'm really excited for the arbitrary transaction proposals to go through. The first thing that we will vote on and that we will use them for is the actual High Guard election itself. But I also see that Jesse joined the stage. So I saw that Jesse has been working on it in the Smart contract. So maybe he can expand a little bit on how far along he is and how he is implementing the arbitrary transaction proposals.

Jesse

Hey, sorry for being late. I don't have a lot of time to talk, but I would love to listen in a bit. I could quickly go over that as well. I heard Lawrence, I think, talking probably mid story, so I don't know what the original question was about. I hope you guys are all doing good. Merry Christmas. Good to see everyone here. So, yeah, to quickly update you guys on the progress there, so I programmed part of the feature already. It's in GitHub. I've been testing it. It's a bit tricky to test because of yeah, it's just quite intrinsic feature to do the multi sick transactions on EOS and it interoperates with this EOSIO MC contract. So I've been testing that out. But yeah, the feature works like Lawrence described. So we will have it in place so the proposal can be in the next cycle, which is on Friday, I think Friday is the 30th already of December. So we'll have it in there and that would assign the High Guard members that have been voted in. So it's a bit of an odd concept now because we already did the voting on who's going to be in there and then the transaction that actually puts the high card members in the EOSIO authority, which is a technical place where they are registered that will be voted on separately. But then after that cycle, the transaction executes and people will actually be instantiated. So that's the process.

Laurens

Yeah. Nice. That's really cool. It's good that you made changed already. That's good to see. And I'm really excited to see it in the next cycle play out. Really fun.

Jesse

Yeah, me too. We're going to see it in action. It's going to be exciting.

Laurens

If anybody has any questions about how it works or because it's like a bit technical right, the arbitrary transaction proposal, then yeah. Then please just let me know. It can be here, it can be through DM. I'm happy to talk to people about it.

DJSTRIKANOVA

Do you hear me?

Jesse

Yeah, I hear you, DJ.

DJSTRIKANOVA

Oh, okay. I thought my microphone was broken. Did you hear my question?

Laurens

No, I don't hear you, DJ. It might be me.

Jesse

I can hear DJ. But maybe learn you were like, cutting him up.

Laurens

Sorry, I cannot hear him at all. Voice. Let me rejoin.

Jesse

Weird maybe it's something with this card.

DJSTRIKANOVA

Can you hear me now?

Laurens

Oh, yes, now I can hear you. Sorry, I didn't hear you at all. So I'm sorry if I spoke right through your talk.

DJSTRIKANOVA

Sorry, no, sorry. I thought my microphone was broken. What did I ask completely for? Do you remember, Jess?

Jesse

You were like mid sentence. I forgot what sentence was. And also you didn't finish it.

DJSTRIKANOVA

I guess the question I had okay, I remember now. What I asked was so the High Guard, they have to explicitly approve the ATPs, right?

Jesse

Not explicitly approved, but they would have to have this period between voting and when the cycle ends to reject it still. So they would have to explicitly reject them if they were voted.

DJSTRIKANOVA

Wondering if that could be like a vulnerability in case there's like some sort of spam attack.

Jesse

Only if the DAO actually votes in the proposal. So if someone would spam it, then the DAO would have to approve. They would need to get majority votes yes to end up in the highguard.

DJSTRIKANOVA

Yeah, Lawrence actually mentioned in a postal exactly. What I was thinking is because the quorum is much lower than majority, right?

Jesse

So quorum is the minimum amount you need to actually even make it to be accepted. And then you also need 51% yes votes or 50 plus I mean 50 plus percent yes votes.

Laurens

Yes. The majority that Jesse is talking about is the majority of the people that voted. And there's a quorum to make sure that you need enough people to vote. But the quorum is not, I think, more than half of all the DAO members. I think the quorum is lower.

DJSTRIKANOVA

The DAO side is 420,000. Right. So in theory you would only need 420,000 vote weight. And then you just programmatically send like, say 1000 ATPs, then programmatically vote yes for them. And then what are the High Guard going to do in that case?

Jesse

If you have the majority of vote weight, then yet it's possible.

DJSTRIKANOVA

You will need only the majority. Like it will basically be a proposal only they vote on. So they have the quorum and they basically have 100% yes because there's only one vote and they could schedule it right before the voting ends.

Laurens

Well, they couldn't schedule it right before the voting ends. Right. Because it's per cycle. Right. So it has to be part of the cycle still.

DJSTRIKANOVA

But then even if it's just I feel like if we don't have to approve it, I feel like it's just a spam could. For example, what would the DAO do if say like 10,000 proposals were made, right, by this one malicious actor with enough vote weight to have a quorum for themselves? Would we all have to manually try to vote no to stop it? Because I feel like we could get overwhelmed.

Laurens

Yeah, we would have to vote no to stop it.

DJSTRIKANOVA

I know sometimes go ahead. Sorry.

Laurens

Well, there's a cost involved to create a proposal, right? It's basically to stop spam. So this can also happen with the current proposals, right? Also to get funding, they could try to spam it. And I think the cost is pretty low right now, but I think we always had this in mind. Like, if there's really people that are trying to spam it, then we need to increase the costs of proposals to make sure that it's too expensive to spam, basically. Right? But there's a couple of things so we have to make sure that the cost of the proposal is high enough so that people won't spam the proposals that show up in the DAO. But then Jesse was also saying in discord somewhere that we will not show basically custom arbitrary transaction proposals right, in the UI in the interface. But it's a little bit dangerous because then the quorum comes into play, because then they could have like, a hidden proposal with a transaction in it that we don't see. So at least the DAO members don't see it in the UI, so they're not going to vote on it. And then indeed, they would just need to have the the quorum and they and they could they could vote on it themselves and then people would even see it in the UI and then it could pass if they have the quorum. So, yeah, it's two things we need to look at. Is the quorum high enough and is the proposal costs to make a proposal high enough?

DJSTRIKANOVA

Would you be able to set the quorum differently compared to the standard proposals? Or does it look all the same under the hood?

Jesse

It's all the same. So yeah, it's a good point. I was thinking it's a good point.

Jesse

Certainly to protect from it, we need.

Jesse

To have the proposal price needs to be high enough and the poor needs to be high enough. And if someone would spam it with a lot of transactions, right, then the High Guard could do a lot of transactions to sort of reject them. But it's a good one. The price and the quorum is the two factors we need to consider. We could actually have it, like maybe start with having the if that happens, I mean, the High Guard could sort of override, I guess, still, it could definitely reject just all those proposals, but it might be a little bit expensive. And for the next cycle it could deploy like a transaction to fix the quorum or change the proposal costs or something. Yeah, it's a good one. It's a really good one to bring up. I wonder how feasible it really is to do that. We should calculate some costs on how hard it would be to do.

Laurens

Go ahead, Rochelle.

Rochelle

I was going to say I have two things I wanted to ask about with the ATPs, but keep going if.

Laurens

You'Re not done, please go on. To finish this part off, then indeed like Jesse saying, we could just try it out with still the lower quorum and the lower proposal cost because we still have the iGuard and we still have the possibility, even if there's many proposals, to make a script, to reject them all. That's all doable. But then we would have to fix it and increase the price and the quorum for the next cycle if that happens.

DJSTRIKANOVA

Well, let's say yes, the price, definitely. I feel like we may want like a way to programmatically end all, like I guess on your end. Like maybe if only one High Guard votes against or tries to stop an ATP, then it won't go through. Basically the High Guards themselves, the High Guards have to sign the transaction to stop it from being done. Right. And it's a majority of them, right?

Laurens

Yeah, exactly. But we could make you a script, I suppose. There's this female tech happening and it's like 1000 proposals, right? Then we would write a script. That all Iguards could use to mass reject them, right, and we could share it with you. It would be like another page on the DAO that's just for the iGuard, where we could mass reject all these proposals or scripts that you need to run. So if that happens, and if all those proposals are passed so people, they may be managed to get a quorum or something and they all pass those proposals, then there's a bit of work involved. But it would not be that hard to create something where we can mass reject all those proposals, for example.

DJSTRIKANOVA

Yeah, I think that would definitely be necessary because we don't want to think about increasing the price by the time if it already happens. Right, then it's too late.

Laurens

We don't want to increase the price right away.

DJSTRIKANOVA

I'm sorry. Go ahead.

Laurens

I was just saying it's not too late, right? Because we can still the highguard can still intervene if it happens. So yeah, it's a bit of work. We have to make the script and share it with the High Guard and then they would all have to use it to mass reject it. So it's not like a disaster if it would happen. But you don't want that to happen every time, right? Because it takes work. So if it happens once, then we say, okay, then the price was too low, right? Or the quorum is too low and then we can still increase the price in the quorum.

Jesse

Yeah, it's a good one. And definitely there is a cost. This is indeed it's a tech sector that we should consider. But the cost of that would definitely be quite high. It will be pricey. But then to counter it, it would also cost a bit of High Guard resources, like they would need to do the transactions. That would also cost a bit of money way less than it would cost a Duty attack. Like it wouldn't cost effs and it would cost no rent, which would cost quite a bit of ram to create ADPs. It's not a cheap thing and the High Guard would only have to sign like transactions to just reject them, basically. So it would be significantly cheaper for Highgard to counter it than for the attacker to do it. So it's definitely going to be possible to deflect it. But yeah, it depends how much funds the attacker would have, I guess. So it's good to at least consider.

DJSTRIKANOVA

It.

Jesse

Like if it happens, yeah, we should make sure that it's at least doable for the agar to deflect it.

Laurens

Right.

Jesse

So I guess we need to do a little calculation on price, cost or something to be sure about it.

David

So I have a question about the different scenarios when it comes to the High cards. How broad of a scope is this first ATP going to be with making for certain scenarios such as removing a high card member, is that included in this first ATP or is that something that we need to define further DAOn the line?

Jesse

The first ATP will include like assigning the initial High Guard so that's I don't know how many people have been accepted now, but they would be assigned to the High Guard permission in one transaction in the next cycle, that's the first B. And then the High Guard, there will be two transactions that can be done afterwards, basically, not in the first cycle of the SB, but in cycles afterwards.

Laurens

I lost Jesse, but maybe it's just me.

Jesse

I had said that I also have to go almost. I'll just try to quickly answer this and then I really have to sorry, but it's unfortunate, but I do have to go somewhere. So anyways, but for this one afterwards, there will be two actions you can do as an ATP, which is to add a member or to remove a member. So that's basically two actions that will be available in a smart contract on the High Guard side. So this means add and remove, those will be in the scope. Afterwards, they will also be open for review before we actually deploy those two actions. So on day zero, we don't have these actions yet, but they are basically already programmed for deployment. So if the High Guards agrees, these two are added, like add member, remove member, and then that will be the initial primitive, sort of naive way on how to remove and add High Guard members. And then we would have to think about some other more advanced cases and.

Jesse

Also.

Jesse

Conflicting cases like we described some of them in the proposal in this court. Like if someone adds and removes one and both get accepted, like what's the order of execution? Like those cases we need to figure out. But that's more like, I think, advanced cases. And also if you would want to quit as a hybrid member yourself, maybe you shouldn't be able to wait for a cycle. You should just be able to do that more straight away. But yeah, those are all topics good for discussion. I do have to drop off guys, so sorry for being really short. I will be in discord, so anything that pops up, please let me know. I hope there's a recording so I can listen back on this and give some more thoughts because I think it's a really important fruitful discussion. But I really have to leave now, so sorry. Have a great dial call and I'll be in this court and discuss more with you guys about this.

Bree

Hi Jesse.

David

Have a good one.

Rochelle

Bye bye.

Jesse

Thanks, speak to you soon.

Rochelle

Well, what I wanted to talk about, Lawrence, with this was with the ATP, this will allow for automatic cycles to start and stop. Right? Just you won't have to manually do this anymore.

Laurens

Yes, he's also been working on that. I'm not sure if that's exactly the same as this. He's also working on improving that.

Rochelle

Okay.

Laurens

That's also a really annoying thing.

Rochelle

Yeah. And the next one was right now we just have the ATPs for planning to add in for adding a member and removing a member. But this should also open up the ability for the DAO to be able to directly create campaigns, like make campaign, add tasks, assign qualification, remove qualification and whatnot, right?

Laurens

Exactly. And that's also I think we could even make maybe a channel for it. What are going to be the next things that we will implement for the arbitrary transaction proposals? It could be a bit more related to the DAO itself, maybe first, like, what's the price of a proposal? That could be something that could also be determined by the DAO, right? So you could make an arbitrary transaction proposal to change the cost of a proposal for the next cycle. So if your proposal will go through, then the price will be changed in the next cycle. There's the only problem is that, like Jesse was saying, suppose there's two arbitrary transaction proposals, one that lowers the price and one that increase the price and they both get voted in. Yeah. Then what are you going to do? Right? Is the price going to go up or is the price going to go DAOn? For a proposal, for example, we have.

DJSTRIKANOVA

To make sure block order that decides.

Laurens

Yeah, but the block would be the same, right? How do we determine which transaction would go first? Right? Suppose they have the same number of votes. Then maybe the High Guard has to decide, right? Like, which transaction do we put first in the block or.

DJSTRIKANOVA

Ends up with? Yes, it would already be something, I guess.

Rochelle

Would it just go by like when the proposal was assigned to the cycle?

Laurens

Yes, but that doesn't really matter. It shouldn't matter when you assign it to a cycle. Right? Because the cycle starts at the exact same time on proposals and then they are either assigned to it or not. Right, so it would be a little bit weird to do it like that. So I guess for now you could say if there's a conflict, the high guard can decide which one goes first. Maybe they can, but that will maybe be a little bit complicated. Technically we have the DAO decide in some way, but that might be a little bit hard. So it's something that we have to discuss. But that's probably a discussion we need to have when Jesse is also there because it's also a little bit technical. So we need to know what's possible also on his side. But that's one of the challenges that we still have to fix. But it's already really fun to think about all the things that we can do with it. So we can change basically settings of the DAO, the cost of a proposal, the length of a cycle, all those things could be changed, right? That's all DAO related. But you could also have, for example, give DAO the power over a couple of effect force settings. Like what's the fee that's coming in from effect force to the DAO? That could be something that the DAO votes on. And we could also have like a DAO requester account on Fxforce with some funds in it. So we could have proposals to fund that account and we could have arbitrary transaction proposals to actually use that account to make campaigns and to add tasks. Right, so you can have a proposal to add a campaign or you could have a proposal to add a task with DAO funds basically on effect for us. So yeah, that's all things that we could implement. Like technically Jesse is going to make it in such a way that basically all transactions would be possible. But then of course, we would need to make it so that the UI only supports specific types of ATPs and that we have good forms to also create those arbitrary transaction proposals. So for example, the changing of the price could be a very simple form, right. Just you see the current price and you can put in a number for the price to create a proposal and then you click submit and then that's your arbitrary transaction proposal created to change the price. But some of creating a campaign, for example, might be a little bit harder in the UI to make. Right, so yeah, we would have to figure out which things we're going to support and which things we need. The first thing, like Jesse said, is the highguards election. That's going to be the first. Basically three transactions, like the initial highgards is one transaction, adding of new people is another transaction and removing of people is another transaction that you could do. And that's also something that we need to support in the UI, I feel.

David

Indeed it is the idea that we'll build out different templates that anyone can use in order to specify certain parameters for certain actions. One that I have in mind that I think would be a cool use cases for the validation committee where with the ATP you can either blacklist or whitelist certain requesters, for example. And indeed we also want to create a way in order to create any kind of ATP. But we're still thinking about that. And like Lawrence mentioned it's a bit. It can be dangerous if you allow anybody to put any kind of action into an ATP and then let people vote on it without people understanding what all of the actions taken might potentially do.

Laurens

So validation is a very valid point. So maybe that's like a good bridge to validation that Rochelle wanted to discuss. But let me quickly expand a little bit on the validation part, how I see it. It's like there's indeed a couple of things that we could involve the DAO in the validation. I've been thinking about this for already quite some time, but it wasn't technically not possible yet. But this is a good step towards that. So one of the foundations that we have, of course, is like the whitelist and a blacklist for requesters, but that's not an actual transaction yet because that's basically just a UI thing, right? So for now, that's just something that's basically a change in GitHub. So that's still a little bit hard for me to see how we can do that in an arbitrary transaction proposal. But we also need to work on validating the actual tasks that people are doing, like the work that's being done. Right, so there should be a way for requesters as well, maybe to reject work that the workers have done, but then we should decide what happens with those funds. Suppose a worker does a task, his funds are three days impending, but in those three days the requester still has the time, for example, to reject his work. But then those funds are there, basically. Looked, we could either send them back to the requester, but yeah, then the requester can also scam basically because he can have the workers do all the work and then ask all the money back. Another possibility is to have all those funds go back into the batch so that the same desk is being posted again so that a different worker can do it. Then at least we know that the requester has to spend the funds on that task, but it would still scan multiple repetitions of the same task by keeping on rejecting it. Right, so maybe there's another way is that if the worker maybe disputes the claim, like the worker might say, yeah, but I did good work, then it might go to the DAO and the DAO could decide either the funds go back to the requester or the funds go to the worker. So that also might be something that we could invoke the DAO with.

Rochelle

Right, sure. Hacking my Google. Workspace. Lawrence.

DJSTRIKANOVA

There needs to be a way to aggregate this all because I imagine it can get kind of tedious if you have to dispute a single task, for example. Yeah, like, say ten or 50 tasks, and then it's like all one process.

Laurens

Yeah. So maybe we can batch them together. So if all those tests are from the same bench, for example right. And he did dispute them, then he disputes them all at the same time. Maybe we can either approve them all or reject them all. And then you don't have to go through all of his work. You couldn't go through like, a subjection of his work.

David

Right.

Laurens

So maybe 10% if it's a lot. And based on that, we can make a decision to vote yes or no, basically. But yeah, that's all things that we still have to build.

DJSTRIKANOVA

Of course, I think we would need a reputation system on top of this. So that way there's like, for example, otherwise there's kind of a little point to it because I'm buying a supply chain, for example, you can just give it a new wallet in like 5 seconds. Right. Like an address based.

Laurens

Of course, what you could already do is go back history and see what the request has been doing, how many tasks he has approved. So maybe like, the approval rate of requesters right, could be already good reputation. How many tasks did he post and how many of those tasks did he actually approve or reject? And if you then see that either, it's like a new requester, you should be careful because you don't know him yet. Right. So it should be a little bit of a proof of requester. And then if you see that he rejects 75% of all the work, then you have to really think hard as a worker if you want to attempt to do his work. Yes or no, right?

DJSTRIKANOVA

Yeah, I think that would be great. That probably should be already feasible, just tracking everything that happened and labeling it to an address.

Laurens

Yeah, exactly. We don't even need too much of that in a smart contract. Right. We could just show it in the UI. We can just retrieve the information from the blockchain and see all the validations that requested this, and then we can basically calculate his approval rating.

DJSTRIKANOVA

And then same for the workers, we can see because with the workers, we can just check if they're like because the workers who have accounts where they have thousands of tasks have more incentive to not cheat on anything because they don't want to ruin the reputation of their account. And then ideally, it also can add a useful qualification, like for requesters, make sure that whoever does this task has the least done certain amount of tasks, not just some like, generated account.

Laurens

Exactly. The only thing is that for that qualification, we would think about how we assign it. Right. Because this information is not like something that we can easily retrieve in a smart contract, but maybe we can make something where maybe we have this validation button. Maybe we can use that to at some point update the qualification of workers with their correct approval rating score and then requests can use that score to basically say, I only want users that at least already did 100 tasks and have approved rating of over 90% or something. Right, that would be great to have.

David

Yeah, I can imagine a cron job that would go through. Then there are recent submissions and or other data points that we can use in order to assign the score.

DJSTRIKANOVA

Exactly.

Rochelle

There's lower ends, but yes, you guys are.

Bree

Okay.

Rochelle

My volume is working good. You guys are touching on a lot of points. And I swear, Lawrence has been hacking into my computer because since about June, when we were getting quite a bit of pace with our decentralized force, I realized working on stuff with the Devs, with what we need on. The UI and things like that. And with posting tasks, what are the things that we had for validation on the old centralized platform that had a back end? What is possible for the decentralized platform with only the blockchain as a back end? And I touched about this a little bit in the thread I had about my high guard election proposal, but I've been working out on a big validation system to employ on effect force and have it in a very decentralized way so that I don't have to manually do it all myself. And also that involves the DAO in a major way. As far as validation, you're right now when a requester post asks, they just all get paid. There's no way for a requester to intercept that three day period to pay a worker if they scan. The only thing a requester could do right now would be to create a custom qualification and on their campaigns add it as an exclusionary qualification and then assign those bad actors that completed work before for them that qualification so they can't attempt and do their tasks again for them. But that doesn't really solve the problem when you load tasks and then they just all get scammed. You can't have to load the task again, but you already have that money now. On other work platforms, requesters have up to 30 days to approve and pay out your work, otherwise it gets automatically paid out. 30 days is way too long. And other platforms, you can set a different increments. Five minutes an hour, automatic payout within an hour, five minutes a week to three to five days to two weeks, whatever, any custom amount you want, up to 30 days. Right now we have the three days. We're going to be building out a way for a requester to be able to flag tasks that were completed on their batch. So say they get one worker that just scammed every task in their batch they can go through on their own, validate on their own, and then flag every single task that the worker did. It's really easy to do that right now, looking through, like, submitted tasks on a batch page and what. My idea for that is anything that is flagged during after that three day period, anything that is flagged instead of going to the worker automatically or going back to the requester automatically, it gets set aside in a separate EOS account to be held in escrow. And then there's a way for the requester. We'll have to add it into the UI, a way for the requester to request a review by the DAO validation team and it'll be like a form they fill out and they describe what happened. The worker ID submit to us the CSV of their results and then the validation team with the DAO that I'm going to help create, we're going to be the intermediary between the worker and the requester. And there'll be some things that I have designed out that we look at and then there'll be a way for us to either assign this stuff back to the worker or assign it back to the requester. This is an important part, I think, for our platform that other platforms don't have. As a worker on other platforms, you really have no say with bad acting requesters and whatnot. A lot of times as a worker, you get scammed out of work, the requesters deal stuff and then a lot of times you can scam like crazy on a requester's batches and then they end up leaving the platform and then it just hurts other workers because then there's no more work for anyone. So this is something I think that we can handle in a decentralized way. I think the DAO should be in charge of it. So that's one thing. As far as task quality, another thing is that right now it's up to the requester to validate their own work. One of my plans, it's going to be a future proposal, is for the DAO to have its own DAP on effect force where requesters can take their data and their results and they can upload it to our dad and pay for us to validate their work for them. The DAO make some money on it. And the tasks that the DAO puts up for the DAO Validators to complete. You know that these are DAO members and only these DAO members have this qualification on effect for us. And they can do these tasks when they want, and it's another way for them to earn EFX and it would be a perk as being a DAO member. So with the DAO open for anyone to join, once we get going with a bunch of requesters and tasks on the platform, we're going to be attracting a lot of workers. And one thing that should set our platform apart from other platforms is giving workers the ability to have a say in the direction of how the platform goes, which is what we're going to do, especially with these ATPs. And everything that we're doing as a DAO, but also be able to take an active part in validating requesters, judging them validating other workers work and whatnot it gives workers a little bit more agency than what they receive on other platforms. So this is Part Of A gap I'm creating and building out that I'll Be working with The Platform Committee on My Teammates at Effect to help build out, because I Think It's Something That's Needed, and I Think the DAO could use an extra revenue stream and have a service, and I Think It's a needed service for Effect Force. Which leads me to part of what? Sorry, it's so hard for me to talk for long periods of time without coughing. Part of what I'm accomplishing with these AI image labeling tasks on our platform is I'm doing a pilot program for DAO members to be validators, so I'm doing a pilot experiment to have us validate these tasks. So I'm doing an open call for any DAO member. You have to be a DAO member who wants to validate other workers submissions with these image labeling tasks to get a hold of me. Tell me the worker ID of your effect force account and if you have more than one, if you have like an EOS and a BSc account and you have both and you don't know which one you're going to use, just tell me both worker ID and I'll assign you guys the qualifications so you can be the validators for these tasks. But I have a whole proposal I'm working on a whole plan I'm laying out. I'll be showing it to my team, but then also showing it to you guys as part of the platform committee. There's a lot I want to do with it. And yeah, I'm really excited because with these ATPs, the DAO is going to need a way to be able to directly publish campaigns for these validation tasks or directly publish campaigns to validate tasks from our clients when we have our validation app and stuff. And it's going to be needed everything from posting campaigns to batches to signing qualifications to flagging workers, blacklisting workers from the platform, blacklisting requesters from the platform. So it's a lot of things, and I think we need these items and, yeah, it's going to just set our platform apart from the others and this is the best way to do it.

Laurens

Nice. I love it. Yeah, it's great to see that you put real thought into this and been working on this, and I can't wait to discuss this more and see the implementation of this. I'm already thinking about a couple of technical things that we have to do to make this work. Of course, but there's work with creating the validators and selecting them and doing all that stuff. But there's also, like, a bit of technical work, of course, that needs to be done. Some changes to the smart contracts to actually be able to even hold these funds and be able to reject it and then either send it to the requester or the worker. That part still needs to be built. But it's great that you already thought about basically the next step. Once we have that part, then how are we going to validate and who is going to validate and how does that process look like? So nice. Really cool.

Rochelle

Exactly.

Laurens

Really help. And I strongly agree with your point that it's not just the workers that need to be validated and checked all sorts of requesters. Right. We shouldn't be a platform where the requests have all the power, like some of the other platforms, where they basically post all tasks and then workers do all the tasks and they reject them all, or they reject them for very silly reasons. The rejection should only be possible if a worker really spammed the task, right? It should be clear like this is not a fair submission and then it should be able to reject. But yeah, it's great that power is not in the hands of the requester. But you can ask like an independent committee. The DAO federators to take a look at that.

Rochelle

Exactly. It's not just say a requester. Flags all work from one worker and they just don't do a great job. But they didn't out and out scam. Sorry. Requester. Mr. Requester. They didn't provide results to your liking. However, they didn't scam. So we've decided to award the task back to the worker. And our suggestions are for you to add a custom qualification to your batches and assign. This worker a block qualification so they just can't do your work again. I mean, that kind of stuff. Yeah, exactly. On MTurk speaking of like rating requesters and whatnot as a turkey on MTurk MTurk is solely only for the requester because that's their client. That's who they get the money from. That's who they care about. So as workers, we had to band together and make our own tools to try to protect ourselves and other workers against unsavory requesters. And there was something started called turcopton. Yeah. And basically we called it to and everyone would send the reviews about requesters they've done tasks for, and they would rate how quickly they paid out, how good the rate of pay for the task was, all kinds of things. And they would also report if they received any rejections from the requester. And these were tools that us as workers had to create ourselves, had to spend the time, resources and everything to create for ourselves and use for ourselves for our own protection so we wouldn't get scammed out of things and that shouldn't have to be on the workers to band together to do on their own with us. This is something that we should have already built in. So that's part of what you guys were talking about. We got to have a rating system. And whatnot that I have in my plan something similar to what you guys were talking about. Something similar to what me and other workers on like MTurk and Navo and on Space God. I spent 15 minutes listing all the other platforms. Each platform us workers created something a little different cater to the platform itself that we're working on to help ourselves. And this should already be built in from the beginning and handled in a decentralized way. And so that's my plans for us as well.

David

Yeah, I'm also really excited to hear how you're envisioning this and how you're using the community to better the community for the community. I think there's a lot of potential here.

DJSTRIKANOVA

Yeah, I'm thinking so validation programmatically isn't that hard because you can just pull all the tasks from a campaign. Right? And so I wonder if you could just I feel like you could have as many levels as you want. Like for example, you can validate your validation on top of that just by pulling the task results and then showing it all to the new validator. And so it seems like from this conversation there's two aspects. There's the validation, which not necessarily has to be DAO member, but it can be to encourage the DAO participation. And then there's the dispute resolution which has to be the DAO because we have to decide and settle who is right and who's wrong. In certain cases, it seems the dispute resolution is going to be the more trickier part of getting it just right.

Rochelle

Yeah, that part is going to be a whole other other thing in itself. But as far as like validating tasks, of course DAO validators will validate the tasks that come out of the DAO validator DAP that we launch. Validation tasks will also come from anything effect Force uses effect Network uses all the Effect Network staps like the image label or DAP the Effect Social. Now as far as like say you DJ with your tasks that you put up a request like you, you would use the DAP that the DAO has for validation to give us all the info. Because say for requesters who have an iframe embedded within their task and the task is done like off site, we can't just pull from the blockchain what was in those requesters batches and validate it because it was done off site.

DJSTRIKANOVA

Like their system, if they upload it exactly. That actually introduces the question how are you going to handle disputes if we don't know what's going on data wise?

Rochelle

They have to upload it to us.

DJSTRIKANOVA

And if they don't, but doesn't that put us at a disadvantage? Because they could easily do whatever they could pretend that they got bad data to get it for free.

Rochelle

They could.

Laurens

That's another part because the worker posts a hash of his work.

DJSTRIKANOVA

We have a hash. Yeah, well, they still like from the class implementation they go to an Iframe which is an external source and they do whatever there. Right. We can't really hash that. We can hash the payments that occurred but we can't necessarily hash that whatever they did on their external site was done correctly.

Laurens

True.

David

Well, assuming that when you do a task, you'll submit some data that will be returned to the requester and there is a bit of a paper trail with that that we can then trace.

DJSTRIKANOVA

Back to well, yeah, if we do it on the blockchain. But for example, if people want.

Laurens

To.

DJSTRIKANOVA

Keep their results private, they can have an Iframe basically embedded in the campaign, like the task context and then on this external site they do a task and then once it's done, they get like a completion code and they send this completion code to them. But I feel like the only way we can really truly deal with disputes is if everything is already on public record on the blockchain because then we know for sure, like this was the requester's task and this was what the worker submitted. That's clear cut. But I mean, if people upload something then it's a lot trickier to know if they're lying or not.

Rochelle

Then I think we take into account, like, the age of the worker account their score with quality on other things that we validated, things like that. And then we kind of just make a case for the worker and the case for the requester. And then as validators on the validator team, we just kind of take a consensus on what we feel as a group.

Laurens

Yeah, but in the end because it is indeed possible to circumvent using the results to the blockchain right, because the requester determines what the workers in the end are posting to the blockchain. And for example, you can still have private data as long as those issue post a hash or something of those questions or the response to those questions on the blockchain. But if you don't do that, for example, with an external survey and it's hard to immediately retrieve the results of those surface. So you just have the completion code and you just basically submit a blank result. Yes. That's a little bit hard to check in a decentralized way. Right, you cannot check centralized results in a decentralized way. So yes, I would either say we could try to say like how trustworthy is the requester and how trustworthy is the worker but if there's doubt and we don't know, then I would say we rule in favor of the worker because it was up to the requesters. The request decided to have it centralized and not on the blockchain.

Rochelle

Then it would default back to the worker and we just tell the requester, look, to avoid this in the future, assign them a qualification and block that qualification from your campaigns when we do a report. But I think in the end it should always go on the side of the requester. I mean the worker cases like that but it would have to have a DAO member consensus too in the open and say it is a crazy high level case like that. That's a crazy outlier. We can put it to the DAO vote.

DJSTRIKANOVA

Well, if we put it to the DAO vote, wouldn't it expire? I guess this is just the whole process we're going to have to figure out for 2023, I guess.

Rochelle

Yes, experiment because yeah, I think after.

DJSTRIKANOVA

The DAO 2.0 updates, I think validation will definitely be a good way to go. Like just improving like metrics. People can use qualifications being able to, I guess, use the user interface to ban workers from doing your campaign or otherwise request that the workers have a certain amount of tasks done or stuff. We can already measure things like that. And then after that we probably will have to go for dispute resolution in a way that is scalable so we don't want too many disputes happening and us not being able to deal with them. So I think, yeah, those are two big chunks of things. We need to start getting more competitive with MTRK and stuff though, I think. I wrote another thread that I think we should try to make an appeal for the AI type market because the data sets they use is usually in the research and so the research data is all publicly available. So there's not really an issue if it's available on the blockchain because it's all publicly available anyway. Right. So there's less friction from having to develop your own server to do basically it seems like it's too much effort to create your own special thing and then only use Effect or EFX as like your payment processor when you can just use MTurk, right. It's like you have a lot of extra steps. So I feel like we are better served going over going to people who want like decentralized AI, they want to fund the creation of models on the blockchain and make it all publicly available. And validation for that will be much easier because we can just look at the results ourselves and make a quick check like that.

Rochelle

There's something fun for you, DJ, it might change your mind on that. Have you ever posted tests on MTurk?

DJSTRIKANOVA

I have not.

Rochelle

Okay. You need to create a requester account and go through the steps to post attached. I guarantee you the everyday person or even someone that's a little more technically inclined like me, it's impossible. It's next to impossible for the average person to post tasks on Turkey brie, no disrespect. You would not be able to do it at all.

Bree

I can barely do it for our.

Rochelle

Own yeah, it's not easy at all and it should be easier than how they have it. It's ridiculous that they're so requested focused and the steps you have to go through and things to post on there, it's nuts. So that's already one win that we have with our system, even with a not intuitive UI.

Bree

I agree with DJ though, about really focusing on the AI and the research.

Rochelle

Yeah.

Bree

Hugely. Because even if you've got a bad worker in that batch, as a researcher, you're used to those outliers, you would just dismiss it. You're still paying for that task, right? Because that whole body of data tells you something so you wouldn't reject it, you would take it in and place it where it belongs. So there's less likely for rejection when you're looking at research bodies.

DJSTRIKANOVA

And with validation, I think I'll just have to demonstrate what I'm thinking, because I don't think it'll be too hard to program, because if they affect SDK, you can just pull all the results from any campaign. And so you can literally just type in the number of that campaign and ask, give me the results for this. And then also, like, can just pipe it back into another campaign and just add, like, a little extra question, like art that shows the results and says, was this done properly? Or something like that. I feel like I could just make a depth like that real quickly and then you could just do that for any campaign for validation because that's like going to be the other aspect. Because for example, you mentioned this with the hands, right? Rochelle image generation can really nail it or sometimes it just gets it wrong and they want to be able to figure out all this stuff is emerging technology. So they're going to be going to want validation of their results, like just masses of individual labeled results where the image is either considered correct or incorrect. That will be very useful to them. So validation won't necessarily even need to be tasks done because they would want to validate the data they already have.

Rochelle

Yeah, absolutely. That's kind of like what we did on that first round of the image labeling test. Like is there an anomaly in this or not? Because the AI had people in there or crazy mismatches of people, then they shouldn't have been, it should have just been only landscape. So I mean, that's an example right there.

David

Indeed. We've also been looking around at the hugging phase data sets and there's so much there actually that we can use and somehow apply the force in order to enrich those data sets. So yes, more and more we're really gearing ourselves up in order to start creating campaigns that are using those APIs and those data sets to try to create a proof of concept in order to really connect with that community more.

DJSTRIKANOVA

Yeah, speaking of the DAO funded ATPs for campaigns, the idea I had is one way we could do outreach is we could publish a data set. And I wonder if we can find one data set that just needs an extra layer of validation and it has like the licensing is all proper where we can just add another column and say, is this correct? Or something? Or something like that. Or is it incorrect or otherwise. Yeah. If we can make a campaign that creates a data set and then upload it to a hugging face, that will be, I think, one way to outreach.

David

Yeah, that's a good idea. And that's also a good way to kind of get our name out there into that community.

Rochelle

I think we're on the same page with that, but lawrence okay, say we have the ATPs done out for like, validation and posting campaigns and tax. Those wouldn't go on the regular two week proposal cycle, right?

Laurens

Yeah, I think so. But for validation, that might not be good. Right. So something that we still need to discuss. Right now we just have this two weeks cycle, right. Maybe we should have another cycle, maybe more shorter cycle. Maybe that's just for these kind of things. Yeah, because two weeks right.

Rochelle

It'd be nice to have a separate voting page with shorter cycles or things like this for effect, for Tao force or something.

Laurens

Yes. Also because, for example, that validation period is only three days. Right. It's definitely something that we need to think about and need to discuss and technically see how we're going to solve that. But right now it would just be a normal proposal which follows a two week cycle. But that might not be good for the validation proposals. Well, it's not really a proposal. Right. So maybe it should be completely separate. Maybe it should not be a proposal or it should be like an arbitrary transaction proposal, but in a different kind of structure, not in the two week cycle.

DJSTRIKANOVA

You mean like dispute resolution stuff?

Laurens

Yeah.

DJSTRIKANOVA

I feel like that could overwhelm the DAO. I don't know how many disputes there are, but if there's a lot you can kind of get and then, as you said, it's the two weeks. Yeah. I feel like we want something just separately designed for dispute resolution.

Rochelle

But in the meantime, I'm going to do a pilot program for validation without the ATPs. And of course, it'll just be kind of me manually assigning qualifications to those of you in the DAO who are interested in doing the early validation tasks. So the image labeling one, the Twitter task, these kind of things, you would just be validating a previously submitted response from a worker and then kind of working with me, giving me feedback on things and your thoughts. So that's going to be the events to kick off the year for the platform committee and the DAO. But yeah, lots of discussions on that. The last thing I put on the agenda, which I think is fitting to end the call with, is I'm. Drafting our two year anniversary effective blog post. And of Course, we're going to recap on all the things that we have accomplished as a DAO. Not Just the team for the DAO, but as a DAO ourselves throughout this Whole year. I'm going to write up a really nice piece, and so I'd like your input on things to add in for this blog post. We only got a few days for me to finish it because it Will be published on the first on our DAO anniversary date. So, of course, the big thing at the end is DJ's task notification system. That's the big one.

Bree

That's cute.

Rochelle

And all the doubt, the DAO 2.0 stuff, the High Guard stuff, all that, of course. But I need your guys'help.

DJSTRIKANOVA

What did you say? Rochelle?

Rochelle

But I'll need your guys'help to remember to list everything. Like we made the DAO Telegram channel, other tools, I think skip Book stuff. Yeah. The developer dashboard.

Bree

Tweaking to the dashboard.

Rochelle

Yes.

Bree

For me, a Big thing is technically the ATP. But I'm happy to see that the DAO is organizing itself into committee where we can focus more intently on getting jobs done, which Might Take some time to develop and get the Right people in place. Because it's not just the High Guard members for the committee. Like, we need other people. The High Guard is just one person of the marketing. Anyway, I'm happy to see that.

David

We also have the apps.

Rochelle

You'Re the launch of.

David

Yes. And the homemade ones.

DJSTRIKANOVA

The whole progression. The beginning of this year was the first hackathon while we still were at Testnet.

Rochelle

Yeah.

DJSTRIKANOVA

I'm pretty happy to see how develop the SDK is getting because even I'm able to build some stuff now. It's pretty straightforward, and I feel like it's pretty powerful already. And I feel once we get validation, like a process set up for and also, like, good qualified metrics for qualifications, so you can kind of like, say, for example, I want only DAO members to do my tasks, or I only want them to do people who have done like, hundred tasks already to do it, stuff like that. Like, I feel once we get that, we'll have a pretty solid, like, tool for micro tasking just and you Can, like, just build what you want. It's pretty neat for the fruit this year. I like the development since, like, the beginning of this year on that.

Laurens

Yeah, I agree. So I'm really excited that it's been shaping up in this way where people can actually build upon it. Right. And that's harder achieved than you think. Like, you cannot just build codes and put them get up and think that people will build upon it. Right. So you really need to make it in such a way that's understandable. It's well documented. You have tools, you help people. And we're not completely there yet, but we're heading in the right direction. I agree. There's still some big steps that we need to take in regards to validation, in regards to qualification. Because I also think that I have many ideas for the qualifications as well. I think if we can filter on qualification values that would also be really helpful, then we can have qualifications for many things. Like you can basically build profiles for workers. Like you could have a value for age or gender or location or language or worker score. And you can say in a campaign, I only want people that have the qualification with the value between like 20 or 30, for example, or a value with higher than 90%. If you can say these kind of things, then the qualification part will also be really, really cool. But yeah, I'm really excited for all the new things and that and also excited for what the DAO is shaping up to be, especially with arbitrary transaction proposals. Yeah, we're really turning into a decentralized network with a DAO effect force that people can build upon. And if we just improve on a couple of little things, it shouldn't even be that hard. Then this can really blow up. And that should be the defect of standard from micro tasking. That's completely decentralized.

Rochelle

If you look at the advancements and developments that the BV has made over the past year and then what the DAO has made over the past year, that's a lot. It's a lot. And you guys will see in my blog post for the Effect networks end of the year one and stuff, and then you'll see like, the DAO one. If we continue this same momentum next year, what do you think we're going to look like this time next year? It's very inspiring to think about.

DJSTRIKANOVA

Those hackathons going again because I think the first one, like, we kind of it was limited because it was testnet. But like, I feel like there are great participants there and even Abdulmad there like really great app there. And so like, I hope Jesse mentioned that this could be this will be part of the road map and he's going to advocate it some more. So like with Pomelo you can submit proposals as well for funding and grants like that. And for them, their goal is to bring usage to the EOS ecosystem. Because it seems like basically I'm not too keen on EOS history. But there was a Block One was the name. They didn't do enough. Basically the perception was after they launched.

Bree

Blockchain, there's been a real community revival for you.

DJSTRIKANOVA

Yeah. So it seems like we should participate in their community. And I think it could be a win win for everyone because we'll be building because any dev built on Effect network is built on EOS. The relayer for buying that smart chain is only a relayer. Everything needs to happen on EOS. If they want applications, we can help run the hackathons and we can maybe invite them to be the judges for the main prizes and then also have dial prices. I feel like we could set up something really cool between us all, and they could get a lot of excitement for everything. And when I look at the grants, the Effect network is completely different from all the other grants. Generally, there's no other micro tasking service there. So definitely, I think, a lot of opportunity. And I think my ideal is, like, if we can get grants for this, we just have every quarter a hackathon, a new set of DAPs, and we're always voting on the best ones, funding others who want like, an infusement because they want to do like a start up for their tasks and stuff like that. And then people building AI, DAPs, data set or data set generators. I feel like that should be our goal for 2023, and I think it's definitely feasible if we can kind of get a good talk with Pamela and then the Effect not affect the EOS Network Foundation.

David

Yes, indeed. Thanks for sending through the grant program for the EOS Network Foundation. I'm looking into it right now, and we're planning on submitting an application for a grant with them. And yeah, it's really exciting. I think indeed, because of the platform that we are, we can potentially really add a lot of users to EOS, so I think it is in their benefit as well to help us grow in order to attract new users. And I think we have a pretty compelling use case here.

Bree

Definitely.

Rochelle

I'm going to put a post in the DAO forum about the two year anniversary blog post. So if you guys would all help me out and add your ideas in there. I just want to make sure I don't miss anything because we've done so much. That would be awesome.

David

We'll do. But I also have to leave. Everybody. It was really nice talking to you. I hope you have a great change of the year. I hope you have a great holiday.

DJSTRIKANOVA

Still.

David

Good luck with the cold.

Laurens

Yes, same here. I also have to go, but thanks for this call. And yes, it's going to be an amazing year. I think if we take a couple of additional steps, then this can really be big.

DJSTRIKANOVA

See you guys.

Bree

Happy New Year.

DJSTRIKANOVA

Yeah, happy New Year as well.

Laurens

Happy New Year, everyone.

Rochelle

Yes. Happy New Year. We'll all be celebrating together on the first anniversary of the DAO.

Laurens

Yes, we will.

Rochelle

Yay. Great.