Token Engineering Commons Lab!, 22 Jan 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: W7 TEC Lab!: DAO Hatch Templating with Python Panel

Description

This week we just take the hour to walk through the progress that the tec lab has made on templating the dao hatch parameters! Next steps are to modularize the code and provide the TEC with instructions on how to modify their own parameter choices.

🙏 Thank you for watching! Hit 👍 and subscribe 🚩 to support this work

🌱Join the Community🌱
on Discord https://discord.gg/DDr5kYU
or say hello on Telegram http://t.me/CommonsStack
Join the conversation https://forum.tecommons.org/
Follow us on Twitter: http://twitter.com/CommonsStack
Learn more http://tecommons.org/

A

Hey good morning or afternoon or evening, whatever it is, for you all of.

B

The above good day, excellent, I'm so happy you're doing this. It dovetails nicely, I think, with the data analysis class santi's doing for us.

A

Oh amazing yeah: I would love to tune into that.

B

um Yeah, it's mondays we'd love to have you.

A

Excellent yeah I'll start dropping by when um yeah when is the is that, like one o'clock or something on monday,.

B

um Let me look at my calendar. Let me see, let's look, data science, it's at 11 a.m, to 12, 30 p.m, pacific um and I can go 30 p.m. I can google the translation of that. Let's go 11 a.m. Seattle.

B

B

7 p.m: utc.

B

Yeah, okay, yeah, that's kind of a goofy time for me, but I uh I jumped in just a little bit late into that one. Where are you at ytg.

A

Oh I'm nearby are you in seattle or.

B

A

I'm on vancouver island.

B

Oh nice, nice! Well, they let us in there they need a passport. Don't I.

A

At this point, uh well, you need some sort of passport, yeah uh yeah or I don't know- maybe just a driver's license that might work. I don't know how the border is doing these days.

B

A

um But I'm sure you could sneak in somehow. If you have the will there's a way.

A

I was hoping my discord. Handle would be enough.

B

There you go, yeah they'd, be like.

A

Bright id here's my bright id all right.

B

There you go, it's me.

B

Excellent well, it looks like we're getting people fabs.

A

Awesome, oh hey everyone uh danilo nice to see you.

B

ah Hey here I am.

A

So I usually just give a few minutes for people to uh stumble in. Usually people are a few minutes late, so um just I'd usually just start with casual chat at the beginning, it's pretty good and um I'm usually doing some last minute prep. So that's what you can see on my screen here and I'll always guide everyone. So the best way to get started with this stuff is just jump in the tec labs channel here in te commons and check the pinned, messages and you'll see a link to the notion workspace. There.

A

And from there, what I'd like everyone to do is open up the attendance sheet and I'll create a new column here and I'll just make sure. Oh.

B

Yeah, I'm also.

A

Recording and also streaming by the way, so if anyone has any concerns about that, just let me know you can even direct message me. um Let's see.

B

Insert left lab.

A

Seven, okay, so if uh what we usually do, is uh you can grab any sort of emoji that you like, and just if you're new to the labs, just go ahead and add a new row here and the data is optional, uh fill in any sort of columns that you would like to and you can pop your favorite emoji or how you're feeling today in the lab 7 column,.

B

I see somebody sassy's got the latest ubuntu 2004. There.

B

Oh, you gotta upgrade.

B

um Yeah, I'm I need to get a new machine, I'm on an old hp that I enjoy a bit.

A

Yeah there's 18.04 was this a slightly unstable version because they were transitioning? I think that was the first lts that moved over to gnome, so it was a little bit. I think the 20.04 has a lot of they tightened up a lot of things. It's a smoother experience.

B

Oh, is it well.

A

That's encouraging.

B

A

B

Want to, I don't want to risk breaking my machine so.

A

B

Yeah, hopefully you can get a get a new.

A

Machine maybe put in a proposal because you do so much for the communities, maybe we'll you could get a machine funded.

B

Oh, I didn't even think of that I mean I'd. Take a you know like a redone machine. I don't need a new one. I'm.

A

I wonder if we could have like a dow asset inventory like a database of all our computers or something, and you know if we have things that we're not using, we might be able to send them around.

B

Oh, that would be so cool. I know I was thinking about that and then I'm here in seattle there's got to be used machines somewhere where I was in alaska. All the homeschool programs had like new computers and, um like every student got a computer and every parent got a computer, so they always. You could always get good use. Computers.

A

B

A

B

Cool it was, it was really cool, but no mas, no mas, we gotta figure out how to get emoji in here. I'm just gonna say ponies because I'm on my computer, I don't know how to do it.

A

Oh, I always just go. I open up a new tab and I go get and then there's all these emojis and you can just copy one.

B

B

I didn't even know about that. That's gonna make my whole experience a lot more fun. Oh look at that.

A

Okay, so, while everyone's stumbling in I'll, just probably a few more people have joined, so I'm just going to say um to get started, you want to open up the tech labs channel in the te commons section here in discord and check the pinned messages and you'll see a link to the notion, workspace, and that should be everything from there. You should be able to find everything that we're going through uh open up the attendance sheet and pop in an emoji.

A

You can use getimoji.com to find one and at the top here we have the github repository that we're going to be working in and in the calendar, you'll see lab 7, that's the workspace, so I just put in the agenda here. It's a rough agenda, hopefully by the the goal. By the end of this, is we have this directory of a clean uh tec, hatch, um dashboard python, module that we can then import into jupyter notebooks and get some very cool widgets.

A

Where we can tweak the parameters, we can get a visualizations of what those parameters mean, and this is like a prototype for a thing that we can invite everyone. The whole community in to repeat the process that we're gonna try to create today, and so everyone can create their own hatch version, uh their their own parameterized hatch, with the graphs that come with that, and then the community as a whole can vote on those outcomes.

A

So pretty ambitious, we'll see uh how far we can get through this. Actually, this first one I was, I have to give a shout out to uh danilo because I'm always working in jupiter notebooks, but seeing uh the lab the interactive git coin modeling on wednesday. I realized oh yeah, it's good to go to like modulize modularize your code and move it into like visual studio and create some clean functions and packages.

A

So that's! uh Luckily, the what we've been working on is in a good state and it's ready for that. So uh if we open up the github labs repository.

B

So real quickly, where am I finding that link again, I was got lost in the emojis.

A

Yeah yeah, so you probably have the notion workspace open, yes, good good, so the github uh is at the top. It's the tec.

A

B

Right where it says where it says, master repository yeah got it. Thank you so much. Oh.

A

Yeah no problem, if you have questions probably three other people have the same questions so um good. So everyone if you want to follow along. Sometimes these labs start moving kind of fast, so they become hard to follow along. You know we'll have different yep. Is that tonga there or is someone a phil, hey phil? I think your mic is a little a little.

B

Robotics so we're on lab seven right.

A

Yes, lab seven, so I'm gonna open up my terminal.

A

And so you can get clone the oops, so I always run this function essentials. It helps, helps my it actually speeds up my keyboard a little bit um and then yeah you can clone the repo.

A

If you want to follow along. I already have it here, so I'm going to go in there, tec labs and I can let's see so I activate my virtual environment tec- doesn't exist. I think it's uppercase. Okay, now I use uh tmux to split my terminal like this and I want to do someday. I want to do like in a session on just a sort of terminal workflow, but not today, so I'm going to go ahead and open up jupiter lab.

B

Is the is the tec contained in the repo you cloned, or is that your own virtual environment that you already have set up? It's.

A

My own virtual environment, so in python it would be like uh you use the venv package, virtual environment and I think it's on the. If we check out the repo here um it's here. So if you want to create a virtual environment, you we have all the instructions here and then I use an alternative shell in my terminal here I use the fish shell and it makes, and then I use virtual fish. So it's really easy commands to create virtual environments.

B

A

Yeah, that's all here in the readme, so everyone you guys, can check that out. Definitely I recommend reading through this before or after the labs or during and yeah. So we've got our jupiter lab going um now. I've navigated into lab seven, also, a little tip that uh I do here is I have this scratch directory and this is ignored by our get ignore file. So you can. This can be a nice like personal workspace. You can put anything in here and it's not going to be tracked by git.

A

So it's like your local experimental workspace and what I do locally. I keep a lot of repositories that I use like hollow views and holoviews is a nice repository because it has all these examples and in fact all of its documentation is uh built out in jupiter notebooks. So I can go ahead and open all these as references and just run them to remember how these libraries work, and so I'm not sure what this one's about. I'm not going to. Oh, it's explaining how all of these layouts are composed.

A

So this is pretty cool, but let's not go down that rabbit hole. Just the concept that anyone you can put anything in that scratch um directory and it's not gonna- get uh pushed upstream to the github, so I'm gonna go into lab 7 and I was like trying so many things here. So I'm going to find uh the first thing we'll do is maybe just clean this up a little bit. So I think the good copy is not the one. That's called good copy, it's the one. That's called culture tribute one!

A

So here we have something missing. That might not be the good copy.

A

I'm just gonna go through these one by one um this I this seems like a probably a lot of code and kind of overwhelming, and honestly it is so that's why the first thing we're going to do today is move this code out of the jupyter notebook environment and into a actual modularized code base, and so what what we have here is we're loading in some data sets. These are this is the impact hour data set um in two different forms, so this is by by user, and then this is like aggregate by period.

A

uh This is uh the c stack token holders, because in order to participate in the tec hatch, um you have to hold c stack tokens.

A

So it's an important um I'm pretty sure one of these is a good copy that should just everything should run and if not, then we're gonna have to do some tinkering to see why uh something's not working here. Okay, I think it was this cultural tribute one. What was this error? We got dandelion voting uh missing, one required, positional argument. I wonder if I can increase my font, can you guys see the text here.

B

Yeah a little bit, that's.

A

I think it's lagging a little bit because of all these.

B

Visualizations.

B

A

Now it's just going to keep increasing in size because I hit the button too many times.

B

A

That should do it. That's probably a little bit excessive.

A

But uh let's see if we can so missing an initialization parameter. um What dandelion voting.

B

A

Tokens so I think I'm getting total tokens from our impact hours formula.

B

A

B

That a piece of of code that is missing so.

A

I think it goes here oops. I have to copy that.

B

And underscore, and it will be the bell tokens.

A

B

A

Initializer here I think.

B

No, no! No! No. I think that it's in the initial function in the in the first function in here here in the third, oh okay,.

A

Weird, okay and then we're also gonna need that, because that's missing in the initializer or the constructor I mean so I have to uh total. No, these are impact hours. Where did I get total tokens from total c stack? Tokens.

A

Let's see here, this is a lot to take in all at once. I've been hacking on this throughout the week and um I should have probably done a refresher myself before doing this lab, but let's see how can we get total tokens? So it's this. What we see here is the hatch so we're talking about total c stack, tokens, hatch, oracle ratio, maximum raise minimum, raise target, raise hatch period days, hatch exchange rate and the hatch tribute, and uh somehow from that I believe.

A

I was getting, I might have a further uh iteration of this. This is called good copy, total total impact hours. I definitely had a version that was outputting the.

A

Let me try recloning the repository, because I'm pretty sure I pushed uh all my changes or let's see the latest commit.

A

uh So you can see my latest commit. It's been a lot of hack and slash, but this is looking really good, so I remember pushing a good copy, so the changed files, even though they're jupiter notebooks, we can see that they were the good copy and the cultural tribute one. So, let's investigate those two files in particular, and I'm just going to re-clone the repository, because it's possible that I had like been working in some different workspace.

A

So I'm just going to create something.

B

To see if it's ahead or not oh.

A

Thank you. Yes, yes, okay, so there's a whole bunch. So, let's just go for now. I could get stash to um let's also check that our latest commit here. Yep I have the same latest commit so I've made changes since the latest commit on my local system, and so I could uh get stash and I think I'll do that, so I'm going to actually close this notebook and get stash now. This brings me to so I have it in my terminal.

A

You can see that I'm now, um if I do get status, it says you're all up to date with origin main, which is the upstream here. So let's try this again jupiter lab and remember from our commit what were those two files? They were like a cultural tribute, one and good copy, so.

A

So cultural tribute one and good copy.

A

And I'm pretty sure this is the one I was working on, so let's go through this um a little bit slower. So this is loading the data, so we have our impact hours um panel. That comes first, so this is modeling okay. So it's loading. A couple data sets historic and optimistic. uh This is actually the real impact hour data.

A

We have this parameter, which is predicted hours. Okay, so let's just walk through this first of all. So what I've done here is we have our impact hours accumulation curve, so these are these are this? Is the accumulation of impact hours, the aggregate impact hours that the tec has been accumulating through um the praise system? So whenever someone is praised, uh it gets logged into a database and gets translated into impact hours.

A

So as of january 15th, there was uh just over three thousand thirty five hundred impact hours accumulated and if we continue that trajectory, so it was this kind of somewhat exponential trajectory and griff actually did this modeling. So what we have in blue here is actually the predicted continuation of the impact hours based on the historic trajectory. Now we had this idea in the tec that optimally. We don't want our impact hours to grow exponentially. It would actually be better if they grow linearly. I mean the less impact hours.

A

There are the higher the hourly wage that people are going to get paid out through the cultural tribute. So we made this sort of optimistic uh projection as well, and this says that by the time of the hatch, if the hatch is in april, then we'll have 6363 impact hours total and then what I did is I made this orange line, which is the predicted.

A

So it says if we kind of strive for this optimistic case, but reality says we're more likely to go for this historic case, then maybe we'll land somewhere in the middle, and so that's what I give the user this option to play with, and this is the predicted impact hours. So it's simply a slider.

A

It's like a factor, multiple that um says how the impact hours are gonna grow over time. Are we gonna go uh just completely? You know the same as we've been going on the same trajectory. Are we gonna strive to kind of you know, slim down our impact hours and and have less total or maybe we'll even almost stagnate, our impact hours and barely grow them, or maybe they'll grow exponentially out of out of control, and then so. This is nice.

A

You can see here that this the total impact hours, and so this number is going to really come in handy as we go later, because this can be used to calculate based on our total raise uh and our cultural tribute that we decide upon actually how many, how much die per hour everybody's getting paid, and since we already have the impact hour data, we can translate that to um total die that people are going to get paid out from this cultural tribute.

A

So I think that can be a really good kind of morale boost for people who have been putting in a lot of work like sem he's number one on the leaderboard um a lot of hours seriously coming from sem, but everyone everyone's been really putting in a lot of work, so it'll be really rewarding to get to translate this into. Actually, what does that mean?

A

What is what is the payout so now that we have this predicted impact hours, we're going to move into the impact hours formula, and this is just some beautiful math that no one saw coming. It's really elegant, sem and griff figured this out together, and so it takes in this number. Okay, what are the total impact hours that have been accumulated?

A

What's our minimum raise that if we don't raise that amount of funding we're actually going to abort and uh and refund the funds? Okay and then, as a person going through this dashboard as a hatcher? What is your expected? um Oh yeah, what do you think for every hour that we contribute as an impact hour? How much can we raise um so the default for that is 25., so for every hour that any of us put in, we expect to increase the total raise by 25, die or x, die or rap dex die.

A

um What's the maximum impact hour rate. So what is if we raised like 100 million x, die, and that might mean we're paying out our hatcher's like a thousand x die per hour, and we just all. I want to acknowledge that, well, that's probably not necessary. We don't have to get paid out.

A

A thousand x die per hour, let's cap it so the default on it is a hundred and then we have a target raise and a maximum raise and the target raise is what we're hoping for the maximum raise is actually a cap, that's built into the smart contracts. So if we hit the maximum raise, then the raise is essentially over.

A

So we can take a look at what this means. So we have our total impact hours, which is gets populated from our projections in the previous example, and we do that, but with this simple initialization, so in our init function we take in total impact hours and we just access our impact hours, data which I had just called I in this case, um and this is a panel object, and so it's going to have this attribute called total impact hours.

A

So this this is using the param library- and this is really neat, because if this is a param- um and you can see that by like we, these classes inherit from param.parameterized and when we give them their attributes, we use this param module.

A

So we say, like total impact hours is a param number and what that does is when we access this attribute from our object. We're actually just returned an integer. This is just a plain old integer, but because we've inherited from the param object, there's some really intelligent things that happen behind the scenes. There's another attribute called dot param, and it's going to have it's basically like an object or a dictionary that has this total impact hours.

A

uh So this would be a different way to access this same attribute, but in this case we're not accessing an integer we're actually accessing a param number which comes with all these bonuses like it gives us these widgets. It basically knows how to visualize itself, and so what the param library does. Is it fools you? It maintains this plain integer, but whenever you try to access it, it has these access functions, getters and setters that mean you're not actually directly, manipulating the integer itself, you're interfacing with this piece, which is a parameterized number.

A

So it feels like you're just accessing this number, but you get all this free functionality like the visualizations anyways we're going to use this integer to populate our impact hours formula to get the total impact hours. um So let's go through this and then we just use our panel library to make this a dashboard.

A

So we have our total impact hours, that's seeded from our previous um and I think we're going to go off course. I like we might get to where I planned, which is we export this code out into a module, but I think this is totally valuable just going over this alone, and we saw at the very end there, there was something that was broken, so I think we have to do this anyways to uh fix our code. So this this is working out. Well, I think so.

A

We have all of our parameters here: minimum minimum raise expected, raise per impact hour, maximum impact hour rate etc, and we get this chart here. So this is from sem's formula and it says the more we raise the more we should get paid out per impact hour.

A

So, as you go up on the x axis, you go up on the y axis, but it it has this asymptotic limit, which is exactly our maximum impact, our rate. So if we increase our maximum impact hour rate, we can see that the y axis is growing and the plot looks the same, but the y axis is growing. So we could turn our maximum impact, our rate, and this is going to grow slowly. Okay, so it actually gets capped there, because we have our uh tar maximum raise.

A

I would I would guess is that our maximum raise.

A

Okay, so this this widget's kind of funny, because these things are like all intertwined, but um basically we can turn our maximum raise all the way up to uh 1.8 million and there's a way to actually increase that cap. I think it's like if we increase our expected raise per impact hour, what happens um so?

A

We have two uh like crosshairs here so and I was working on putting a legend in, I feel like I had it, but I made some it required some sacrifices of other uh simplicity and in blue here we have our actually our expected raise based on this expected raise per impact hour. So that's simply a multiplication of this number times our total impact hours, but then the red cross hair is we have our target.

A

So, regardless of what's the expectation based on our on this, these these kind of inputs, we also have the freedom to say well, what would be that? What should we strive for? uh What? What is the target that we're aiming for? And so, if we set our maximum impact hour rate to something I think 100 is pretty reasonable and we set a target raise to. um Let's see what is this 500 around 500 000 die, then we get this impact hour rate uh target impact.

A

Our rate is around 60 65 uh die per hour with a target raise of just over 500 000. So we can play with this and there's totally another uh chart here. That's off so maybe I can change this from. I can use this panel functionality, so I can go panel dot column.

A

This is a lot like bootstrap. The panel library is a lot like bootstrap if you're that's, where I'm familiar with working with like this row and column format. So I just added this panel.column.

A

So now I get these two charts in a column and the beautiful thing about the holov, holovez and panel stack is like everything's composable. So all of these visualizations are independent and we can just throw them together and and kind of loop them together really easily using this uh parameter inheritance system, and if you want to get a deep dive into that, just check out right here from the lab workspace check out the lab playlist and you can see lab 6.

A

We go into a deep dive on panel, and so here we have actually the expected and target funding pools, so we have the minimum. We have the expected and we have the target and something that's really interesting for me. That jumps out is like this.

A

The cultural tribute seems huge, so either might be a bug in the code, or this might be just an unexpected outcome of the parameter setup because I think usually we're thinking in our data and and we could do some further data science to really get to the bottom of this and even visualize it. But in our data that we were playing with to come up with these models, I think the cultural tribute was usually between like one to twenty percent of the of the rays.

A

But what we see here based on the conditions it's looking like it, you know- maybe these colors are inverted so that could require some. That would actually make a lot of sense, and but I had checked, I thought of that- and I checked it and I it seemed like nope. These colors are not inverted. This is correct, but they should essentially be the opposite. The cultural tribute should be rather small, like a maximum of 20 percent of the total funding, and the funding pool should be the majority.

A

But let's take a look at if we say what's our expected raise so as we increase our expected raise per impact hour, we see that there's more funds going towards the funding pool, especially in that expected section, and if we increase our target, raise our total target raise. Then this starts to grow. Now, as we increase our raise, we start getting the properties that we would hope for which is the cultural tribute being a smaller and smaller portion.

A

So this is this sort of incentive aligned aspect that we pin down to the cultural tribute where the it's, the more funds we raise, the more the hatcher's will get paid. So if, if the amount of funds raises anything, then the hourly rate of the hatcheries will also increase yet at the same time, while that constraint holds there's another constraint that says the more we raise the less the piece of the pie goes to the cultural tribute.

A

So this is a double incentive: it means for the hatchers the more we raise, the more we're going to get paid per hour, but also for the investors or the co-vestors we like to call them the more. They invest the larger the piece of the pie they're getting because the pie is growing which helps the hatchers get a larger payout. But the piece of the pie that goes to the hatcheries is actually shrinking, as the pie increases, so the co-investors get a larger piece of the pie as they invest more.

A

So it's this like doubly incentive aligned way of maximizing the amount of funds that we want to raise.

A

So I think this is going to be a really fun dashboard for hatchers to play around with, and it's very dynamic. uh You can get into uh weird kind of situations, um depending on how you play with this thing.

A

uh Yeah you can get into these like you can like basically break it, which makes sense if the maximum impact our rate is zero. Okay, that makes sense. So, let's not get too stuck on that now, let's keep moving on, and um so this is sort of setting the expectations, and this is actually how the final impact our rate is going to be determined. It's based on our expectations.

A

This is a key input into the smart contracts expected raise per impact hour and maximum impact hour rate. These two attributes actually go right into the smart contracts and determine how much people are going to get paid out and then the the third variable that plays a role is the total total impact hours, oh yeah, and then the fourth would be the total raise.

A

So the things we can control are these two parameters. We get to actually tune these for the smart contracts. Now this we don't really get to tune, but we get to kind of guide it.

A

We can sort of guide how many impact hours we want to accumulate, but by the time we launch the dow, this will be absolutely determined, it'll be fixed and then the un there's a fourth variable that will determine our impact, our rate and that will be unknown at the time of the launch, and that is how much we raise it's the total raise. So we can only speculate on here and that's what we're doing with our our target and our expected so really interesting dynamics going on here. There's a whole blog post on this.

A

uh I think it's linked here and I definitely recommend people play around with it. There's a whole nother math widget that actually exposes even more uh sort of parameters and windows into how this is working, and I really encourage people to play around with it here.

A

So maybe I'll just drop the link to this forum post in the tec, labs.

A

B

A

So that's the impact hours formula. Now we get to the actual hatch.

A

And total target tech tokens. I think this is what I okay so total target tech tokens. I think that's what's going to get set in the initializer.

A

Oh, we don't have an init function: okay, no it's the next one that was uh had that issue. So let's check this out.

A

So this is the actual hatch. Now this might look scary at first, so let's walk through it. So this is a fixed parameter. This is the total c stack tokens now, keep in mind, only c stack token holders are and allowed to participate in the initial phase. One. The whole tec is going through two phases: phase.

A

One is the hatch and then phase two is the initialization of the bonding curve and the conviction voting, and so what we're doing with our initial hatch is we're going to initialize a dandelion voting dao, and this is like the god mode of the dao. It can override any parameters of the dao.

A

So we have to be very careful with how we tune this at the beginning and we're going to use this mechanism the dandelion voting to vote on all the parameters of launching phase two, which will be the bonding curve and the conviction voting, and then, from that point on after phase two most decisions will all be made using uh conviction. Voting unless we really need to augment the dow uh for some reason at some point in the future, then we will have this dandelion voting mechanism to do that.

A

So, let's so, this is total c stack. Token whole total tokens. Now this number gets multiplied by this number to give us a total possible raise, because what this is saying is, if you hold one c stack token, you're allowed to invest. uh 1 100th of uh wrapped x die in in the hatch. So this is a multiple on your c stack tokens that you're allowed to invest. So we could turn this right up. It could say if you have one c stack token, you can actually you're allowed to invest.

A

100 x, die or wrap x, die per c stack token, and what you can see here? It changes the y axis.

A

So if one c stack token means you can invoke, you can invest one x dissent, then the total possible raise if every c stack token holder invested their maximum possible capacity, then we we would. The max we could raise would be just over. 400 x die, and this is the kind of thing because we're actually doing a dress rehearsal, um a test uh test, tech hatch, soon uh coming up in the next few weeks. So these are. This is actually the kind of range that we want to work with in that case.

A

So if, for the test hatch, this is a realistic number, but for the actual hatch it might be between you know, maybe 10 is reasonable. That means that gives us an 800 000 cap, okay, maybe uh 20 you know. So this is a really interesting one to play with it changes the maximum possible amount we could raise during the hatch now keep in mind, there's also the bonding curve coming in phase two.

A

So this isn't the complete picture of funds being raised for the token engineering commons, um so there's that hatch oracle ratio that we can play with now. This drove me crazy. um I I spent a lot of time thinking about this you're because it seems weird we have these same parameters showing up again.

A

We had maximum arrays and minimum raise in our previous dashboard here, and I struggled so hard to like unify these and only have these in one place, but I eventually came to the conclusion that it does make sense to repeat those parameters, so we have them here. We have our maximum raise and our minimum raise as part of our impact hours formula. We're also going to have them here uh for our hatch visualization, but they're different in this case, they're, not absolute amounts of x die they're.

A

Actually uh relative amounts from zero to 1 of like or in this case 0.5 to 1. And it's saying are we what happens if we hit our max capacity? What what happens if we hit the maximum raise possible or if we only hit half of it, and so now I'll explain what you're seeing here in this chart, so we have in blue the maximum possible raise and what what this is. Is it's actually a distribution of the c stack.

A

Token holders so, in a ranked view from most tokens held to least tokens held, so this index 0 on the x axis is the person I think it's griff green, maybe he's in this call right now, I'm not sure, but um griff green probably has the yep. So this is him here. So this we can see how much griff can actually invest in this hatch, and so, if we turn this all the way down, oh yeah, this isn't a total raise. This is actually per person. I believe this index.

A

So if we turn this all the way down to 0.01, then griff can invest up to 400. Just over 400 wrapped x die into the um into the hatch or the test hatch, and as we turn that up it increases yeah. I kind of forget actually what this axis is. If it's per person or if it's total, it would make more sense if it's per person, because on the x-axis we have per person um but anyways, let's keep going so we have the cap raise now. The orange is our expectation of like okay.

A

Do we expect only half of the c stack token tokens to invest, or do we expect all of them to invest and then or somewhere in between, and we can slide that range and what we see down here is there's. Some of these numbers are changing, so the max goal is changing.

A

So if, if a hundred percent of the c stack token holders invested as much as they could, then we get this case here where there would be a total of seven thousand one hundred and twenty x distinct- um and this is, I think this is pretty clever. I took the geometric mean or okay, let's just forget, geometric. Let's think about the mean this means your average c stack. Token holder has to invest, 28 x, die okay, and then that would get us to this case.

A

But this is not just the mean it's actually, the geometric mean, which means you're. Gonna have a distribution of some people investing 10 times this amount and some people investing one-tenth this amount. So it's like an order of magnitude average. So this would. This would essentially be the middle of this, because this is like a you can see. This is a power distribution.

A

This like grows exponentially or diminishes logarithmically, and so that the middle point is going to be this geometric mean, and so the people with the geometric average of c stack tokens are, are expected to invest about. 28 x, die in this case, and griff green would be expected to invest 10 times that which would be around the 280 mark.

A

You know, the exact amount in this case is is 800, but following the geometric, oh because he'd be at the end of that, just the very end of that distribution. So a lot of these people would be expected to invest around 280 x die and a lot of these people would be expected to invest around 2.8 x die and then we get the geometric mean in the middle, um and then we also get the hatch tribute. I think, there's another: is there another hidden visualization here, or did I not render it? Let's see?

A

Yes, so we have another. So let's change that into a column as well. So we have this.

A

um I think I just have to try changing this into a column.

A

A

Okay, so then we can't quite see the table. That's there's always trade-offs. uh So let's keep looking at that. um We get the I we're already. uh I think, for the for this whole hour. I'm just gonna go through this notebook. um I think that's we're already on that trajectory. So that's, okay and we'll get this all modular later, so we have the xdi funding pool cool. So we have the total x distance and that gets broken into two categories, which is the hatch tribute and the funding pool.

A

So these two numbers would sum to be the total x distance, and so, in this case, with these parameters, where we reach our max capacity goal, um we're going to have 133 x die in the in the hatch tribute. So this is going towards.

A

um Let's see where does the hatch tribute go, griff towards the pool for public token engineering, goods, yeah.

B

So the the hatch tribute the main thing here for the hashtag is that it's non-redeemable. So when, during the hatch, the people who put funds in there's like two tributes, there's the cultural build tribute and then there's the hatch tribute and uh they can pull out the the their share of funds that aren't in the hash tribute.

B

So uh when you upgrade to a commons you'll, be able to choose to add more funds to the other um to the conviction, voting pool if you'd like so because the commons upgrade you know you can send money different places. So uh it really. The biggest impact is on how much you can redeem. As a token.

B

Yeah, there's two vaults: basically, there's a vault. You can redeem from and a vault that you can't and.

A

I guess that plays into this next section, which is going to have this. uh The rage quit hours. Is that sort of the? I don't want to spoil this next section, because it's the juiciest part, but essentially um hatchers, have the opportunity to pull out of of the hatch um and but save for this hatch tribute. You know if you, if you invest in this hatch and then later you decide no, this isn't for me. I want to. I want to take my money back.

A

Well, the hatch tribute is um it's non-redeemable and what's cool here. Is you know at the end of the day this these things it doesn't matter how many tokens there are because it can always just be normalized, but it's fun anyways.

A

So in this case, if we hit our match max goal, then we're going to have a total tech tokens of 26 million 26 million tech tokens so, and I think we can change that based on how much we that's based on how much we raise so I'm going to change this back to a row, so we can see that table.

B

There's also the hatch exchange rate, which.

A

Changes, oh yeah, that's totally gonna yeah, so maybe people want more tokens, and this this is really. This is all the hatch exchange rate does so and like in theory in all the math. This should never affect anything. It shouldn't change the outcome. It shouldn't change the amount raised it shouldn't change, anything it's just going to change the number of tokens and the price per token, but in practice we know that humans are really sensitive to numbers and what they look like so in in practice.

A

I would imagine this would actually be very important and will change the the outcome. So so it's going to be a fun. This isn't going to change how any of the visualizations look, but um still we we encourage people uh the hatchers as they're playing with these parameters. You know put some thought into this because, even though it doesn't change the plots and the numbers here, it probably will change the the total end result and the outcome so it'll be interesting to see what people choose for this exchange rate.

A

um Oh yeah and then the hatch tribute here so we could turn that way up and that will change these visualizations over here. So you we could put the majority as a non-redeemable lockup, but that might that's probably not a good idea because it might be kind of scary for the investors. So five percent is the reasonable starting place for that, and I think I have to change that parameter a little bit. I'm gonna change the step.

A

um Yeah, I'm going to add a step to this step, equals.

A

A

And then we can it's more granular.

A

So this is the hatch view. This is probably the hardest one to grok. You know I recommend everyone just stare at this for a while. I think it'll start to make sense if you really read through these uh headers here and what what is the outcome and just keep in mind that this is c stack token holders, and this is the amount that they stake per holder and yeah. I've answered my own question that I had earlier. This is per holder because we see that the total raise is like you know.

A

The cap raise here in this case is 3560., so for any like calculus nerds. uh This number would be the area under this blue plot. This, like the area of the blue, so you could kind of do an integral and you could you could even ch like for future, to generalize this to all dow hatches. This could actually be like a probability distribution in this case. It's real live data, but this to generalize this to a theoretical dows. This could just be a probability distribution.

A

So so that's cool, that's the hatch, and let's see uh this is my favorite one. I think this is so cool the dandelion voting, so we're going to have total tokens which we get from now. We want to populate from here, but I've filled this in with a default of 6 million. Maybe I'll fill it in. What's, let's see what the default is, when I run this, I think it should have default parameters and our target goal is actually 17 million.

A

So I'm going to change this to 17 million as the default, and let's run this- and I can take away this or I can set this so total tokens and I can pass in an initializer the same. That is the default anyways, but I'll just pass 17 e6 with 17 million.

B

A

Constant parameter, total tokens cannot be modified yeah, so I'm trying to oh that's because I have this constant thing here. Let's just take that out.

A

Okay, it works.

A

Now we can actually change this, let's see if it's reactive. So if I put um a hundred thousand okay, so it changes the axis. So that's cool. So same thing like I was saying before this, like total number of tokens doesn't actually change like it won't change how this plot looks. It's just gonna change the axis here. So let's uh turn that back to the default 17 million.

A

Now this is fun. This is like. Finally, we can understand, uh support required and minimum quorum for dandelion voting, and it took I made so many really inaccurate visualizations that didn't make any sense until finally, I I worked with uh griff and jake a lot on this which uh helped so much. uh I wouldn't have been able to figure this out on my own, so we can change this minimum support.

A

So what minimum support is? Is it's the.

A

Percentage of token votes that must vote yes to pass a proposal, and so what we see here is like, as we grow, the number of tokens that have voted the amount that have to vote yes, which is this line. So the important thing to consider is this boundary line between red and green, because that is the boundary between a vote passing and a vote failing, and I want to do, there's a few future projects here that maybe some of you are imagining already, and I can definitely imagine some.

A

We could actually now run a simulation with this, and what a vote would look like is a vote would be a line, a curve that we could superimpose on this plot and it would grow as more people vote. The curve would grow on the x-axis. It would move to the right and as people vote oh, this should be I this might be a bit confusing token parody.

A

I think this could just say yes votes, um as the number of people who vote yes goes up, the curve would go upwards, so how a vote gets passed is first of all. Let's, let's check check out this minimum quorum and I'll explain that in a second, but first of all, the number of tokens would have to surpass this blue threshold to simply be eligible. That's the minimum quorum accepted. It means we need at least um 12 percent of all tokens to vote.

A

uh Yes- and this is a very subtle detail- that's like really confusing too, um and it makes total sense if you like, really flesh this stuff out and think about it. It makes sense and it's an improvement on past past dandelion implementations, but we want this minimum accepted quorum to be the percentage of all tokens that have voted yes, so this means 12 percent of all tokens need to vote.

A

Yes on this proposal in order for it to pass that's a hard constraint, and so in order to do that, it means we have to be in this green area. So, if we're in this green area, it means we have surpassed minimum accepted quorum and we have surpassed minimum support. It's actually possible.

A

This visualization could be a little bit deceptive because it's possible that we're like way over here on this chart like a whole bunch of tokens have voted and we haven't we're not even past minimum quorum because minimum quorum, like I said, is total tokens that have voted yes. So when someone votes no, it does not contribute to the quorum. It does not increase the quorum towards this threshold, so we could be like way over here on this chart and not even past minimum quorum.

A

But if we're anywhere in this green area, then it means we know that we have indeed passed minimum quorum and we have indeed passed the support required. So what we could do, we could play out a simulation where we see this curve grow over time, and it would kind of wave like a very controversial vote, would actually dip into these.

A

You know passing and then maybe people would come to block it, so it would dip into the red and then people would show up to pass it so dip into the green and then people would show up to block it and it would pass into the red and it would be like a football game or a u.s election. You know you'd have to get your popcorn out and uh you encourage all your fellow hatcheries to vote and and in that sort of simulation you see there's some parameters here that are not being visualized yet.

A

um So we have the vote duration days, which is, um I guess, the default here I put three. I I think I need to go through this and update all my defaults to match the tech spec. So that's one thing to be done, um but vote duration day. Maybe it's three. Maybe it's seven! This means a vote can only and I'm I'm going to go through these.

A

I might get some of them wrong and griffy can kind of jump in maybe at the end we only have four minutes left of you know I apologize, but also. I think this is good stuff sets the foundation for uh the next iterations of this, but vote duration is like the total number of days that we have to vote, uh so we, it would be a seven day simulation that we could run on on a vote moving through this space, the vote buffer hours. It means if we initialize a vote.

A

We can't initialize any more votes for at least eight hours, so we can have votes running in parallel at the same time, but we can't like start two votes at the same time. We can't start you know or unless we turn this to zero or okay. Minimum looks like it's one. So if we start a vote, we definitely can't start another vote for at least one hour, so we can't just spam votes and that ties in I'll skip this one for a moment toll gate fee. So this is the cost per vote initialization.

A

So if we say hey, I know I got all the right parameters to launch the bonding curve and conviction voting and I'm gonna put a proposal forward. Well, you need to at least stake. Some x die uh to show that you're serious about that proposal. So I think the default is three. I think it's pretty reasonable, but you know this goes all the way up to 100. Maybe we want it to be very expensive to propose a vote um and that this this requires people to put in a lot of thought.

A

I think this is a very important parameter to consider um and then finally, the rage quit hours. That means it. Okay, so we've put forward the vote, we're upgrading the the tec and it's gonna. It's actually passed it's already. The vote has passed well. This creates a buffer of 24 hours saying that vote, even though it passed it's not going to be executed for for 24 hours, and this connects back to what we were talking about before um the redeemability investors can pull out. They can say, okay. Well, I didn't like that vote. I voted.

A

No, I'm not happy with the way this is going. They have 24 hours to to redeem all of their initial investment and pull out, um and I think that's a great opportunity and it it we could re. This has a more formal name. uh We could rename this, but I really like this term ragequit hours, it's so clear um and griff did I get any of that wrong? Is there anything that should be really corrected uh like.

B

I don't I don't. I don't want to say wrong, but there's always more details.

A

Yes, there's a lot more details here.

B

Nailed it, this is amazing. This is incredible. I I mean it looks done. It looks basically done. I assume you just want to do a couple more iterations, but well.

A

That's a good so I'll go quickly. We got two more minutes I'll go over. What the next steps are. uh I want to pull all of this out inspired by denilo from the wednesday lab awesome hacking on git coin, and I encourage everyone to like stay tuned, because these labs are expanding. It's really exciting, but the next step here is. I want to pull all this out this is we have these nice classes.

A

um This is like really great. We should pull these out and put them into a python module, so I would go like a new. um Let's see, we can go like new file uh rename and I'm just gonna call it like t tech. I was gonna call it tech for tech, hatch, tech, dot, pi, and then we can go in there and basically we just want to copy all this code and take it out of the notebooks, and what that's going to give us is we're going to be able to make a new notebook.

A

That's super clean we're just going to go from tech, import impact impact hours and then and then we'll go. You know display impact hours and we'll get that whole visualization without any code. uh You know just this like one line of code, so it's going to give us a really clean interface and then, finally, that will be basically done that people can play around with it uh future iterations there's things that could be improved, but this could be. This could be the test.

A

Param vote for the test. Hatch, basically- and I want to find a piece where I can export this as a pdf, so anyone can play around with these parameters, get the visualizations and then just click a button at the end and it exports a pdf and then it's really easy. Anyone can just read the pdf like straight from the github repo. You can open up pdf files and it would make the voting process very easy.

B

This is so incredible and I really love the idea of making it a module, because other people will want to use create dandelion dowels without any of the extra stuff. So if they have this, they can visualize what their, what their requirements are for things to pass and better understand what these, what these pieces are. I mean, maybe eventually this could even get integrated into one hive's um deployment. You know uh so that when people are putting in they see these graphs exactly.

A

Yeah- and I want to share with everyone that this this has been- this is kind of the end of a series or towards the end of a series. This has been about four labs that we've been collaboratively working on this. So if you want to really see the process of how this was built out, um go check out the youtube playlist for the tech labs.

A

And shout out to jake and griff for sunday param sessions, that's where and anyone's welcome it's in the tec calendar so stop by on sundays, for some hack sessions on this sort of stuff and we're a couple minutes over. So I want to thank everyone for coming out and stay tuned lots more lab stuff happening wednesdays and thursdays check twitter. I guess.

A

Thanks phil get a new mic.

B

Thanks john, that was incredible. Thank you. So much thanks.

A

B

A