Rise of AI Virtual Chat 2020 · Berlin

Building a Money Printing AI

At a Rise of AI Virtual Chat recorded in July 2020, Fabian Westerheide interviews trader Raul Glavan, who describes building an AI system that automates and, he says, now outperforms his own decades of manual trading. Glavan traces his path from childhood chess and an early home computer through winning Germany's Iron Trader competition and the Forex World Cup to a system trained specifically on his own trading behavior rather than generic market data. He explains a mean-reversion strategy grounded in behavioral finance, buying into panic and selling into euphoria across stocks, commodities and currencies, using scraped sentiment data and LSTM neural networks with peephole connections built in Python and TensorFlow. He reports around five trades a week held roughly 30 days each, a 93 percent historical win rate, a profit factor of 24, and a Sharpe ratio near 20, which he calls unsustainable given his small sample size. He also describes overruling his own system during the March 2020 Corona panic and losing money as a result. The conversation closes with a discussion of universal basic income and equal taxation of AI and human income as responses to automation-driven job loss.

Key takeaways

  • Glavan built his AI trading system specifically to replicate and improve on his own trading behavior rather than training on generic market data.
  • His mean-reversion strategy buys into market panic and sells into euphoria across stocks, commodities and currencies.
  • He uses LSTM networks with peephole connections, built in Python and TensorFlow, and combines them with scraped social media and insider-trading signals.
  • He reports a roughly 93 percent win rate and a Sharpe ratio near 20, but calls both figures unsustainable given his small sample size.
  • He and Westerheide argue AI-driven and capital income should be taxed the same as labor income to help fund universal basic income.

Detailed notes

From chess and a home computer to algorithmic trading

Glavan describes starting chess at age six, getting an early home computer as a present and teaching himself to program on it, and making his first real stock trades as a teenager on gut feeling before building his first rule-based trend-following trading system at 16. As a student he built an early robo-advisor, though its exact name is unclear from the transcript, and won Germany's Iron Trader competition and later the Forex World Cup, where he says he generated returns of several hundred percent virtually over a few months using high leverage.

A mean-reversion strategy built on behavioral finance

Glavan says his strategy exploits behavioral biases such as panic and euphoria: when other traders are panic-selling, his system buys, and vice versa, a move he calls counter-intuitive because it goes against instinct. He scans stocks, commodities and currencies rather than single stocks, citing the roughly eight trillion dollars traded daily in foreign exchange, and avoids concentrated single-stock bets because of bankruptcy risk, citing Wirecard's collapse to zero as an example of a stock that kept falling after an initial panic signal.

Sentiment data and the model behind the system

Glavan says he began scraping Twitter for trading signals in 2013, moving from simple keyword spotting toward natural language processing that can handle negation, and also tracks high-influence accounts and publicly disclosed insider trades, weighting a chief executive's stock purchases against their known salary to judge how meaningful the trade is. His model uses LSTM networks with peephole connections, built in Python with TensorFlow, which he found more accurate for forecasting than the alternatives he tested. He runs the system on his own physical computer rather than the cloud for security, and trades about five times a week with roughly 30-day holding periods.

Performance, risk and a costly human override

Glavan reports a roughly 93 percent historical win rate, a profit factor of 24, and a Sharpe ratio near 20 for the period discussed, figures he calls insane and not sustainable given his limited sample size and the difficulty of beating an efficient market like currencies. He does not use stop-losses, arguing they get exploited by market makers in the illiquid options he trades, relying instead on diversified position sizing. He also describes panicking during the March 2020 Corona sell-off and manually reducing his position sizes instead of following his system's signals, which he says cost him money he would otherwise have made.

UBI and equal taxation of AI and human income

Westerheide raises concern that AI concentrates wealth and power, and Glavan agrees that automation and digitalization risk mass unemployment, arguing for universal basic income guaranteeing free healthcare, education and housing so people can take risks. Citing an Oxford paper on large tech companies' low effective tax rates, Glavan and Westerheide argue capital and AI-driven earnings should be taxed the same as labor income; a live audience poll cited during the discussion found 75 percent in favor of taxing AI and human earnings equally, 8 percent wanting AI taxed higher, and 17 percent wanting it taxed lower.

Quotes

After now something like 20 years I recognized I think it's the hardest way to make easy money.

Raul Glavan

My system is now better than me, so far better.

Raul Glavan

Every entrepreneur out there needs his own UBI.

Fabian Westerheide

Questions answered in this talk

How does Raul Glavan's AI trading system decide when to buy?

According to Raul Glavan at a Rise of AI Virtual Chat in 2020, his system looks for panic or euphoria across stocks, commodities and currencies using scraped sentiment and historical trading patterns, and buys into panic and sells into euphoria based on behavioral finance.

What win rate and Sharpe ratio does Raul Glavan report for his trading system?

Raul Glavan said at a Rise of AI Virtual Chat in 2020 that his system had a roughly 93 percent historical win rate, a profit factor of 24, and a Sharpe ratio near 20 for that period, which he called unsustainable given his small sample size.

What happened when Raul Glavan overrode his own AI trading system during the Corona market panic?

According to Raul Glavan at a Rise of AI Virtual Chat in 2020, he panicked in March 2020 and manually reduced his position sizes instead of following his system's signals, which he says cost him money he would otherwise have made.

Mentioned

LSTM, TensorFlow, Python, Behavioral finance, Sentiment analysis, Universal basic income, Sharpe ratio, Kelly criterion, Profit factor

Keep exploring

More from Raul Glavan

More from Rise of AI Virtual Chat 2020

More on Research & Frontier Models

Transcript

Auto-generated transcript, not edited. · 13,064 words

Show the full transcript

good morning good afternoon we are a life again today it's episode 12 and it's a special edition because it's a hybrid event what happened is during the briefing call I normally have with my guests we started discussing and talking talking and our technical team was listening in they will like bond that's already so exciting so we decided like hey don't you want to come over to Berlin we do the sly because we have much to speak about so we are using swap car today and there are a couple of features first you can connect with everyone yes you just can chat with them you can do meetings second is where the poles section so please answer the poles I will read them out later again and third is you can ask questions I will seem on a separate screen the questions another thing is and some people watch this life if your life your advantages you can't interact with us some people watch the replay we'll publish this on YouTube or here on swap card

and others just listen to the audio via podcast later um first disclaimer because we speak about trading today I have two official tell you or I would will do this trading is always on your own risk you will not receive any trading advice here today and second past successors never count for the future so when someone was very successful and if in the past doesn't mean he will be successful in the future just as the disclaimer here so who's our guest today well it's got will I will describe them in three bullet points near chess and Commodore 64 and that's his brief story so he was a nerd in school I think they did a lot of mobbing on him and I know this case I think many of us may be nerds in schools by the way just meanwhile the game of coins today it is my topic and he got his crews computer and started building working with it and that's how he got into computers and machines etc second he is ambassador

of recipe I and the story behind there is how didn't know about well there was one person who was always buying the first ticket and later told me he even brought a stripped scrapping our website to see when we published tickets so he would be the first one to buy a ticket automated I was like man that's a dedication that's a passion let's talk to him and then when I could talk to him and I talked to him and then we figured out hey he's great he's like a skis you know crazy um he became an ambassador for ice I began second then he became pure World Cup champion for its trading so I've no clue but this seems to be like a one-year program where you know you compete against other traders with real money and if you win this you receive I think a golden egg or something nevertheless he is currently number one in the world championship we will talk about this later and third is what I like about he would

build in AI after himself so I call it an arrow AGI he is building the eye which is copy of his trading skills so Linda AJ is like but in there what really really focus um and saying that oh yes I forgot as you introduced yourself so use the chat to publish where you are with your base you can share your LinkedIn profile just say hello to everyone so everyone has a moment and my guess is already here he's pacing back and forth so I would say this is the moment and I will please come come the screen Thank You favi on for the invitation I'm very happy to be here with you today yes we will have to practice I'll look at you but I also will look for the camera so you're all involved you're all our imaginary audience out there please I hope you have you have applauded when he entered the screen um let me say a couple of more words about what rules who you have a better picture first

what was interesting as he started to study and he has a meteor copy true from baba vs bavarian okay so he said he called the Bavarian happy to either push lucky tour but you know this is excused he didn't went to one of the top business schools just saying this because he's a smart guy but he didn't went to the top business school he went to number I didn't even know that they have a university at number Tokyo yeah but interestingly his professor figured out his skills and hired him right away in the second semester became his mentor they started a consultancy together in the area of I would say digitalization in the financial industry and huh being a student already went to the top levels of all global banks are told them how internet works as a student and how to have build a website etc he told me confidential who he already advised and it's impressive list one example he then you know kind of left his job and as students started to

be a free consultant earning a thousand five hundred years birthday I mean honestly that's pretty good money being a student and he didn't stop there he built by the way one of the first robot buyers are out there and you can google it it's money filter poor but de money filter and still out there so he he invented Robo advisor before the term was already there so that's his area anyways he coming from um he won the iron trader for Germany so it was a competition when competing as two-and-a-half thousands traders and within six I think two months he made over six percent virtual on his account so that gave him the block of PR who will come to this and the keyword out there's behavior finance yes I mean so that's his old strategy um and here's a story I mean he's joyed I think to startups slash headphones which fail he worked for large corporations he did an asset management for a couple of billions he worked for smaller banks I mean he

did in the financial industry everything and he's just thirties so I like this and it's been a four-second just look at him look at his pants and his outfit he's a financial guy and doesn't look like one I think that's why I like him that's why he's in the show today um let's start with with your you know past or a weight from so first your your kid from the rural areas of connect joking yeah yes yeah me too by the way um how did when did you start trading like when did you start you to invest money while others are out there and shrieking itself you know and you started how did you think about oh I can use the computer to make more money yeah yeah first the very first at the beginning you know also doing the disclaiming posit on going so I just just always in mind that at the end trading and especially trading with futures and Forex involve substantial risk of loss and it's not suitable for everyone and

also people also in mind past performance is not necessary indicator for future returns so okay and I'm not a financial advisor be another financial advising show so please keep in mind you're doing everything on your own behalf and yeah this is just a briefly disclaimer actually you have to okay that was the disclaimer so now the question is how did you get into this well yes it started at a very young age I was su SS you said at the beginning I was at this time I was starting and no one has this term nerd but I was one of these kids who was in a chess club I was in the Chess Club I was six years old so very young after the kindergarten in the first class I was in the Chess Club Zoho and the average in his chest that was fifty years or something in beige and I read a lot of books and I was very into astronomy and then I had something like an uncle I'm not saying it

was video it was a friend of a family but he said well okay you need something and this helps you for your life and this was one present at the Commodore 64 and at this time I cannot speak English I cannot type really because I was very young and but I have was so interested in this Commodore 64 that I wanted to learn how to do it and how to how to start a game it was easy to start a game okay so because you have just a commando like what operating system do you have on a c64 it's a good question I was too young right I'll be really honest I don't know how to figure it out to do everything of this because there was no SS today there was no mouse you have know something like an user interface we have just command lines and it was very hard physicists and just two starting game was very hard and at this time I started to learn basic it's a very very very

easy language and you can do something like go to line and how is your day yes no and yeah this was the first thing but my parents just just said well okay he has fun with the Commodore 64 and then I got the 368 and 468 and then the first windows 3.11 briefing user interface and this was very cool because ok far enough I'll make it shorter how did you start training okay hello start reading well I just wanted to figure out an easy way to make money and after now something like 20 years I recognized I think it's the hardest way to make easy money so to be honest it's not easy it's really hard and I wouldn't recommend anyone to do this but at the end I thought it was newspapers and I just cooked at one stock and next day if the stock was higher than the day before I said well cool I make money and that it run to my mother cool I made money and yeah that's cool and

what what to buy next and this is a very physical buyer just imagine you're imaginary just email I was 14 years old fortunately yeah but I thought wow this is easy you just buy something and next day you can sell it and it's easy money and know the basic rules of businesses by them self I pay late collect early I think this isn't benefits future options yeah yeah it was just my imagination but I started then trading the first systems I booked my first system when I was 16 years old and participating on the first competitions and the bank was available I think it was something competition and the cool thing about is I won this competition regularly with my first trend-following sister I was 16 years old I think it was the problem because I have not aged a little age for when you used your computer right away yes from the beginning so for me I started I mean my parents always had stocks and my grandfather was in a bank so I

already had this feeling and one was born I got like a portfolio of stocks which I didn't I didn't have access to but I could always follow them for a couple of years and I started a good stroke competition with spa hustle and I won I was like top 100 in Germany at the end so I want there the national regional thing but where traded neutral with stocks yes but I didn't have a system I just did it by gut feeling well this is why what I said before I was starting so so all of the programming then it was just naturally to say okay let's make a system for kids so you get a trend system a Tran system at this time I was trained following and I just got laughs so my system was not good missus it was just crap okay but yeah I was just lucky it is fixed and because I got a lot of presents and I won all the time and I always get next delivery and my

parents what is this oh this isn't PDA and what is this this is gold and what is this this is money and why you are getting is because I've been in waltz competition in the internet and with this money yeah it was crazy at this time because it was really it was 20 years ago 15 years ago something yeah so and this was just very unique oh you're always getting these packages with cool stuff inside and I sold it on eBay it was the first email and we sold it and if this money has started my first account so when I was 18 then I was straight over to be money based on this woman's I made from the competitions okay so you started dating your first trading I guess and then you continue to use your cell phone system always to enhance this and then we jump into you then you being students yeah I said the new him back yeah I didn't know that they have a business and then you you're with

your professor you started a company because he saw your talent and he did the sales and you were just telling the wunderkind he used to say oh that's the genius was building a system I assume well I think he really really in the early stage of my Wolfman he is have seen some talent in programming in everything about internet and also about capital markets and he was said well okay why we don't have an F in Makka Saudi company and helping banks with digitalization and at the end is company was extremely successful and and we made a very very great job for his banks the only thing to be honest I was at a straight years old or 19 or something when I've met the first people on sea level sometimes or or very high management levels in banks and my sister and this was WOW okay it was a crazy situation but at the end the things we have done we're a great all the clients Wow what you have done this time is

unbelievable and so it was a very cool experience and it helped me to get more confident can you show me what you sold them was it like building a website or how to digital processes or really like investment strategies how to do this ah it could be everything no it's just the the it's fine you just don't you excited you can talk a little bit quieter or cuz the microphone is very sensitive okay okay okay okay okay think nobody we had a very large portfolio of different projects so it was sometimes for for internet it was how to let's say optimize your channel of sales but it can also be optimization process in the bank but it can also be let's say a very very specific about trading how to help traders to get better because some traders have to do risk to lose their money very fast and how to protect them how do you educate them for example when you are at a broker how to get them to a metal lever that

they are not dying because they they are dying faster they don't make a venue so I I jump now to 2011 when you published your own robot visor still a trending topic today it's called as I said money filter and then you want the German version of the island trader because it's a literature competition but you want the German one and then you compete it international you trade it I looked up the numbers it's like six hundred percent virtual and two months so pretty impressive and but most interesting is the PR you got you said like you've invited all over the world and still being a student right yeah that's correct but just to say money for this mock my think go system the company of my professor just to make it clear but as you said in iron trader yeah it is crazy situation because there was a show and this show has two parts in one reality part and one where a lot because at this time I was a consultant for banks

i builded my own consultants I was also working for the University and traded and so I had no time for the reality show but I because I have not the time to be there was this time sleeping was very very rare and so I made a bit robot and this is correct and I won this virtual part I was invited to Frankfurt and one this also and the thing is the people asked me how you do it how was it possible to generate this return in this short amount of time and now I want to know this too how did you do it yeah to be honest in this competitions you are using a lot of leverage and a lot of risk so it was to be honest a lot of lot awesome but I was using and this is what I'm using also today and all my my academic work and research papers I made was about behavioral finance so I'm very much into how people behave in capital markets because people have some

biases they have anchoring for example they are overconfident there - Hall bias and so on it's on and despises the most people are just discussing about these biases on a very very academic level but I was more how to use these biases of other people so give us an example example is just just say there are two extreme values one is panic and one is extremer forward and if it's possible for you to track these levels of a forum or panic then you can make on a very very easy way on trading system just scanning for a Fourier and just scanning for panic and if there is panic you have to buy and this is very tough because to grab the phone yeah right yes and this is hard because it's against your human nature yeah it's against your instinct because you want to go away of the situation this is what the most but already robot said this like you you buy when everyone is selling I mean your mother's saying this - like

like you you have to buy in the panic i I I purchased most famous stores when everything was going down like you just never know when the red point is so against your instincts you buy when everyone like how do you measure it how do you know there is a panic and how do you know it's the right panic for the right stock that's a good question the coolest thing is with the rise of the internet we are getting more and more data and this is where you can use let's say crawlers for collecting this data and people are communicating let's say on Twitter on boards and so on and so on and this data just just for example they are also surveys day there are a lot of data you can use but for me it was hard to collect this data would you analyze it and based on this data I know we're in the market on a very specific point where is in a panic to go in or where is the

situation I have to go short so take a different position when the market is two or thorak and that the question is is this happening all the time or is it this happening every couple of years from this issue to crisis do we have panic often yeah we have we have it off because the most the most people are just looking on very very big equity indices like S&P or Ducks okay and when you're looking on this okay the problem is you have just okay to probably saw the problem I'm happy if there is no crisis in this world but at the end you have all five or ten years one big crisis and you have some minor cycles of crisis but at the end you just get one signal a year a very good signal and what I am doing and this is something I think very unique is I'm scanning all asset classes I'm scanning commodity markets so commodity markets means resources slice a gold silver bar a food everything garlic etc so

old which is I think comedy America's huge it's like bigger than everything like it's I think it's bigger than stops or it depends I'm I'm not sure because there are some parts of let's say gold it's extremely extremely high but stocks I would say old yeah so there's a market in commodities are very large okay so you look you look at commodity market yes second this then then the next thing is I was in the currency markets because it is for Australia this is forex trading and the coolest thing is about they are uncorrelated to each other they have some correlations - but the most important thing about is you can generate much more signals when you are in all asset classes and and this helps you to not to have one singer a year then you can have 100 signals if you are scanning so let's assume you have a trend building system or like a signal spotting system at the beginning because not ESO training that's also it's just giving you on your

dashboard hints like hey there's something happening so let's take the example of wire card because I move use the small fry a Kurtz panic I mean there was a foyer yeah and then there was panic but it crashed to zero so if you buy it when there's panic you still lose money how can you buy something falling down and expecting that it will rise again yeah 100 percent correct um to be honest it's very hard on a single stock and we already I will switch to this here there because now I think we are not speaking about money we skipped a little bit of your CV that's fine we are now right in the topic yes it is very hard as you said I was I would not recommend everyone to to go in this single stock area like you said because there is an risk of bankruptcy of single stocks you have this risk not in this case for currency and commodity come what is not tend to be zero okay there is VTE

and you got negative Fred but at the end it is very uncommon but they are just like the Google lira like the like the Turkish mod didn't they go down and they stay down for really one thing I because I have stopped Russian stocks and they are done for the last six years because of currency I mean it crashed and never came up yeah I think this is the very hard thing about this you have to look out for very very extreme points and just what I'm doing I'm not too long into a situation like this so my average for our trade is from this ratio side is just a 30 day something like that because I'm not too long into this as you said if you are too long into Russian ruble then it can be very very painful so you have also go to the right horizon yeah so um I will continue with this so you said you look at Forex commodities and stock and because you have so much bandwidth you

have up to a hundred signals a day your citizens spotting the signals and then it's even a buy or sell signal so you do it and you go after mainly after panic or for you these are the main signal to use these are the mainstay from the behavioral finance circles behaviour friends and my work means you look at human behavior which is irrational and you try to make money at the end it's money on the wrong emotional behavior of humans in a normally rational market exit exit I'm something like corrective for the market because everyone is selling someone has to buy and I'm a guy who buys so it but the problem is it's against our instinctive it's very very hard from an emotional side so if you want to do something like this because it's very hard you usually have to do it automated so because it's too hard to do it as a human yeah so this leads me to we already have two live questions oh we have questions Brum Brum from

john piper one from attila and both are similar and Atilla asked what is the best place to scrap the sentiment of your in your experience likes when you scrap it and the other one is with Twitter accounts does your I'll go scrape or how many followers do they need to have so a little bit let's talk about your sources how do you get the feeling like this sentiment like this this purchasing signal yes okay that there are very very good questions and that sources do you use yes well we can stay in Twitter because I think I'm a bit famous about Twitter it was 2013 and one conference and no one had strategy at this time because fit was very young as a source and I was building the first or one of the first crawlers for for Twitter messages and it was not a good system but it will work for me and in this time I just tried to get them sentiment based on the Twitter message so I'm searching for filters so

for example the most are with tickers so you know what stock they are talking about and after you you have to sticker then you are searching for some keywords because good grade want to buy so this is a positive sentiment and of the other side you have these negative words and this is a very simple approach today we are about far more it's a keyword spotting it's it's getting about any natural language processing yet what's cupid spotting yeah I was keep on spotting and day we are we are far better because today we can also use negotiation so if someone says I don't like added a stock and for me it was before it was oh like editors it's wrong okay today we can work with this and some saying you can also getting other informations like ironic stuff I mean someone will say something in Iran akin Twitter but I cannot do it but some say that they can do so how do you do you screen whole Twitter or do they have pre-selected

everything which is related to try adding a financial industry yeah um well not whole Twitter because the the volume is that far too high it is not possible when I was starting it in 2013 it was possible so because it was not that that not that because today and today is no chance for me because I am a one-man show but to be honest there are two things I would recommend for but if you want to do it there are two things first you build a sentiment on the broad market for example for the SPE for gold for rates for okay so this is one thing and you're looking as you said like or your baseline a little bit so this is very bad yeah this is my best I'm searching for outliers I'm searching just for our clients and this is something very very very useful but and this is a very good question from from the other guy there are also special accounts in Twitter with extremely high influence and they have to

be covered on the other way so they are there some training searches you can build on them one trading Sergey I was using at this time was searching for accounts of high influence and if this account for example promotes one stock and you have to buy whatever this kind of a stock in Canada and there are more most of the time penny stocks they are reacting very very hard and then there was assessed by the way with quickly this is just a movement that this moment and I don't want to buy the stock this is not what I wanted but the cool thing is when someone like this for something new yes just have our Jim Cramer and if Jim Cramer says ok this stock is cool you have to buy it then it skyrocketing something like half an hour and then it after half an hour it gets in a correction so if you are fast enough and you are automating you can go on the long side so you are playing the first

rush and also I mean you're very good you know yourself yeah and then you know it's your system doing this it wasn't it yes it was doing this now it's it's something like this this situation I'm searching for people in Twitter or in other sources who have alpha okay I mean please what alpha alpha is risk adjusted over return so and some people in this world generated some systems or they have a very good a good feeling or whatever and they they post their trades and you have to find these guys and you can follow them so it's something like collecting the Alpha from the internet collecting the the wisdom from the best of the best and this is very interesting strategy you can use them you follow them or do you have an not disclosed account which is following them and after not disclosed account for it okay so you can't just follow the rules followers or fans but other way back account when we are in between because we have from Max Petrus

and John another question Wow okay first um do you have your old NLP model for the sentiment analysis or do you and do you have the back test to your strategy second what are other financial sources versus do you have like libraries api's you're tapping into and the third is would you be able to go as far as coding and I'll go that rechart annotations or Twitter stream or with be impossible so three questions a little bit together well okay first do you have your own a tepee model or do you use libraries existing forever no no no I'm just using libraries and to be honest it is not the most important thing for for my my situation and I'm one person I am NOT a big company with 100 people so if there are libraries for example I use also the back-testing library they are here and I'm using a lot of api's for data for the basic data and you can use quando you can use yellow finance if you have nothing else

or you can use Tingo or so there there are a lot of resources and they are for free but if you want to do it more professional level you have to do by like Bloomberg well yes of course the blue megasis is very very expensive oh so do you do pay yourself for resources for data well yes I'm paying for data but I'm not paying for for Bloomberg just because we make this is very expensive so yeah but I assume so you pay for data I would say for label data yes because it's better you get this label data because if I have to classify all this data it's very very hard so I just want to be more pragmatic as I was before as a child I was always this guy who want to be perfect and code everything on my own yes and it's just stupid so because it's because it's not it's not the main thing I'm doing the main thing it's not I'm I'm buying I'm building libraries for doing the

best and opie or doing the best back testing I'm using this I think my my part or my job is to build all these things together that they are working and they are on a pragmatic side I can manage us in one person ok so I tell us for example asking do you scrap several days at several times per day or just once how fast is your system at the end I am trying to improve this because I've seen that information efficiency is rising just because the speed of information is so much faster than then was one year ago and 10 years ago so I try to be better but at the end some informations they have not that much data points so you cannot have this information often then then once a day or sometimes once a week so I was using some economical indicators and they are there just once a month so I cannot push it to seconds so depending on the source and the whole updated to this you scrap several

like very fast you know very slow yeah I want to speak a little bit more about the system so I understand you have libraries do you have your frameworks you you get the data in it does your system run in the cloud or is it on a physical computer on your on your dorm room it's on my physical thing we don't know it's just I don't know how to save a bit I don't know where he's living I mean just doing kohonen he was living ex-girlfriend on on the countryside somewhere random yeah it was just a situation at the end to be honest I have to be paranoid a bit because then when you're building holes like this and you're doing this I'm always very protective about source code yes and I don't trust at the end of this level clouds so I don't need it to be honest because I have not that extremely high problem with big data so because some some data I'm using they are not that often I'm not not

on high-frequency traders or something like that at the end I'm doing on Macs five trades a week so it's not okay in those high frequency and five trades a week how long do you hold a position in approach something like 30 so you make five trades per week 30 days but you have exceptional returns with the strategy yes and this comes to a question by Slean honestly I hope I pronounced it right to your life yeah and I should ask you about the trades you mate or can you ask all to talk about the trade he made that resulted in losses based upon the strategies he used for his winning trades okay he wants that and exposing losing trades well there there is Forex I just I've heard that it's not about winning all the time it's just winning more than you lose so you lose money I see women explain it about a wedding you lose money yeah and what you learned from it and feed it into the system yeah it is I

just posted also my my metrics from my account what is life tracked and every day is checked by the worker chief of training and in this account I have 93 percent roughly are winning trades and also seven percent are also losing trades so keep in mind I had also losing trades and this is interesting because I looked a lot of the hedge funds a lot of hedge funds started with the eigth so they came to me a little many many AI driven hedge funds and they were just like a we had 55 percent winning yes and there was a very good because they were beating the market by at one two points yeah 93 percent winning trades historical and seven percent you're losing yes and that is enough to make a surplus that's enough because I am running from profit factor of 24 so just to explain the profit factor every euro I'm putting in I get 24 euros out of the system so the profitably agreed again so we assume you have 1,000 years

you put into the system in 93 percent you get money back except percent you lose money but do you get each trade 20 trades the profit factor also includes my losing and what on a day or a week no on the whole time period so on the time horizon you're I'm doing like 30 days per trade I have a separate trade per trade yeah I I I like it yeah it's no no no no no no this comes back to you because you are beating the market here yeah and I would say again it's risk of my it's risk he's one-man show so he's speeding the market with his own money and sometimes virtual and the past and that's good you're using money he also trade it by the way for large banks and corporations did you use your own system or did you use their and just optimized it well to be honest I'm I'm doing training systems now roughly where roughly four for 20 years when I was a teenager I started it

and with AI I'm developing since four years five years I tried it before but to be honest it doesn't work out because ten years ago there was no good libraries there was no good things you can do it okay and before I I think it's just four or five years and well at the end it's just the cool thing is I have to not to called everything on on our rhythm because I've written is also learning and I'm a lazy lazy person okay and when you're a lazy person machine learning is better for you because it learns for itself and I don't have to learn all the time and watch the markets and adjust the systems and this is where started and this is what I'm doing now I have not done for anyone before I always wanted to be protective and this is what I'm just doing for myself in the comments already read from Reiner that he says seven percent losses that's a crap system you should be president which I think if

you want to be present in United States I think that's not a positive but I get it so I would say you're very risk-averse or yeah embarrass I mean you make very very few trades yeah second question there's can i scale so could i put more money into the trades and still have the same output or is it only working because you have very little money depending on the size of this we like penny stocks if you buy too much you are influencing yourself to be honest and this is the craziest thing about but I have to to get this more in a context this is just a result of this year and also this year the sample size is not that high because to be honest I'm running far far above my expectations and this is not my long-term winning Greg at this markets it is not possible for me to do this on a long term I was just kissed by variance and I will be in the next month I don't know

but I think this is not sustainable but as you said yeah it is scalable it is extremely scalable just because I'm trading the car market and the currency market we are talking about something like 8 trillion revenge every day so this is not a problem to to use billions in one trade you can do it it depends on how what what is your currency pair all we ask Allah is working US dollar yen is working but of course if you are in the Swedish Krona it's not so this competition thus for its trading more championship it's they tap into your portfolio they see it so it's real every day they measure it and it's only your money yes do you leverage or is it a hundred percent equity no I'm also using memory using leverage using options swaps called puts like all these fancy terms you should either know if you want to use it or if you don't know don't put your money in it yeah well I'm leveraging it and these things in

in Forex it the most of this stuff is OTC over-the-counter so you are I'm directional in this competition because options are aren't it is a cloud and so I'm directional and you are leveraging it with the money from the bank so the bank or the broker is providing you leverage and I'm using not high leverage for Forex because Forex sometimes on the Vita side they are leveraging up to 100 so 100 times your equity but this is just crazy you can imagine 1% in the direct in the wrong direction and you are dead and my average you ever lost like painfully much yeah of course I I mean I'm trading since I'm 18 years old and I had also big losses and was not cool because if you're losing money and in a casino I think it's a bit cool and a cool atmosphere and not for me I hate yeah yeah it losing yeah I also hate but I think everything is cooler to lose money but if you if you tuxedo and it's

shining your liquor yeah then then it's cooler then you're sitting at a screen and you wake up and something is happening in China or or Kimmie alone and so on and you see what the yeah and and this is something it's not cool I think it's the uncoolest way to lose money because no one sees no one is applauding you're just losing money and of course I was so let's jump back to you AI so today at the beginning it was earnest at a trend predicting signal it was giving you signals and you were trading this was at my teenage time but but now I'm on the other side I'm mean reverting so this mean reverting so I'm searching for panic or or foria right this is what I'm using I think the most of time tranfering was just at the beginning okay and you're looking for the six but today you you a is trading for yourself yes first question I know it's but is it legal that AI is doing all this work

it's legal it's legal so did you know no regulation no protection for human rights no one's going to the street and say hey I is taking my job they I took a lot of jobs out there you know we all know how the stock used to do you know yelling and selling and buying today it's all about AI and for me so the u.s. system is mimicking your baby er yes so you you build like I call it a copier for yourself yeah the thing is um I trained the system very much how I have traded so what I will start I've done manually or later with trading systems I just trained on these times years I'm I'm looking at so I think that most people are doing is wrong they are just going on all data so they have extremely extremely large sample sizes of different data sets but it doesn't make sense so this is what I see a lot a lot of in articles I've seen also on github and I'm looking

at Ossining what are they doing for example they're using technical indicators and these technical indicators are based on the time series of the price but this is also in so this is implied and you don't have to do edit again it's just stupid so I just put all this alternative data and this alternative data has some hidden information and this hidden information is not priced into the markets and we said what you are looking what you can go as we as we talked before on Twitter you can go on insider trades for example what insider is buying what stock and if you want to do it really properly now now if you want to get one time series of me I can show you this but it's very complex because the most people say okay let's jump it back to the CEO of wire card he purchase stocks before everything went bust so that were like it mostly said inside a trade because he didn't announce it and public something like this at first how

do you recognize an insider trade well you get databases in the US they are public public and in Germany they're also public so you can get the state of Isis and as I said for me it's not important to get it in the first second so it's the seat is not that crucial for me not for me no I am NOT in the high-frequency a sector because their competition is very high and I don't want to compete with these guys then so if you want to do it really professional then you are not looking on what what's your board and itself so you have to go for example what size is he buying so this is very interesting it's assassin CEO of a big company and who has a salary and just you know the salary because they are all public you have to grab the salary of the CEO empty amount how much he is investing and this gives you the signal of how confident he is buying and if someone is earning tenure

is doing this today yes so you have trained in the eye to get all these sources yeah and you don't even have to double-check it anymore it's doing a lot sold then I have to double check it because it's always a big problem but my the e is using the trade without asking for your permission or no no I get a signal so I'm doing this I'm sure you still execute it as a trade yes I am executing all trades I am as a human doing all / - correct maybe it's not any eye it's a nipple is closed it's in Hebrew day I but at the end it's very very tough to get all this information on this space you can use them because as I said before someone who is earning 10 million and buying for 1000 stocks of the own company the signal is very weak and then you have a very stupid system he buys it because old it's so - so your system is a lot of learnings that how

you wait all this information later then sort them and then your system and then let's think about the impact of your system you told me before you were able to monitor it 100's of classes let's say stops comedies etc that's a lot but you you monitored them and your system was telling you where signal is but that's when you started using way Isis self-improving this scale to F over thousand asset classes and he told me with even a better recommendation so when the system tells you to buy something and you do it it's better than before when you did it on your own without the AI that's pretty good because my system is now better than me so far better and I made this year some mistakes because I was overruling my own system in March I got panic like every one of Corona and I think we also had communication at this time yeah spoken you were everything is ready was going there it was a tough situation and and then I just reduced

my position sizing extremely because I don't want to go broke at the situation because no one knows exactly what was happening between Corona at this time and much and I reduced position so but now it was now it's always easy to talk about the past but you got did you loosen so what's your human you're human decision to reduce the position that amount of money you have invested the right or the wrong one in Contras so you shouldn't have done and you should have followed your system and it would even have made more money yes but I have not done it but you were human yeah and you failed I the problem is at the end I think I'm not a talented trader I don't know if they are to lend the trader because it's against our human nature to trade interesting but yeah I lost money I left money on the table because I was reducing the size of of my trades okay so John asked another questions which is fitted there and it's

another screen no no I will do it it was from Chris this one crucible and I do you use Lauren networks yes in what case like or what forms of machine learning to use and what language you used for example fight item okay I'm using Python of course I think it is the most universal language you have to use I would recommend it to you if you want to start with this I'm using primary at the moment Enzo flow and I have some older systems in Cadiz these are the libraries for machine learning I have all the other libraries you need like pandas numpy and so on and of course sometimes you have to a TI library so this is also necessary and machine the machine learning field I'm using tensorflow primary with LST MS so because they are mid to what would be happening yeah they're the best for time series and casting so I have to test a lot of things and what what was the best for forecasting if the highest accuracy

and for me it was LS CMS and especially if you want really technical me but but I don't know if we are going to do too much detail I'm also using the LST I'm not in a basic version I have the most of the LST MSR with peephole connections so just because they are better when you have rare but extreme outliers and capital markets have rare and extreme outliers so but when you can also use Alice I'm in the basic architecture at the end this is not not the most important thing yeah I think the most important thing is to have you need to do main knowledge and you need the data and you have a compiled it this is extremely important this is the most important thing so I don't think that this is a big difference if you say okay Carol attends a flaw tens of row is faster for me it's better and and I think is the best framework at the moment on the market so use tensorflow but at the

end it's more about today so what I like about it all is you don't sell the software you don't work for another a bank account you do it on your own with your own cash in you learn but your options aren't you just increase this but I also know from our past works you could have made a fund out of this it would be by the way hedge fund I think that regulation so you have a lot of options out there by the way if you want to work with Raul we might do it wavy now one day will doesn't train this in detail we figure this out so if you interested to learn from him we might figure it out a webinar where we will go through all his learnings and maybe help you to build your own system this is one option but yes it's all about money at the end and then we will see maybe in a year or two years where we'll be in a zone setup I want to

jump touch another third topic and first talk about you then about your system now we want to speak a little bit about your passion which you called ubi unconditional basic income so we will touch this if you have questions the meanwhile I will collect them and he might even over an extent so let's say you have to go after one hour you can work quickly any time he wants walk on because as a youtube link we've also to cut and edit the video and we'll publish this on our social media channels in our playlist so you can share watch this whenever you want to and you can contact on LinkedIn Instagram all the tic-tock private about every day so let's talk a little bit about the bigger picture because I have following problem which I see with a guy AI leads to a combination of power and wealth and if we don't have in regulation system which is distributing this you only have guys like this we spoke about in the beginning he might be

the first trillion year just for trading he could his vision his ideas but if you want to have less people like this and more you know sharing the wealth because if you're smarter with the eye you can excel everyone we need something so you say ubi unconditional basic income could be one way to distribute a little bit a little bit of the surplus to society tell us yes tell us you what scientists yeah well if you just look at the numbers you can see there is even more and concentration on income and the concentration on wealth is far higher than from income and digitalization and optimization is for me leading to society in in in the next years with mass unemployment so for me it is very very important to to get this topic into politics to get this topic as it's a number one topic just because we are going this year and I think Corona was something like an catalyst for a lot of cut offs and layoffs in the workforce not not

in Germany especially but globally and automation and also tequilas ation will lead to a very very bad situation for many many people in the workforce and we have to mitigate this and we have to approach this this is very very important for our society and this is what I am what I am fighting for what I am really into it what I'm committed to use AI for good and this is also what I why I have done this speech because I hope that someone out there is imitating me and building a non-systemic building its own universal basic income it would be cool in one year this is your ubi yeah I said every enterpreneur out there needs his own ubi you can be rich parents it can be you made a fortune before you living on this you're building your system we all have a certain income source which is it can be real estate etc which is giving you minimum guarantees so you are free in your mind you can experiment you can go

all then you take risks yes so my vision of ubi is just to take I am little bit like Sanders there everyone should have freewill health care free education and free housing and a certain degree so I don't want that everyone is afraid of being homeless and I think if you're losing your fear you can taking risks you can be creative and you can take a job you know just learn from it or you can do an internship etc so first you say the first step is you build your own ubi but that's your exceptional I mean you wear red pants yeah you're breaking your edgy and I would say most of the people today here yes your guys are think you unlike us you're edgy you're creative you might be brilliant in your area but not everyone out there so the question is how can we make an ubi for everyone how can we help those who are not smart like this well there was I think the last month was a cool paper

from Oxford it's about pinfall Claus men there are profits from big tech using for example aí and aí therefore gets extremely extremely high the best case is epic you Apple is the highest tax payer in the United States and they I think I look today they they had two hundred and thirty billion in cash outside of the United States the trips were they paid an average ten percent of this and the F Moses problem big tech is for example in Europe not really paying the fair share because when you are a restaurant owner in Germany you have to pay your tax and you are paying a fair salary and all this stuff but when you are too big and you are a tech you are not paying and I think this would be one of the first things we have to arrange on an European level they have to pay tax okay this is the starting point okay this is just a starting point first when we talk about AI we need to say i

regulation to Texas I mean by the way for a data export X so if that say Facebook is using my data to sell advertising either the earnings are text which is not that easy or at least the data which is used might be text so something like this and by the way everyone asking questions out there we will collect them and I will ask I will ask a little bit later because they're technical so we need taxation but still ubi I mean this is your traffic so taxation fear I'm fine how do you lead this to ubi because my problem is most parties thinking you have the right to work and that's your the goal of life is to have a job I would say no the goal of life is to be human to live freely and if you need a drop to have more money or if you have a job because it's your passion you enjoy it that's something else but why do we have to have a basic right to work

because we were eight comments up there and you know we will have new jobs created it's not about lot of drops and there won't be new ones it's about why don't we work from society way I is doing the jobs and the work and we have the option to work but not the need let's say I built this super AI which is training and then the AI is all by by the pension funds and the pension funds is just using all the surplus to pay retirement money which we have in this kind of idea but I would say a pension for retirement fund it's great because it is just like conditional income for all the retirees out there although all people benefit from a better trading system absolutely so I think we have to do it more on a governmental level that we are going to get so close because of the productivity and we can we can look at the data and set up for all Western countries shows us that the productivity is

rising but the problem is just the share for the employees is is declining in the last 10 20 years there is declining day there are a lot of effects you can see in the US you can see in a lot of different countries that at the end the owner of capital getting every year more of this of this tank and this is not fair so and within ubi we can make it more equally let's see I don't see the logic there but I will come to this so from my own experience I have a venture capital firm I've investment firms but also hire myself my own company if I pay myself a celery celery I pay up to 60% texas total in germany if i sell shares I pay 1.5 percent taxes different level but what I mean is capital gains little text labor is text a lot so I have a poll there and if you have not answered the poll yet you can do it for those being live here and the question

is what should be text like should AI be text and similarly and well eight percent say said hey I should be text higher than humans 17 percent said AI should be text less than humans but the majority seventy five percent said no earnings from AI and humans should be taxed the same and honestly that was my answer to I believe we need to have a system where it does not make a difference because what we currently have is wittman system which gives us a huge incentivized to let people go and use AI and optimization to have an increase because they I suppose is text less but I think that's a wrong incentive system if it would be equal people would fire people just for productivity but but just for for text productivity but we will have to figure this out so we are official coming to the end and but I would like to come back choose I want to questions which we still have and I think this was um do you use stop

losses yes which is from John if yes is it a physical or mental stop-loss if not is your trade-in invalidation point or how do you decide a trade has gone wrong so yeah okay that's cool I'm not using stop-loss it's just based the most of my my accounts are running on options and in options you have the problem the markets are very illiquid so if you have stop-loss or if you are orders like this you get ripped up from the market makers so I don't want to have the situation and I'm I'm more from the risk side doing the the BET sizing in a very adequate way so if a position is running against me it's not not a very big heart because I have a diversified portfolio of positions and the size of every bet in this case is not destroying my my whole system so I would not recommend to use stop-loss especially when I'm using options it doesn't make sense I have to learn a lot I use a lot of stop

loss but I only trade stocks and I do it like long and if the crash at least I what I'm do it why a car best case if it's losing get out because humans not realized losses like I already lost fifty percent so I lose another fifty percent it's interesting and another question which we have there and they've got three likes is have you tried attention networks in comparison to LS TMS well I have tried a lot of different networks and architectures but to be honest at my point where I'm working now Alice TMS was the best network just just for my cases and there was also this question about how to use to the classifying models for example in charts and so on and on yeah you can do a lot of a lot of cool things and a lot of stuff but as I told you before my problem is just I'm I'm limited in time so for me it was a pragmatic way and so I'm stick to Dallas TMS and I'm

also as I said I'm using in trying other systems and architectures for for better results so you use the title you wanted this how to make money while you sleep I know that you first I think you don't make money when you sleep because if your system is giving you a buy or sell signal and you sleep you don't need money what you see yeah that's correct so at the end you make an active choice and second is what can others learn from this I mean we know we have people here from Deutsche Bank and we are from Softbank we have a couple of private equity funds from the offices but also technically put built there how can they benefit from it outside of hiring you okay um well to be honest do we do it I never take the most important thing is the data search for data what is not not like pricing data volume data and so on search for alternative data this is I think the most important thing don't go

in to large sample sizes that don't go to long ago back in the history I would not recommend I have seen some system designers who are using data from 1985 this is Bush okay the the market structure has changed a lot since 1985 and we have a whole more you have more faster if the information efficiency and so on it's on so don't use that old data I would recommend use data for example for 10 or 15 years 15 years is a bit better if you go back like this because you have to financial crisis in your data set and this is not bad to train with this data so 15 years is good but sometimes you don't have some alternative data is all along so it would be a bit hard for you to do it in this way the next thing is don't and this is this I have seen so often and this is so wrong the people say if you're building one system and it has to work for copper futures

can also for the Euro stocks and also for let's say a US dollar that's okay every every asset class has some very own characteristic and has own behavior so if you want to trade currency you have to build a system for currency and even when you want to really go into detail I want to make it really really good for me every every currency pair I'm trading as an own system so because we I have a different data and this is something you have to keep in mind you build the system for very specific asset and I don't build one system for all assets because it's not it's not working good so it's just next question and how often do you check your portfolio several times a day daily basis push notification new systems like closed ecosystem communicating with you yeah at the end the system variegated signals is related from my trading account so I'm getting the the signal and then I can make the trade and because I'm a bit lazy and I

don't want to be someone who is oh I missed this trade and if I'm not doing this in the next five minutes it's wrong I have the advantage using options and I'm an option seller so in this case I'm I'm not so extremely OnPoint I don't have to be there so if I get a signal I can also make the trade tomorrow sometimes of course it gets bad worse than but it's not so time critical and I'm checking my messages every day of course and sometimes every hour but there is no necessary T to do it it's just because I'm interested person and love what I'm doing okay this is of course then I'm checking it but it's not necessary at the end I can just look one one one time a week on Friday and place my bets on Friday so it could also work I think that's a very interesting system because it's opposite of flesh trading and say yes coincidence German or German asked about your money management system yes he called

it how you do stop losses etc or trading stops but at the end you're your keyboard is my money management system yeah okay ah but I'm a very risk-averse person so I said risk-averse first I'm not optimizing on maximum profit and not optimizing or for example for the highest risk adjust profit so this is very curious because normal assets doing this but I want in cash flow for me is support to get in the best case every day on income and every week and every month so my system is a more based on constant cash flow instead of running for highest profit and this is this is something different because of what you are asking is I think the best money management to get the best risk adjusted returns and just make it clear I think the best way to do is is with Kelly criteria or vocational so you should use the character turn but I would recommend don't use the karyam in the original version because it's too aggressive you cannot hold it

with character tarian sometimes you have three or four positions in your performance very very hard and one position is 60% or something like that Kelly Killoren can be very hard so practically if you want to maximize for the gains you use something like a half helical turn or 132 clearly catania something like that you have to figure out and with this you get your position sizing for the right positions but to use category you need your win ratio and so on but I like this criterion but I like this only when you are maximizing for profits and I'm not making much of a profit so I understand you have trading account one of many and your system is independent from this you manually buy it or if your system has maybe proposing it but then cash there and you take out some cash to live from this so it's not about having the highest portfolio value but to have a real return on cash and this cash you are living on yeah and the rest

you used to reinvest of course the the best thing is my my returns to see where to hi I'm running too much above my expectations so when you're building a system you get an expected value so in this respect well you long term you see okay this is my complaining is making too much money yeah this was crazy because I know that I am going to get a draw on I don't know when but I'm going to get a hard-on because I'm running extremely too high so just just to imagine at moment on my account was covered every day I'm running different Sharpe ratio of 20 okay this is completely insane this is not sustainable this is this is tell everyone else what a Sharpe ratio is and why 20 is insane yeah it's it's one of the most famous risk or let's say return to risk metric and just just to to get an feeling when you are hiring you are on the allocation let's say your insurance company family office you are whatever

pension fund and you are searching for great managers so insure pressure is one of the selection criteria and you can say something above two is exceptional this is crazy this is this is the killer okay if someone is here in the microfilm three it's okay this guy is close because he's got money and two runners twenty is just doesn't make sense that this twenty means means what you are very risky are you just very successful to successful to successful photos so that's what we speak about lying a lot of very successful trader we're not always able to do this with a couple of larger amount yeah but at the end of my mind my sample size is too small I just have several hundred trades and we've several trades or two its actions in this case my sample is just too small and to be understand I'm going back I don't know what what is my real rate so hopefully it will be something like 10 percent of what I'm running now because this is

crazy enough because I'm trying currency market and justice to to think about the currency market is the most efficient market so it's the hardest market to gain some extra time so if you have there this market and sharp edge of to you are crazy you are gone okay so yeah it's a problem I know it's like what is ten times more gold I don't know it's Broadway I like the crazy thing is someone I have worked a long time someone who I had worked a long time with me and called me just days ago and have seen an interview with me and I explained metrics and it just said to me no one will believe your numbers they are just too good and I say yeah they are too good okay last question from my side moral one you're making money and you don't care with what right at the end it's you're making money and someone lost at the end maybe the currency is going down everyone is starving and you making a fortune

well yeah it's not not support because there are people it's a in there isn't panic and these people who are panic want to sell and if if are not people like me who buy this stock or give them money then the market will collapse so I need to exploit weaknesses of other humans I'm helping them because I'm I'm buying the asset I heard this before oh yeah stop-loss is great or like buying options because hey I'm helping them to to sell yeah I am helping to sell because I'm on the other side that's very nice of you yeah that's very friendly good everyone out there um let's I will have some closing words so for you I will thank you you may now leave the stage to everyone else there if you want to get in contact with hole contact details out there we will do a full up session or maybe we do a webinar with him so thank you all I guess you clap for him vitriol thank you and this is what's

written here my screen it's called the final slide so if you want to follow us if you want to stay updated for upcoming events whatever we do first is you get an email from me regular but also check us on LinkedIn Rise of AI Twitter account insta we would do some behind the scenes and Facebook the same ah and second is we will have the next virtual chat in two weeks it's again on Thursday again at 11 a.m. it's in two weeks and it's what member of the European Parliament called damn embezzle Agra he's the founder of vault vault is pan-european is the first european party not a national one it's a European party he got elected he's there and he's working on AI data strategy so AI slash data tragedy I will invite him in the show I will grill him a little bit of what the AI wants to do with the data or what is the big master plan behind the I what is the exit will be like today I focus

but it may be more on the political society side as usual we have this mixture and that is the words from me I hope you enjoyed it I can't tell you give a thumbs up too because we have no writing system just for losses you stay tuned if you have recommendations who to invite who I should talk to just sent me a name I will hustler as long and then try to get them in the show so it was a pleasure as usual I wish you a great day enjoy your day and we hopefully see again in two weeks

Rise of AI 2027 · Berlin · May 2027

Be in the room for the next one.

Ten years of talks started as conversations in one curated room in Berlin. The next edition brings 300 selected leaders together in May 2027.