The Pyth Effect: How the Oracle is Reshaping DeFi - TOKEN2049 Singapore 2023

foreign thanks so um Mike Cahill and I'm going to tell you a little bit about pith the Oracle that is powering High throughput defy so if you work at a trading firm or an exchange you've already probably heard of fifth if you're building a high throughput D5 application you've also probably heard of fifth but if you're not in either of those camps you probably haven't but if you've used over 200 High throughput D5 apps you may have done so thanks to pith and just don't know that I'll take two minutes now to First tell you what an oracle is and why pith is needed blockchain applications have no way to access external data they can only access the state of the blockchain data like the price of Bitcoin the price of Tesla shares Sports weather all needs to be brought on change somehow and that's where oracles come in they update the state with exogenous data and that state needs to be trusted or that data needs to be trustworthy so the first generation of Oracle networks solve for trust at the expense of speed think averages of averages slow updates Etc and this was fine for first generation D5 on slow blockchains but it wouldn't really work for high throughput D5 or derivatives on Fast block chains so the pith thesis is that defy needs trustworthy data at low latencies the unique part of this thesis is the low latency or high speed so why is speed important isn't this something that liquidation Bots or high frequency trading firms used to trade against retail well the answer is no the reality is slow oracles create adverse selection for retail and for Passive liquidity the reason for it is it encourages a private a market of private price discovery that can be used to take advantage against someone who doesn't have that same information it's faster oracles that put people users on the same even playing field I'll give you an example let's say that you are staked as an LP in the derivatives protocol synthetics and it's using an oracle that's only updating every 10 seconds a profitable trading strategy against you would be to trade here so you got your slow Oracle a profitable trading strategy against you would be to watch the slow price and the fast price and trade when they're dislocated so if your passive liquidity is updating only every 10 seconds and the market happens to be 17 15. but within that 10 seconds there's a market rally and it jumps to 1719. a strategy could be to buy from you and sell at 1719 locking in a profit at your loss if you do this on a fast Oracle which is updating every 400 milliseconds the likelihood of that to happen has been mitigated is much lower now there's some important Nuance to this which is how old the data is by the time it gets onto the blockchain and it's worth exploring in order to do so I'll make two general points about financial Market data number one Financial Market data is expensive in 2022 users willingly paid 6.6 billion dollars for financial Market data in the real world and number two all that data becomes free after about 15 minutes so it's pretty simple to deduce that the value exists somewhere between time of zero and time of 15 minutes in fact you can write a simple formula or function that shows that data is maximally valuable at t0 and decays as a function of time sorry so when you go to a website like coin gecko or corn market cap typically the data is about 60 second old old and updates every say 30 seconds the reason for this is that they have to aggregate the data using apis from centralized exchange and decentralized exchange average it together filter out exceptions and then publish on the website it's even worse with real world assets when if you go to Yahoo finance the Futures feeds you see there are typically delayed by about 15 minutes first generation oracles solving for trust added in a additional layer of latency where they would incentivize reporter nodes to take averages of the averages from coin Gecko and coin market cap it's a little bit like your mom putting on 10 pairs of gloves to keep your hands warm sure you're warm but you can't move your fingers this is a general diagram of what a legacy or reporter node Oracle Network looks like pith solved for trust without compromising on latency by having three breakthrough Innovations number one pith data comes directly from the source this is what we've referred to as first party data inside the pith Network there are 90 institutions who submit their own proprietary data this consists of trading firms and exchanges having such a wide robust group of independent Publishers makes the price stronger more accurate and robust number two fifth data Aggregates at the fastest block times in the world pith runs on a Solana virtual machine app chain called pithnet and Aggregates at 400 milliseconds there are 350 symbols that aggregate every 400 milliseconds and are distributed to over 30 blockchains and finally all pif data has confidence intervals at any given point there's no price of Bitcoin there's a price of Bitcoin on binance on coinbase on up bit sometimes these prices agree sometimes they don't typically when they don't agree arbitragers will do risk-free trades at the various exchanges to bring them back to equilibrium during periods of high volatility however they could have a wide dispersion in prices confidence intervals represent this phenomenon in real time a wide confidence interval indicates a wine dispersion in price consensus whereas tight confidence interval indicates a very high consensus price first generation oracles do not have confidence intervals therefore the users have to assume that the confidence is zero this is wrong and in order to preserve trust they would typically update less frequently so mom discovers North Face and now you can move your fingers here's a diagram of what the pith Network looks like in order to have a successful Network you need to have data providers that actually want to publish their data so you may ask well why do the Publishers in the PIV Network want to publish their data so as you remember I mentioned the beginning that Financial Market data was six and a half billion dollars in 2022. well it turns out all that money was paid to exchanges from the previous slides you'll remember and is probably intuitive exchanges without Traders have no data so one of the key breakthroughs of the pith network is it is it encourage a new set of financial data sources that never existed before in a similar way to Airbnb providing more rooms that competed with hotels pith enables an entire new category of trading firms to now be able to monetize their data so you may wonder what is the incentive of a trading firm to be able to publish their data on the PIP Network if you wanted to think that they were doing this for profit you'd be correct they're looking to profit from their Market data for the very first times trading firms aren't the only participants there's also exchanges why is that well six and a half billion dollars is a lot of money nobody knows how to monetize their Market data better than exchanges and the blockchains represent a new place to do this so I've told you a little bit about how first party data has enabled pith to have low latency or high or low latency updates but low latency updates is not enough to be able to provide the fastest possible outputs you also need to have a high frequency blockchain so because it's a crypto conference I assume some of you are technical for those of you here's a schematic for the rest of you I'll explain what this does so hit operates on a Solana virtual machine called pithnet which is validated by the 90 Publishers in the pith Network it has all the same transparency characteristics as Solana mainnet you could use a block Explorer and check the Providence of every price there's 350 symbols that aggregate every 400 milliseconds through Wormhole they're then distributed to over 30 blockchains let's look at some of the statistics of how busy that is there are 300 million transactions on pithnet each day that then creates 50 million Aggregates so 300 million transactions is a big number this is from these 90 Publishers publishing 350 different symbols at various times compare that to Solana which has 25 million transactions or ethereum which has a million transactions per day or optimism or arbitrum which each have about a half a million so pipnet is a very busy blockchain you can understand some of the technology that is behind pith that enables it to operate at massive scales larger than some of the largest layer ones in terms of the number of daily transactions so all of this is to validate this thesis that high throughput Define these trustworthy fast oracles so how is this thesis playing out pith launched on Solana in September of 2021 and then launched cross chain earlier this year this year alone there have been four new application Integrations on pif per week it includes lending protocols trading applications derivatives as you'd imagine but this alone doesn't tell you all that much another way to look at the usage is to look at the daily prices that are delivered remember there's 50 million Aggregates created on pithnet you can see here in January when it was launched cross-chain about 2 000 were distributed each day but then something happened we saw a spike and so what's beyond the details of this Spike there's been about two million updates per day last month turns out there's a new generation of applications that are now possible on fast blockchains that include derivatives and Perpetual markets that were not served before with Legacy types of oracles here you can see the growth of this segment of the market that pith has been empowering and here's where high throughput as a trend is emerging and some of the blockchains where pith is prevalent and finally I'll finish on the pith ecosystem which shows that Defy is in fact looking for high throughput trustworthy low latency oracles thanks for the time foreign