ETH Rio 24 - AI & Trading: What's myth and what's real? | Guruprasad Venkatesha

Transcript

for um and now I'm going to switch to

English um I would like to introduce you guys to bcube they are a AI driven platform um but not driven in like trading in charts and so on but rather using AI to know more about sentiment and how um people are trading towards an specific asset and for that I'm going to invite Guru here to introduce us to B Cube welcome Guru thank youon thank you hello everyone uh let me introduce you to bcbe and I

will also be speaking today about uh what is possible with AI and what is not possible with AI uh because right now the very big trend is going on about AI so everybody are interested in this topic uh but there are certain things which is true and certain things which is not true at least uh for today it's evolving very fast so that's what I want to present you today and I will present how exactly we are doing it um we're also showing you some architectures like how exactly our AI works so you will get good information about it so um I'm the co-founder and CEO of um bc. aai um we call it as a next chapter in autonomous

AI trading because completely it's automated with no human intervention so uh the common things um that that we see uh people using Ai and which is not true is assuring you some returns like every month you will get 5% 10% or like every year you get this much trust me uh no matter how strong is your technology no matter how smart you are you can't assure any kind of returns because it's unpredictable anything can happen so anybody if he promises you some kind of return then then most probably it's a scam so that's one thing what we need to

know just because there is technology which doesn't mean like someone can really promise you any returns um second thing is like uh you know like now as I said like AI boom is going on and lot of projects who had nothing to do with AI they all said like no we are also AI projects we are also AI projects we are into this um but have you ever gone to that community and asked like is there any proof that that you belong to AI category or what exactly you are doing in AI um because they don't show any documentation nothing so that's very important that you need to have a proof uh to say that they really have an AI there so that's

what we need to know and coming to the third Point like blackbox Solutions which means like they say that we use it works and they don't give you any more information about it there should be a product page there should be some documentation for you to know like how exactly it works and on what basis it's making the decisions or even if they're using AI for different kind of things they need to have some kind of information about that how how exactly they're doing it they don't need you um you know to see all the things what they are doing but at least like they need to give you some bird's eye on the subject uh if that is not there then again it's a kind of Pitfall and it's a

myth what they're trying to um fourth thing is like U many of the trading companies they try to show huge Returns on their webites and how they do that is by mixing the back testing results with the live testing so in the back testing any strategy can show you 1,000% 5,000% or even 10,000% heavily overfitted models and in reality it will not even do 100 of it so when you see those charts like you get excited and you want to buy the product uh but like never trust back test uh trust only forward testing trust only live results that's what matters but back testing is just a kind of uh um bookish

story so um what is the ground reality a AI is very good to recognize the patterns um like like nobody else this is how like our system like beginning in the year like in January like when the new bull market started to come it was predicting and it was trying to make more bigger trades with more bigger quantities because it was able to recognize certain patterns like which has happened in the past and also there are some reasons like why it has to make the trade uh when we go to our own dedicated slide I will explain to you a little bit more about it and even if you provide a large amount of data to AI like just not just the math models but um even giving them

you know s sentiment data exogenous datas um well valet moments volume data any number of data that you feed it will be able to consume all this data and will be able to give you the results way way way faster than anybody else can do it so AI is really good for that and it can process the data very very fast um then this enables AI with powerful Insight abilities which means uh it consumes all these data and it tries to interpret this data in the form of traits executes that one so if there is a human being then they can't do all these things on their own in such a fast pace and with such kind of like uh

accuracy like the way AI can do so that's what AI is very good at and and fourth one is like uh whenever human is trading like he always tries to make uh his own assumptions um the assumptions may may be like okay um I don't know my model is saying it's a sell but I feel it's a buy so there is no such um um need or there is no such kind of um uh proof for you to make that decision this decision is purely based on your emotions and emotional trading is not going to work and when you lose a trade you will go after a trade you will again go after a trade to get back those losses so this is how we human beings work but uh there

is no rational behind why you have to open a trade so that's what AI is good at so um crypto trading is very uh everybody of us know here like everyone is into crypto trading or at least done some crypto trading here okay so so people who raise their hands they know it for sure like it's very rewarding it can make you a lot of money for sure but uh at the same time it also needs a lot of technical uh expertise I used to be an Institutional Trader for Morgan Stanley I it took me a long time to learn all the tricks of the trades and how to build a strategy how to you know optimize a strategy all these

things so it needs a lot of expertise and you need to spend a lot of time it's a full-time job you cannot be a very good Trader like if you are just doing part-time or like a hobby um that's that's the point number one and second thing is like uh crypto markets are very volatile it's good and it's bad both uh if you're a very good Trader if you're a volatility Trader you can make a lot of money but if you're not very good Trader then you will not be able to make money because you get scared you go short markets are falling down it again starts to move up 2 3% you will again like go long so this kind of monkey business of short long short long long short long short this keeps going on because you're not sure about it so that's why um the

volatility can play a role if you're not a good Trader if you're not a professional Trader and the third most thing uh what we mastered and what I will speak later on is about the consistency because if you just deliver returns in bull market anybody can do that um if you deliver returns only in the short Market anybody can do that also when markets are coming down but irrespective of the market condition if you keep delivering consistently um then that shows your expertise and that's very hard to do and that's what a very good algorithm is supposed to do and fourth thing is like it's unpredict and risky right um as I said in the beginning like whoever promises you any kind of returns like

it's not possible and it is risky and it is unpredictable anything can happen so now uh is a moment like where I will present the ground realities of uh what we do and how we do that um first of all like AI is a very ere exploited marketing Buzz world that we all know if something is not selling in crypto these days they just put an AI element into it they can sell their tokens they can sell their nfts they can sell everything now um that's why it's been very overly used word and also uh it is true that uh recent developments in AI uh especially after open AI chat gpta and all these things have come it's

become much more easier to build strategies or to test something so the advancement of AI is definitely going up and it is not the same like what we started six years ago because we started our company six years ago uh in AI research and development but back then like we we had to do so many things but now it's getting much more easier all these things we had to build some infrastructure on our own to test these kind of models but like now many kind of open sources and many other things are available which is very easy to make it um so what is bc. so we Pro we provide AI driven trading agents U which which is an

Institutional grade of very low latency which works on your own accounts so we have both custodian and non- custodian bases but uh the one what we are talking about here is like a non-custodial basis so uh as we say that like it's a self- adapting we any any Market condition it works like whether it's a bull market bare Market swing Market or range bond market we have a track records for the past 5 years consistently performing um it's institutional great because we have a very low latency rate we can execute 100,000 traes in less than 5 to six seconds so um and if you see the industry generally with the same

um models or same kind of things uh they work for some time and later on they don't work right like any kind of algorithmic trading that you might have seen but uh we have maintained this consistency for 5 years uh with very good train track record and going forward also we try our best to maintain our consistencies and that is very important in trading and efficiency in performance that's what I said about very low latency um and very quickly analyzing the data like we use sentiment datas also uh we built our own sentiment analysis engine we are not using nansen or any other kind of things so even if there is a slightest tweet which is quite indirect by Elon Musk like BTC is

going to move up or something we will be able to interpret like he is going to give a hint towards crypto so that's how we are going to assume that and that will have a power on our sentiment analysis engine so this is one of the strategies that we have so this is having a track record of three and a half years uh the sping spot trades only in spot Market no leverage absolutely nothing zero leverage just going long it trades on only top 25 coins very liquid very able one and it's consistently performing uh it makes the decision like how much it has to buy or you can set it up on your own dashboard also you can either make it completely autonomous or you can make

some choices of you no leverage nothing absolutely spot trading so this this strategy has been there for three years and uh if you're investing all the 100% of your money in every single trade then you would make 26x of your money so um why we are the best here because our our our goal was to give a tool where you didn't need to do anything on your own there are people like who like to um make some modification on their own they want to set up their own things they want to set up their own uh stop losses this and that but uh we had a different Vision like we are for people

like completely hands off approach which means like just plug and play just need to get your API Keys just do it and it's fine it's doing everything on its own so that's why we just focus on like that autonomous nature and we are the only platform like which provides you this kind of flexibility of being completely autonomous which means you can carry on with your everyday life you don't have to watch your account every day setting again like markets are bad like okay I put down that AI will do on its own if the markets are really very bad if the markets are really unpredictable it will be just using 10% of your money if it's really very good or if it feels like something is very strong it will take 40% of your money so everything is

completely autonomous and dynamic at the same time it's adaptive as I said like uh it doesn't matter uh what Market condition it is um it will understand and it will take the right decision according to the market Trend and also no actively involvement required as I said like uh it will do on its own until you really want to do uh we provide that option as well but we still prefer you to completely you know focus on your life and let it do its thing um does not take custody of your funds because uh we we we will not take your money into our account all the money is in your own account you will just provide us the API keys and we have permission only to trade not to withdraw your money so everything is happening in

your own account you can see it you can track it that's how our system works and it it minimizes the potential risk and maximizes the potential gains and when we say that um uh as I said like when markets are really bad uh it will adjust your portfolio based on that one it will take like only small amounts of money make small trades but when the markets are good it will really take the big portion and make the most of it and it will do everything uh dynamically and as the market condition changes it will change also so uh this is one of the most uh complex technologies that we have built which is patent pending uh so this is a very high level architecture the detail

architecture is super complex so uh the strategy Factory and strategy selector as we call as sfss um what it does is like provide n number of strategies Trend following strategies mean reversion strategies swing strategies um Range bond market strategies any type of strategies we provide even sentiment analysis strategies based on sentiments we provide even exogenous strategies which is based on like wallet movements uh Wales wallet volume unusual moments and how many times this one repeated in the past and when it happened like what happened all these analysis happens instantaneously in real time so we give that one to strategy Factory and then AI generates

the signals thousands of signals based on these conditions and then these conditions will be sent to the strategy selector and which is what is a secret Source like it's adaptive real time intelligent and it makes autonomous decisions it will select those strategies which which is relevant for today's market condition and those strategies is what like will be benchmarked once again with various other things just to see if um it's worth the trade by taking the risk and everything into consideration and finally it goes into your trade so this is this is one of the things we we have filed several patents and this is the one uh one of the one so when we see

like when we compare other solutions to our Solutions so what we see is like we are highly highly adaptable like for the market conditions and we are highly autonomous which means like no human intervention is really necessary and it's very adoptable which means like you don't need to select uh bull market strategy you you need to select like bare Market strategy and swing strategy or range Bond markets it does on its own it will select it will do on its own so uh we have 27,000 users plus users we are growing fast um we manage more than $7 million asset Ender management which is connected and we have made like more than $750 million in

trading volumes and we have several status like U trust pilot we are having very good ratings we have backed by laf French Tech which is the French startup initiative and we have been funded by the European Union to make the project we have been funded by polygon uh bitat and also we have Junior um Juni Enterprise innovate means uh Innovative company status which means it's an official accrediation by government of France as a research and development company in Ai and blockchain so we are the first company to receive this accrediation they give that only after reviewing your codes going through

everything and thanks to this we have 60% tax refund from all of our employees salary so we are paying them only 40% salary so we get back all the 60% and we are regulated in Italy as a virtual asset service provider so we do even custodian Services brokerage OTC and very soon we are starting Market making as well so um this is my team um so I'm the co-founder and CEO um I come from equities background and another co-founder comes from the quantitative Finance another co-founder comes from the the software engineering AI that is his expertise and uh who am I so I first

made my company when I was 16 uh which was automating the trades in stock markets I sold five years later to the company uh that company went public quite recently and then I wanted to have experience with large fund so I joined Morgan Stanley as an Institutional Trader on developed certain several strategies like 20 of my best strategies are still in use by the company and similar strategies I try to develop for my own company as well so um I am also like editor sometimes on fors magazine I write about crypto in Hungarian fors magazine um so this is about it so my most of my uh career was on uh trading investment I didn't do

anything else uh besides being a Industrial Engineer so we were on the fores uh magazine several times s uh we were also invited to be on the for stage uh to showcast our startup uh because we were we were one of the first ones to go into Ai and to become successful in Europe and today by volumes we are the third largest one uh in Europe uh these are some of the other events that we attended and represented the company and our Solutions uh before I say thank you I will say be Cube to the Moon our token and thank you so much [Applause]