EthCC 2: Paul Kohlhaas

cool hi everyone my name is Paul I've kind of been exploring the area of how we can use bonding curves I'd say really over the past Simon first started working on it but then intensive lis over the past one and a half years and as many of you kind of know bonding curves first emerged as as a way to curate attention and information and specifically around around memes and kind of from this first idea around using using bounding curves as memes I think we've seen an entire subset of projects emerge that specifically just look at how can we use cue Asian markets to to curate assets and objectives of all of all type so what the goal is for today is we need to walk through a few engineering parameters that that have emerged in bonding curve design and see and apply those to two different use cases and then specifically at the end of the talk we go a little bit into the challenges that we're trying to alleviate with bonding curves in the pharmaceutical industry yeah it's kind of the the goal for today's talk maybe just a set of hands who is fairly still to the topic of bonding curves and it's kind of like diving in for the first time ok ok cool that's great because I have some intro and recap slides so you guys can actually yeah like dive into the topic for the first time a little bit so I'm seeing like 60% of people explore this for the first time cause so I'm gonna give a short intro in what our bonding curves what can we use them for why they're exciting then look at the different engineering parameters that have kind of emerged and then we're going to look at really how can we use these parameters to re-engineer incentives and and apply them in systems design and then in the end I'm going to go a little bit into Marco and just kind of given industry examples of some of the big problems in in pharma cool bonding curves condom curves what are they so they're really an economic incentive their economic incentive coordination tools so we can use bonding curves to coordinate people around a common goal acid there also ought on market makers what does that mean for the first time we can kind of use the smart contract type to deploy an acid or create a essentially create a market around any type of acid and why I think that's really exciting is we can take an asset that's worth $1 or we can take an asset that's worth 100 million dollars and and have something traded around that we can even take assets or even objectives where we don't know what it is worth and just put it in there and let the market decide and this is really I think a very revolutionary principle because before whenever we create an a market for something they'd always have to be a by a by side and a sell side we still have that to a certain degree the burning curve but we can bootstrap any kind of market very quickly there are also Q Asian markets circulation markets the goal is to curate information from the crowd by by harnessing different signals and through that producing information asymmetry so these autonomous markets that we're creating here they incentivize people to behave differently and and maybe they allow systems to behave differently if we think about one of the I think like one of the biggest the biggest things that blockchain networks really achieved at scale were to change people's behaviors yeah and maybe as a last point the markets currency is a proxy for attention in that specific market and that rewards early adopters so this is probably the most common running curve kind of display of like what a banach curve could look like it's essentially a it's essentially a supply and demand curve taken out of microeconomics and we'll get we'll get into what actually happens cool so what is a bonding curve do exactly it's actually really really really easy it's a small contract that accepts in its most simple from ether or any other a or C 20 token and in exchange it mints tokens in its denomination so you send ether or or die to to Ivana Cove and in exchange you'll get tokens of its denomination according to where at where which point of the curve you're in the the contract holds that ether or E or C 20 in reserve so that's the bond and users can now freely trade up and down the curve with our contract what's really important here this is not an IC o---- it's not a fundraising mechanism because no one gets the money in the curve the money always stays locked up in the contract and kind of builds this permanent liquidity pool that anyone can trade with there are exceptions of this possible and we'll can it we'll get into that later yeah this is the most typical typical design so you have a price per token which is the y-axis and then you have the number of tokens and circulation which is the x axis that's predefined by a slope formula in this case this is kind of a quadratic function and that's something really exciting is well there's many different functions that we can play with for these designs cool this is kind of a little overview that I did a while back that kind of just shows all of the different parameters that that are in existence today and how we can use them actually not how we can use them but so let's start with the token the token issuance so the token issuance is essentially what type of token does this contract issue in the most common design it's just a simple IOC 20 but it could be an A or C 7 21 so we could have a bonding curve that issues non fungible tokens it could issue security tokens it could issue anything so that's a first important thing the second important thing is the token supply how many of these tokens are there is it an infinite supply or a finit supply many of the early designs that came on kind of just assumed that this curve would go on forever without actually capping supply and that can make sense for a variety of use cases for other use cases it does make sense the other and other aspect of supplies that you can build vesting and so vesting schedules into into the contracts themselves then the next thing is the collateral that the contract accepts so this will be what we're sending to the contract or trade with it a lot of early designs were kind of just assuming eath but it's a problem if you have I think it's a problem if you have one volatile asset that you stake into a bonding curve because now you have double volatility so we've been playing around with implementations that used I kind of as just as a stable collateral that you sent that you used to trade with a bonding curve you could also use other ERC 20s as collateral there's a lot of options around the collateral then the next one is the asset or the objective and we'll get to actually wait let me jump straight to that so how do we act what are we actually trading here because we need to define what the objective of the bonding curve is like what is that asset in in early days it was a lot of talk about memes so let's take an IP FS link and attach it to the contract and now we're trading attention in that meme and then Billy Rena camp who's also just one of the leading like thought figures and defining defining the space came in think early last year and came up with a concept of refund tables and we fungible basically work with onl just random work so we function well so basically we take an NF T and we can stake it into a bonding curve so now the bonding curve contract is the is the owner address of that NF t which now means we could trade attention in a crypto kitty so it means we can take any non fungible asset and make it fundable again and the way that we're that we're applying this is essentially saying patents and intellectual property is a timestamp data claim let's attach that to let's attach that to an NF T and then we can take that NF t and move it into a boning curve and now we've made we've made that that market tradable with an underlying asset yeah there's a lot of different asset types that we can stake other people using this for example for it to trade attention and value in a community so then we're more trading the asset that's traded is run and objective it could be for content and something really important here this the asset will define what the legal classification are the bottom curve issued token is if we're trading a meme probably not a security if we're trading any form of asset could likely be a security so so that's important how that relationship is defined then the other thing is the the function a lot of people have written about different functions we'll get we'll get to some cool functions us now and then a last parameter is also how pricing is implemented in the curve so specifically the guys working around continuous organizations are working with different price in and price out mechanisms which essentially means you buy into the curve at a certain price but your sale price is always lower because it's a different curve and that means that bonding curve is not generating revenue yeah cool so kind of getting back to this example of like of refund tables it's one real example how we are like a first design step how how we're using the the traded acid are the bonding curve in our case to represent a piece of intellectual property because it means if we can trade crypto kiddies why not why not use n FTS to attach to attach patent data to and then essentially allow to trade attention and value in an ax patent yeah some common curve functions of bounding curves include exponential functions they are relatively easy to implement but also extremely volatile the price stays lowest for about 80 percent of the curve and then accelerates very unmanageably linear functions so the ocean team has done a lot of work around modeling and experimenting with linear functions they're quite straightforward but in the end you end up with enormous returns there's war based functions and we'll get to sigmoid functions as now which is some of the functions that we're we're looking at using there's a great article from Wilson Lau who who does a lot of analysis on these different curves if you're if you're interested to read more about that so sigmoid functions are cool because they were best suited for markets that stabilize after a certain inflection point so if we think about this there's very slow growth in the beginning and low capital accumulation but then there's a point where the market interest around around this particular asset starts accelerating and once the acid is de-risked the the the burning curve stabilizes again at that upper level these these are curves that are commonly found in population growth and yes so there's quite a few natural processes as well as like the the typical technological development cycles that different technologies go through that use sigmoid functions cool so how this could look like for example is we have someone in our case we would take an N ft and stake it into a bonding curve the original creator of that curve could allocate himself a certain share in the market so kind of so that the first set of tokens that are issued or collateralized with the NFT the NF t value itself so imagine if this was a crypto kitty I would say cool this is the market cap that I'm starting at but I no longer own that crypto kitty I'm staking it into the bonding curve and in return for example I'm getting 20% of the other total tokens issue another really interesting mechanism is I think we just mentioned that as price in versus price out and there's a lot of different ways that people are starting to be engineer there's an experiment with different kind of building revenue models around bonding curves through through different price and perhaps how this could look like from a very simple perspective is if I buy in at Point a I immediately would make a loss if I sold at point B so this in a way disincentivizes pump-and-dump behavior because I'm already factoring in that I'm gonna take a loss which means now I'm actually committing to holding this asset to at least point C where where Point C would be my break-even price and kind of the difference between the curve could now be used by someone who is for example actually using that funding to build out that asset cool so how are using this with with molecule so molecule protocol essentially enables the distributed research development and funding of pharmaceutical IP and our fundamental design goal with building molecule is to distribute the high risk cost and the ownership centralization around drug development and so maybe just so some of you understand a little bit better at what the kind of what the context is pharma is one of the it's a huge industry but it also has an enormous amount of problems if some of you are familiar with the problems in the US healthcare systems the the prices for drugs continue rising production costs for drugs continue rising it takes very long to bring something to market because it's such a close to us development system so we when we first started looking at this we started asking ourselves like what are the key what are the key systemic problems in that industry and I I personally I I got very interested in the pharmaceutical industry just after high school because I started realizing there was just a lot of problems that seemed unnecessary in the way that that industry was structures from from a systems design perspective and so we look at some of the different problems and we'll make the loop later how can we apply those to how can bond occurs be useful here so and it's it's actually common in the pharmaceutical industry itself isn't an innovation crisis because they're bringing less and less drugs to market they have to raise the prices on those drugs and it's very problematic and so one of the fundamental issues with the patent industry is that patents act as monopolies and they create closed source research and development negative data doesn't get shared this causes rising drug prices and the market produces drugs often that maximize revenue but not necessarily impact so we came at this from a mechanism design perspective and looked at what do all of the different stakeholders in the pharma industry actually want and if you think about pharmaceutical development it's really IP production then producing data about that IP that is funded through through investment into the IP so we looked at the different stakeholders and then this is kind of a typical diagram that's how the drug development process works today so you move through these different development stages there's early drug discovery then clinical trials phase one trials phase two 12 phase 3 trials and then a drug gets approved to go to market and this below here are like typical innovation mechanisms within a company where a lot of different ideas are gathered and then you kind of you curate the best ideas or and in this case all that from a suitable development is is curating curating valuable IP and taking it through different development stages each development stage needs more data to take the drug further to Margaret so we started realizing that cute markets principles could be extremely interesting here but because in the discovery stage there's an enormous amount of problems that you have about liquidity of these assets and getting funding for an assets it's a chicken and egg problem you have a piece of valuable IP an interesting molecular structure but you need an enormous amount of money to prove that it's valuable and because it's so centralized all of that value needs to be created in one company so what if we broke up that market and the way that we essentially break it up is by the by the mechanism that I explained earlier we take we essentially allow anyone to to take a piece of intellectual property and put it into an open market with very low friction design and then let the market curate information around it because now that you have an open market where people can trade any research information in any research data about that asset becomes valuable because as soon as I for example as a researcher published something out in the open other researchers or speculators might see it and thus allocate more more capital to that market on the other side negative data also becomes valuable because you can short these markets we've been working with the ocean team a little bit run thinking cool how can we short bonding curves at the same time which would alleviate some of their pump-and-dump like behavior and also make negative data valuable which is a huge problem in the in the scientific community yeah this is kind of like oversimplified again completely I was simplified but what a what the drug development process mapped over kind of a bonding curve could look like we've we've realized like we can't we can't map the entire development process through that but and that specifically the mourning curves as price discovery mechanisms are specifically valuable in these early stages but it is a similar mechanism in that the amount of capital required really comes in these later clinical trial stages cool so from a reengineering economic incentives what Brownian curves in this example in this market really allow us to do is it enables distributed price discovery which moves power away from like from individual actors and and lawyers and evaluators that try to try to guess the price of more or less especially if it's in its early stages and puts that into an open market and that in return incentivizes open source research and it makes negative data valuable because anyone can contribute to - increasing the value of an asset it also curates the most valuable IP and it distributes ownership around it and that in return we believe could lead to a much more competitive pricing mechanism for when a drug comes to market because at the moment what a patent allows us to do is essentially just set an arbitrary price and with this mechanism the pricing the pricing can become more market driven we're also looking at with the governance functions that would come out of these markets so combining bonding curves bonding curves as a funding mechanism with governance functions because at the end of the day you'll still have to decide what how this drug goes to market and how its licensed and we believe an open system like that could lead to much more diverse treatments coming to market and the last point is it's never been possible before to create frictionless markets like this so I couldn't I could never secure it as an asset like like a patent that is that is only worth $50,000 and at the same time I can't prove it's worth that much but with this with this design I can actually from very early-stage IP to latest h IP i can create markets around it and access funding and liquidity last small point what are the key challenges there's a lot of challenges that still need to be overcome with with these designs i think we need much better UX and much more gamified experimentation in order for bonding curves to really be able to slowly chug ahead we're not planning to take this the system that we're building live probably maybe it flows better within within a year at the most because there's so much experimentation to be integral integrals and logarithms in solidity are really really hard so if you actually intend to take some more fancy designs on chain it can get very gas intensive there's a lot of challenges there we need much more simulations and modeling and tools block science and incentive I are really building a lot of cool stuff there we also built a bonding curve simulator which is open to us online if you if you're keen to test out different designs and see what they do we need I think off chain experiments because if we can model this off out off chain we can much better design an on chain system and not get stuck on implementing complex algorithms and solidity but actually really test the system from a game theory perspective then I think curve governance is really really interesting because once an assets live if you ever want to change its valuation that has major implications on the market and then the last thing is also like dynamic making curves more dynamic being able to perhaps react to outside Oracle's that changed that again would change the curvature cool I think I'm on time if you guys are interested please reach out if you have anyone working in pharma want to share experiences about bonding curves anything please feel free do we have time for questions [Applause]