Value based AI - Conscious decisions on investments

Welcome to the unDavos Summit - A community-organized series of interactive panels, talks, and networking taking place in Davos, Switzerland - and online - in parallel to the World Economic Forum’s Annual Meeting 20-24 Jan 2025. Our mission is threefold: 
• Democratizing Davos: We open the doors to diverse voices and ideas, breaking down traditional barriers to participation.
• Humanizing Davos: We foster genuine, relationship-driven connections that go beyond transactional networking.
• Bringing Action to Davos: We turn meaningful discussions into tangible, real-world solutions.

Join us for the "Finance and Investing" session, where we delve into the rapidly changing landscape of finance in the digital age. This insightful panel discussion will provide a platform for leading economists, seasoned investors, and innovative fintech entrepreneurs as they unpack critical topics such as sustainable investment strategies, the impact of digital currencies, and the democratization of financial services. Attendees will walk away with a deeper understanding of how traditional financial models are being disrupted, along with valuable insights into future investment opportunities that prioritize both profitability and social responsibility.

Don't miss this chance to participate in thought-provoking discussions that challenge conventional wisdom and reveal how technology is reshaping the future of finance.

For unDavos 2026 Sponsorship & Partnerships:  
Exclusive collaboration opportunities: Contact Mark here: https://bit.ly/417TrB9

Or catch the full video here: https://www.youtube.com/watch?v=EM9o8tVStiUand browse all podcast episodes here: https://undavos.buzzsprout.com/

Join the Conversation:  
Love this content? Like, comment, and share.

Homepage: http://www.undavos.com  
Linkedin: https://www.linkedin.com/company/undavos/  
X.com: https://x.com/un_davos  
Podcast: https://undavos.buzzsprout.com/

And join our global community and let’s spark the change together!

Transcript

(00:01) [Music] a couple of weeks ago a friend of mine called me and he said to me Wilford why is it so difficult to find good private funds to invest in our family has now invested in like 10 of them and six have done worse than the the stock market is it a myth that there are those private funds who do three times better than the stock market is it uh is it a joke we own had like two small firsttime funds who did better than the stock market so should we actually go into this at all as a family office he was talking on
(00:40) behalf of his family um and this question actually made me think because a couple of years ago when we founded Heights the company I work for uh we saw this a lot because actually the answer is no there are a lot of good private funds you can invest in but you just should have the right data and uh the right tools and I come back to that later um first let me uh introduce myself I'm realr I'm one of the founders of heights height is a firm which collaborates with family officers to build technology bespoke tailored
(01:16) technology to optimize their investments in private funds both in terms of their returns and in a way which aligns with their purpose their their their way they want to make impact um I do together with Pascal so maybe Pascal you can quickly say who you are that's really quickly um and today we're here to talk about the topic uh conscious investing in the age of real-time AI I'm really happy to see so many panelists first of all and uh the guests so conscious uh is a word which is mentioned and that will be really Central today it refers to the
(01:55) purpose what makes you you as a family office as an investor what is the the thing which drives you where do you want to make impact and this goes beyond uh profitability it can be like uh really unique to your individual um preferences your values second it's about real time data and we're going to hear more about that and I think that's a real Innovation how can we use real-time data and AI on top of that and ultimately as we see in the later panels how can we combine those two so how can we have real-time data
(02:31) and AI in a conscious manner which aligns with your purpose as a family office that are topics we uh want to discuss today um but before we do I give the word to uh Jean Philip um with whom we organized this event together um who also share a couple of words hi everybody I'm Jean Philip you can call me JP if you want so after for the networking don't pronounce the whole name is quite long um so um together with Wilfred we organized this day um and to discuss about Investments and Investments the way you do Investments are dramatically changing
(03:11) because within here in four years 5 years we'll have a completely new generation that will take investment decisions today also more and more women are taking investment decisions than in the past so other elements are playing a role in there so elements like not only performance indicators but also other values so later in the panels we'll talk about values we'll talk about which leads to Consciousness more Consciousness and be more aware and we're going to talk about how we can do that and how technology can be an ally
(03:48) in there um I'm the founder of uh axj axj is a scaleup company global scalop company that is transforming data into signals and a cocktail of signals is predicting a company Behavior or a fund Behavior or whatever Behavior you want to predict based on that cocktail of signals and then the end user is giving a feedback and this is retraining the algorithm so you have a complete Loop so that's why it's in real time but I would say enough about us uh it's about the panel we quickly going to present the panel uh too bad for the panel we're
(04:25) going to do it to save time so uh so so that's a smart move we would say but if you have the word you can tell a bit more about yourself of course but I let Wilford start presenting the panel okay I pick a couple of panelists and JP will do a couple um there's Carla a global strategy leader at uh the boardroom uh Network for female leaders um then I skip to but we come to the other panelists um yoges yoges is investor and founder of msis msis is an IT consulting firm with a global presence in Europe Netherlands in Us in
(05:06) India um mahol director of research strategy at versus a cognitive Computing company and then the others I give to JP um Greg yes that's Greg he is based in Berlin and he's working with invest Hong Kong so invest Hong Kong is an organization ation that uh facilitates and promotes investments in Hong Kong and I can tell you Hong Kong is a really thriving community of Investments uh over there so uh I was there two months ago and I can tell you that's really vibrant over there then we have John uh John is a CEO former CEO of software
(05:56) companies and uh he's now working for Landers Innovations uh and uh he has quite a lot of experience of Investments and uh managing uh tech companies uh then we have uh Martin Martin works for a wealth company and for several single and multiple family offices and they acquired a heavy user in Practical terms of AI we have uh Alo Alexander Alejandro and uh he's based in New York and he used to be group CEO of sdg one of the biggest consultancy in data and analytical companies in the world uh that speciality and he's now launching a
(06:42) new Venture called propeller Tech which is uh enhancing or using deep Tech to get some uh operational uh advantages and the facilitations and then we have philli Philip is a Serial entrepreneur mostly in AI coming from the defense department where he worked into you know like the hacking community and how to work against that and with that so so we have quite a uh eclectic panel on purpose uh so we can talk about the subjects and the first subject is about value based AI so about value based AI uh we had a first question about
(07:30) value based Ai and that first question is very simple how can we democratize AI how can we work with that because AI is 99% of the time a black box so that's what is happening today you don't know how the results are made so I'll give the F Well you already have a mic Martin so I'll give the first uh uh question uh or to answer to Martin so do you have in your experience any practical examples how you work with democratizing AI or making it more transparent uh yeah thanks JP um so the transparency thing I'll start with is
(08:18) that clear is every okay um because it's really really important to us um in what we do and we're using AI to invest in the stock market basically and we simply would not be putting our own Capital at risk or client's Capital At Risk if we didn't have some degree of transparency the issue is when you're working with machine learning and you want to leverage the power that these machines have for decision making as well as processing the data by definition you can't really know what the machines are thinking so you have to
(08:45) find a way to encourage transparency in your process in your inputs and your outputs so you can get a sense for it and then have the confidence to make one of the the four decisions you can ultimately make um as an investor which is buy hold sell or short any given stock at any given time and so this is one of the things that we had to really build ourselves um starting in 201 17 and we tried to buy something off the shelf but you know we weren't at the stage we are now with AI and machine learning so there was nothing and the
(09:16) key to us was to build in some some means of transparency and luckily because you know we we come from a kind of fundamental analyst background we just said look at all the things that we want to look at as human analysts and these are your input and when you're giving us your opinion as to which are the best stocks in the world which are the worst stocks in the world show us all the outputs show us how those have changed so that we can really get a sense for why you've changed okay I'll give you an example so
(09:44) in the pandemic we um had been holding Airbus for 3 years till the beginning of 2020 and was on Bloomberg TV guy Johnson said why do you hold airb and we had this whole spiel about you know it's like 12 times earnings all the kind of usual metrics um there's this kind of thing going on between the US and China so there's going to be a pivot towards Europe very happy with the answer came into the office the next day and the AI said sell it so there was IM immediate red flush to the face hugely embarrassed but because we trust the AI on the kind
(10:14) of the cell decision we sold it immediately but you know normally we would have kind of had these emotional biases as a human and rationalized why we should still hold it we have said you know all those things are still true and yeah maybe it's not going to be a proper pandemic and so on so forth but because we had this output from the AI we were able to go and do that and we had this transparency of the outputs that showed us that the smart analysts who weren't so lacad isical about the coming pandemic had started to downgrade the
(10:43) airline manufacturers and there's therefore a lag in the performance of the kind of sorry the um the travel industry and there's a lag to the performance of the airline manufacturers we were able to gain confidence to actually get out of the stock and you know thank God we did because it dropped like 65% in the following week good move so you see emotions yeah and having that transparency exactly gives us the ability to navigate those emotional issues that we have as humans all right cool thank thanks Martin um next
(11:11) question is we talked about transparency of AI systems now but that is connected to that question and I don't know who wants to answer that question will AI will aiops sorry will AI remain a closed system or will it go open yes you can take that one if we stay down the same path right now the technology that we have we're trying to patch their explainability problems we're creating these explainability layers and we're hoping for the best so if we stick down this path no but we have other paths there are other technologies that are
(11:45) now scaling up that allow direct explainability within the system you understand the cause of certain decisions being made just like yours you understand how the data affects the parameters um so you can go down these modelbased approaches which um allow you to read and ultimately affect the system as well to reduce the biases uh so these are generally basian based for example and these Technologies are now scaling and are now much more efficient even in terms of data than the blackbox approaches which have been built just to
(12:18) throw more and more more data at it well we're hitting a data wall there's no more data to throw at the models so you now need models that are much much more efficient and that have this underlying structure structure that allows you to understand what causes a given effect in the world and then you can just simply fit your model to uh the the data that actually comes um through this agentic based structure cool thank you um Greg you want to react on that or uh because you have a little bit more experience in the
(12:52) healthcare as well so so so how is this this transparency and also what I just said is you know like will it be open or not because in the healthcare it should be more open exactly everybody should benefit from that yes exactly so the thing is um in the nutshell normally we want to have a better communication with the best researchers on planet Earth okay and then we have several countries and then we have the industry the farmer okay they make it for a profit right but I was just amazed that we have so much we have so many tools we
(13:22) have so many things we have human data we have cohort studies with healthy not healthy volunteers and I wonder because it's about regulation this is the health thing that we always have the same medical uh medicines and I say why we not approve a newer ones because I said no it's already working it's about the price cut but I think uh it's not only about the medicine the treatment is actually also about the diagnosis so I can imagine that we use a medical AI for certain biomarkers and biomarkers is for example for the early detection of CNS
(13:55) diseases like Parkson Alzheimer and eye diseases we have now the special age the the Aging Society silver population I should I think we should we could do much more when the industry also open up because they they they they have close systems saying you know what I still do the profit but they should create more organizations now there are certain um associations where they say we have to team up so several Farmers they teaming up and say you know what we have to work on new things even rare disease so I think there's a kind of trend is good
(14:25) but I think we need um very special evangelist we need people to discuss this because uh it's it's better for the quality of life of the human beings so I hope so yeah thank you uh yes yes um so we have one more minute yeah go ahead special on on the healthcare I I fully agree and I think there is an other element that that could be very important on on Healthcare there is so much data Healthcare data which is sitting around in all kind of silos and diagnostic information um after defect information and I think that's important
(14:59) important and specially in healthcare that you bring that data together but there the whole the whole privacy data privacy which is very important for patient data there are tools and techniques to overcome this like Federated learning Federated Computing where the data resides at the owner and that's the thing what we try to push as well forward in in Belgium that that those diagnostic information clinical information patient information bring that together to do some research and and advance in in in the in the world
(15:27) yeah I I wanted to add on on my side in terms of democratizing the the whole AI it starts with the evangelism or or education uh behind I think the the biggest stopper of getting democratized AI is the the change that is in every industry is it's going to be so disruptive on everything that we do like even if we put the hellur how is going to change like healthcare organization Pharmaceuticals is we're talking about improvements efficiency that is going to affect millions of dollars in in the industry as we know it today and it
(16:04) could go to banking it could go in the way how we invest uh because it will give decisions to the masses like kind of um what happened with the robing hood app with uh GameStop it was the masses the one we were using in this case wasn't really AI used by the masses but they were using the messes the information so I I think goes into losing the fear of of using AI like I meaning how you can blame with that and it comes with education like like if you know something then you're you don't have the fear if you keep it away or it
(16:40) becomes more like a a thing of a few people then the inequality will go bigger uh in this case I I would like to add just one point here in that uh we have to go and take one step back understand what we are talking about it is intelligence at the base and we have seen how human intelligence has evolved over the centuries and and millions of years now um it is all about information if we had if we have closed information then it is limited to only certain people so people will not be intelligent enough similarly it goes for the
(17:15) machines as well fundamental aspect is how much we can open machines to what type of information as and when we open it more we will have better intelligent machines which will be augmenting augmenting the human efforts so that is what my take on that is so we should have the open systems for the AI to learn more true I agree with that indeed but also I don't agreed as well because you need to get better leverage of your data you're already having having all the data open is an utopy so we see what's going on today with open AI some data
(17:51) sources are being blocked so they are stopping sharing the data because they always think in the oldfashioned way I have more data than you so I'm smart than you so I'm always I'm not a big fan of that and I'm not a Believer scientific believer in that I think you can leverage data that's why what we're doing in for instance We're translating data into signals and the cocktail is predicting something on that so and sometimes that data is come from the left side and sometimes the DAT is coming from the right side so this is a
(18:18) different this a it's a very interesting discussion actually so so how far how much data do you need do you need a lot of data do you need to be 99% certain in your prediction also a discussion is 30% good enough so so yeah but that's another topic so um I would like to ask the last question for panel one because we have like three panels but we going to stay so is and that's a very it seems to be a simple question but it's not that simple is um uh can AI be creative in the beginning we said AI will never be creative and the first thing it
(18:52) disrupted was creativity but if you think further uh we as human we comp pause with you can say oh I don't want to think today or I don't want to be creative today or I want to sleep or I want to do this uh or I can take a break uh a machine can cannot do that so you you have like several things of looking at the creativity by machines and I would like that Carla would would answer that one I think you're the best place to answer this so um from my perspective I think AI can be creative to a certain an extent uh but yes I mean we cannot
(19:25) deny the fact that it can create fantastic script scripts can actually put together music compositions but there needs to be a human guidance to that and so that's the important part because the data doesn't have emotion and so in order for creativeness to happen there is definitely like this inspiration motivation um encouragement that needs to go involved and I also think of like creative Industries um in that regard with artists and how they're also that composition with regulations when it comes to ownership copyright
(20:00) these are things that need to be also considered from an ethical perspective um so it's definitely I mean complex in that regard and to to the extent of it so I love that perspective and I think I'm going to add on to it if we Define creativity as two fundamental processes which is you add categories in the world or you connect existing categories in new ways deep learning was the dream because it allowed us to not have to know the pre-existing categories in order to get actual results so we believe that it could do something like structural
(20:36) learning unfortunately it is somewhat limited so that's why we think of AI as not really creative because mostly we think of it as it connects existing categories and to some extent because of the statistical regularities it can sometimes output something we don't necessarily think of right away which feels like creativity some models though are able to do structur learning realizing new patterns in the world and as soon as our models are able to do this they will be creative and perhaps even as creative as we are now the
(21:06) beauty is going to be mixing the human entropy this natural tendency we have to create new categories in the world both materially and uh virtually and mix it with the entropy we'll get from AI cool thank you for this we're going to panel two where we get a bit more so the aim is uh that we're going from from uh value based to more Technical and then to the conscious decision taking so we're taking you to a little journey towards that so that's the red line within the discussions that we're having so if there's a also a question from the
(21:43) public please raise your hand and we'll if something is not clear or whatever so we're like in a small environment here that's the aim to be a little bit interactive as well so let's go to the next panel and will you will Anya will ask Greg and zot to change so so um yes Wilfred okay so we heard a couple of things about value based AI let's now go more into technical aspects there are two basically I would say uh one is realtime data um so not the data of an industry report of two years ago not even the data of yesterday
(22:20) but today like real time and the second is AI this technology we hear so much about which is maybe also a buzzword to some extent um and how can we use those for investing and I want to start really simple um by asking a question to yoges um you're an investor seasoned investor and I'd like to ask you what do you think is the role of real-time data uh currently for investing for family offices or for any other investor thanks so much very appropriate question from in in the new era when we are talking about because uh usually we
(23:00) have to understand the concept of uh what is the intention of a person when he is going for the investment uh the first thing which goes in mind is definitely one thing is the growing the wealth but then very important part is with what purpose that family office would like to associate themselves with so there is an emotional aspect goes into it um as in when we grow the AI systems there is only one way to grow that is do have appropriate data available for that learning and to draw those conclusions which are emotional
(23:36) based are not that straightforward with the limited data so what Wilfred you are saying is that the realtime data definitely it will help a lot to the machines to think on those lines but that is not something which is 100% foolproof so that's where the intention of a person comes into play and those particular data points which are not available in the data in the real time which will come from the human mind and those will have to be fed back into the system that type of feedback loop will definitely help systems to decide better
(24:10) that what type of investments will have to be there on in what form form and which particular area of uh basically business they need to invest in so that's what I would say on that thank you I see zai nodding maybe you want to add something to that or just agreeing I just want to agree completely also because we just talk about data has no emotion whereas families like human Minds have a lot of emotions so I think for better decision making the emotions part to be looped back it's essential so that that definition is
(24:48) limited the real data yeah I just wanted to add to that as well I'm I'm quite new to the family office game running our own family office but um so far we've seen limited use for it in the sense that um you know you have the saying that when you want to see a rational man go broke give him more information um you know that's that's the whole risk that we're seeing basically when making the decisions of where to invest how to invest mainly looking at private funds fun investing in private funds as well definitely
(25:18) because we have an appetite for early stage uh private funds that go into early stage for example there are very few data points right like when you go in stock there's quite a bit of data points out there different risk profile different material level we're rather in early stage and there we see that it's it's very difficult actually to get some signals out of the market to you know to make it worthwhile of of making decisions on that um even more so it might even become a risk right so the way we try to approach things is going
(25:47) more into a venture Studio model because we believe that actually the whole VC model is a very poor track record employing Capital extremely poor um it's like a functioning drunk um well with all due respect to VCS but um now a venture Studio model for us works better in terms of der risking or having a better grasp on the risk um um you know profile of of going in early stage basically so that might be a contrarian view I don't know but yeah I I kind of agree with I agree with that in that sense is that uh we work with a
(26:26) lot of family offices and we see how decisions are still taken with data that is a year old and we saying what are you doing I'm not saying that that data is wrong but it's just one stream that needs to go into the equation because a lot is happening in one year especially when you go into the startup world and the scaleup world you it's a it's a century a year so this is things that are getting now with the new AI systems and I'm not talking about llms and uh nlps I'm talking about predictive AI I take this as a white name uh broad name
(27:04) but in this predictive AI when you work with real-time data it's completely different perspective I I wanted to to agree with Philip like like the the the approach when you're working with startups or companies that are just studying it goes to the emotions like you need to meet who is the funer team what is the energy behind do they have this to run this company uh otherwise like I'm not working with them like it it's you work with a business plan but it's just it's not real data everything is created it's an idea it's more is
(27:37) this person or this team willing to take the the idea to to closure in terms of real data and AI the mix for and I think not just for investment for for everything I think what it what is changing is in the past we were having real data big data and they were calling oh we have real data but we didn't have the tools to analyze all that real real time data now we have it so now the data is becoming more operational so we are making decisions right away there with with technologies that are allowing us to to do
(28:10) it yeah yeah I can add maybe just from the public markets perspective um having heard and agreeing with all of your points in private markets private markets right the fact that it is real-time data or pretty close is really really important when stock prices are traded practically every day some information changes the market is in a way racing to kind of get ahead of that and predict the future better than the next guy cuz our performan is a zero some kind of game basically so you have to be really on point with the data um
(28:37) and so for some sorts of invent investors in strategies it matters more than others you have um hedge funds that use high frequency trading strategies and those guys it has to be second byc data you've got us who have more of like let's say a 12mth view on a stock where if the machines change their mind a little bit it doesn't matter if we trade it today or tomorrow you know it's not so key but as long as you're kind of more or less on top of the new data that's coming out day by day um you can you can stay current and to your point obviously
(29:06) this data is ubiquitous and it's more a question of how do you then boil it down to stuff that's useful for your decision- making process um and that as Alo said actually is becoming um very very doable um with the machines that exist in the the the programs so and I just uh would like to add that you know to enhance the point that real data can be your true differentiator going forward that is um kind of like the data is the backbone of decision making and it's incredibly important going forward even more than before and so of course
(29:45) integrating that data interpreting that data is your GameChanger but without that data um you know um you're pretty much going to stay behind so real time data as a differentiator I think that's something that needs to be really much enhanced and kept in mind okay thanks everyone for the additions let's go to the the second question so we talked about the role of uh real-time data for Investments of family offices or investors in general um another angle you could take is the skill set or maybe put it wider than
(30:23) just education the awareness that family office should have now in this new time of of of real time Ai and investing um a lot of family officers that I meet even if they manage big portfolios a billion plus they have just a couple of analysts maybe a data scientist and that's sometimes bit weird to me because they manage such big portfolios um so that depends a lot on these small amount of people who who make the decisions and that led me to the question like which skill set or wider phrase which awareness should
(30:58) family offices or their managers have to do uh proper and optimal investing um maybe John you want to kick that off maybe you could use that example of uh reducing number of people yeah yeah so this like the democratization question effectively uh so um so when I used to work in the institutional Asset Management World our USP was the opposite of democratization it was you know we can afford to employ 8800 analysts sitting around the world to help us gain some kind of information Edge over our competitors since having a new kind of
(31:37) Boutique Etc and since we've been able to build our own machine Learning System we simply don't need to hire those analysts anymore we run um our portfolios with a team of three people because the way we've designed the system is to do the job of all those analysts um do it better do it in more real time without the cognitive emotional biases turning up to work hung over being sick being varying in quality all of those kind of things you know and they can do so much of the heavy lifting that it makes our job you know a lot
(32:06) easier and we're actually outperforming those big guys you know by by a lot um as a result of embracing this technology and this is something it's a process um that's going to put that sort of decision-making power into the hands of certainly family officers if they've got a couple of analysts they can do this they can work with um a machine Learning Company themselves to build it the way they want to do it and that's kind of like it's not what the man in the street call democratization but you know it's it's a kind of progression along that
(32:31) line to really give the power to um individuals to kind of partner with these machines and make smart investing decisions which is not always the case can I add something to this yeah John and then we go to Mahal and I think it's it's that's certainly valid point what you're mentioning but um I'm working from time to time as a senior advisor from for some private equity and a couple of them have made as well um like you did in organization an own data model AI tools on top of it but they use this concept as well as a selling mechanism
(33:04) to invest into new companies when when you're sitting into meetings they use this offering they um to their let's say prospective investment companies they want to invest look we have this model not only for us but we will help you as well to do uh additional Investments Acquisitions for your company and you see that resonates with with the people sitting on on the other side of the table so you might want to have people who have this skill to understand what constitutes noise versus signal so then what you can do understanding that is
(33:35) have people who have this skill set to just scrape just scrape everywhere relative to to the signals you're looking for and then do feature engineering like what you're doing or you can find Dynamic causal modeling which will take a lot of data and do that work for you of identifying what constitute noise and what constitutes signal and then continuously scrape and get this um realtime data and more than real-time data you also to factor in your perspective relative to the data and how your action is going to affect
(34:03) this real time data because you do exist in a a second order chotic system so these are the key uh crucial skills and ultimately that can be a team of three people like in in reality kind of what I was saying before with the lobes in the industry inside the technology there is also a Lobby in in the way how eort and productivity are are driven like if if you can get get more money by doing something why you want to do it for less and that's the lobby that that we were talking before in reality there are technologies that can optimize the whole
(34:37) thing so it goes more to what you were saying it goes more into the knowledge the context the business knowledge and you can leverage technology to do that and and gain productivity thanks um I never it's that time already for the second yeah you have time for one more question okay okay so we discussed uh uh realtime data and the skill set then maybe let's think about Technologies which type of Technologies can be applied on top of these realtime data to um well to do two things we discussed the purpose side also but also to get
(35:18) just optimal returns um yogish is that maybe something you want to add a comment about yeah sure um when it comes to technology the there are uh definitely I would not go to much technical as in technical no brand names it's just yeah that's so so there are um I would say I'll go back to the previous conversation when we talk about open and closed ones right so um there are so many open Technologies which are available these days and important part is that uh how we improvise on top of those instead of going for the close
(35:55) Technologies so much so um intention is about again going back to the information exchange which is very important so we have to pick up the technology which is more information um I would say conversent and it can make more sense from the information which is coming from the realtime data as well as from the feedback loop which is coming back to it from the the investors or the home offices and that kind of thing so um the humans are the best people to judge about the technology not the technology will drive the use of
(36:30) technology I would like to say that thank you I think I'm going to point to specific technology so because you deal with systems that have high uncertainty and volatility and you don't want to make a decision just to make a decision you want to make a decision that leads you to a good outcome with high certainty you want to deal with system that can deal with uncertainty and have that baked into the system so any kind of technology that gives you these bars the second part is you want to deal with time series so the
(36:56) technology I would pay attention most is rsls which basically is a recurrent um linear dynamic system so basically what you do is you have your system you have uncertainty about the linear Dynamics and you understand when you switch from linear Dynamic to linear Dynamic and this gives you a very good result yeah that's exactly what we're doing in our company we're using a dynamic time series indeed so um we're going to do a little break in two minutes so to and then we go to the last panel and to do the bridge between what we discussed now
(37:28) to the last panel which is the most interesting one I would say is that we discussed now we heard stories oh we had 8 analyst and now we're dealing with three people because the machines are preparing everything and are digesting and are telling us when to go into contact when to do the investment when to approach this or that so this is what is basically today so the next level of discussion will be okay then what do we do as humans the machine are doing quite a lot for us so this is starting to look into more of
(38:03) ourselves how we take decisions in a more conscious way because it will be all about relationships because the machines will tell you when to contact when to invest how to invest uh how to approach etc etc you will be giving all the elements by the machine so you will be giving the signals and it will be almost or in real time so and this is the next level we're going to discuss them so so let's have a little break yeah 20 minute break and then we come back [Music]