Session 2: Advances in Computer Applications in Data Science
5:15 pm – 6:00 pm GMT
Round Table
Nikiforos Panourgias, Harald Stieber, Dan Laidley, Sudip Patra – Regulator and Practical Challenges and Opportunities of Data Science (Chair Sheri Markose)
Transcript: –
Sheri Markose
So that was a very wonderful afternoon of talks. I learned a lot. I mean, I will, I will start with Harald, because he was left, you know, unable to answer many of the questions that we were all very keen to ask him. So, part of the, the, you know, you worked on you said, the relationship between legal entity identifiers and smart contracts is all about precisely sort of identifying agents involved. Various clauses in contracts and so on, your data set your data set, you first look at, let’s say in the Airbnb data set, but what I really want to get to, to understand is, the, the extent to which entirely smart contracts, in other words, self executing online contracts exist in the economy. And the Airbnb is where you as a researcher, take the data and then analyse it. Whereas pure Self executing smart contracts that exist entirely online, that’s no Counterparty, Sort of existence in the analogue world. What percentage of all contracts now are pure smart contracts? For one? That’s the first aspect of my question, as opposed to, you know, the analysis that you’re doing on existing contracts, that could also have an analogue? Avatar?
Harald Stieber
Actually, it’s been a while that I have not looked at to the sort of, say, volume turnover. Any metrics on the main platforms that I see, I think Ethereum is still the most popular platform for hosting smart contracts? I would answer in two ways, we see that we are in a way, maturing into two kinds of use cases for smart contracts. One that is the classical one that that with the ERC standard that Ethereum has launched. But that can be implemented on other blockchains Of course, as well. I mean, there’s no reason why you could not launch a private currency on Bitcoin. In addition, I mean, I mean, it can be self money, but do you see what I mean? So, I mean there’s the one smart contract that you mentioned that basically lives only in cyberspace and that is only also efficient, when it is purely connecting to address the first time, but what we see more and more and that is maybe for supply chain management and risk management, the more interesting type is a sort of say a perversion of this one, where you strip away the money item, the monetary function of the smart contract. And you may even strip away the, what was initially the accord defining feature of automating payments. Yeah. Because the payment is, in terms of what you do, it is just sending information from one address into another. Right? It is it is different from it. I mean, cash is not a meaningful concept. And it’s just, it’s just changing the value of one address and another address, this is all you do. So instead of a payment, you can just send any kind of information. And the smart contract basically the second type of use case is what we see more and more in supply chain management and contract management like asset liability management, where you’re not interested in a transfer in a payment that changes the balances of two addresses into wallets. But you just want to basically exchange other kinds of assets the assets being some relevant information, or simply you change the status of two addresses but it’s not corresponding to a cash or a monetary balance but to something else. You know, and there are numerous use cases in the context of, of compliance of regulatory or supervisory reporting, where the payment function that was originally the value proposition of the smart contract is not the most interesting one, but rather to, to confirm to a third party to and to any, to any third party. So to principle, unlimited, unlimited number of third parties, that certain information was provided at a particular point in time and to confirm this with mathematical you know, in approximate certainty, to a very high degree. So, I think these two universes have been diverging for a while. And I’m, I’m personally not very upbeat on the currency use case, I think in the long term the analysis carried out early in the early days by people like a black hole now, Garrett, at Cambridge, that, Bitcoin will have also in the medium run a value for jurisdictions with poor governance, weak governments and poor rule of law and these kinds of things. I think this is correct. But in the medium term in European US. I mean, I don’t want to speculate on the US but in Europe, in Europe, Bitcoin has no future, clearly. Because the digital euro is on the doorstep. And the digital euros is the end of Bitcoin for European markets. I’m sorry to say that
Sheri Markose
So is this insider information?
Harald Stieber
No, it’s not insider information it’s all public domain. You can just Google ECB, digital, digital Euro, you see, this is progressing very fast.
Sheri Markose
what do we talk about, what time horizon have we got?
Harald Stieber
we talk about five years. So in those five years, of course, you can see a wild gyrations still in the cryptocurrency market, and they will continue to exist because there are a majority of jurisdictions on this planet that are actually very poorly governed. I mean, it’s the majority of countries on this planet that have very poor governance, very poor rule of law, where people have strong economic incentives to hide their, their assets to transfer their, belongings across borders, without the government being knowing about it. So there is there still in the medium term, a very bright future for cryptocurrencies, but I’m working in a policy institution, European Union in the European Union, it is a dead end.
a supply chain use case, which does not involve payments. So you can have a supply chain management using smart contracts while keeping all your payments in the standard banking network. It is highly effective, it adds a lot of controllability, manageability, and transparency. And it can prevent you from paying fines for honest errors. Because the compliance and reporting environment has become so complex, that you pay mostly fines for honest errors.
Anindya Chakrabarti
It’s a new story, first of all, also, you know, an interesting point to begin with, but this was not part of your presentation, I know but given that you make this comment that Bitcoin will be essentially replaced by this, but that also means that the nature of monetary policy would also have to change. Because the usual way of setting up the you know, the borrowing rate and the lending rate by the central bank will not probably work. So can you shed some light.
Ankur Sinha
my question is very much related.
Sheri Markose
Actually Anindya Can I just point that let’s not directly relate it to a monetary policy
Ankur Sinha
My question is more on the government structure of the digital Euro that you’re talking about. And you see that the governance structure is very how different it is from euro and how different is it from the bitcoins?
Harald Stieber
it would be I mean that it would be not different at all from Europe but actually some there could be some commonality I can explain to Euro there is zero a difference in governance by definition, because the digital euro will be an application, yeah, it will be a software application, it will not replace the euro. So, the monetary policy will be defined by fiat money, the euro and basically you get an application that will have a small positive cost for using it. And I think there is an economist at the Bank of England. Yeah, of course, Michael would have worked a lot on this concept, why actually a positive user cost in four of the digital currency is really also a very interesting beneficial concept actually. So, introduces a price element that has very interesting features actually also from a monetary policy point perspective, I can highly recommend Michaels work on this, but there could be some things that we can learn from the Bitcoin. And that is, the way they use. They allow basically every user that is ready to invest into you know, studying the protocol and looking for glitches and weaknesses. So I think the open source philosophy, the philosophy of having an open ended user community that can contribute to improving the protocol of the app if you want. I think this is something that we should probably adopt from the private cryptocurrency space, very much like in telecoms they tried to close models, and they had to go back to open contribution models, because they were just producing better results, more stable, more stable applications, more stable systems.
Sheri Markose
Okay, can I move to Niki’s talk today? So, Niki, you’re concerned about various aspects of the legal entity identifiers. But I also know that you work on innovation and recombination, it’s another theme in all digital systems in one of the great advantages of a digital world is that you can recombine things, it’s easy to do that with software. It’s very hard to do with analogue components. So do you want to just throw light because I’m widening the scope Because I know your backgrounds, where you’re coming from and how you contribute to this theme of the digital economy and so on. You want to throw some light on innovation and recombinations? Niki?
Nikiforos Panourgias
Apologies, the dreaded mute button. Yes, of course. And I’m probably going to go back to something I think Harald said about Schumpeter, you know, and recombination. So we were it’s interesting, I think one of the interesting things about the sort of digital economy, we use that sort of catch all term is that, you know, how digitization does allow this kind of reconven ability and, and generativity as well, and one that was probably stay out sort of still at Warwick, before I moved to Leicester, we were working with a couple of my colleagues there on, you know, kind of trying to understand value, you know, the value of digital, of digital innovation in terms of recombinations of digital resources, that could be sort of combined and recombined. And, and how that sort of changes our view of value from something which is sort of like a number to something which is more, you know, like a like a field or a space. I think we, in that initial paper, we call this open value space or something like that, which is sort of very much then sort of changes your perspective with the sort of viewing how you sort of channel value paths through various combinations or re combinations of these digital resources, which of course can include, you know, human as well as kind of technological elements there as well, you know, you could argue the so, you know, cognition is a type of, you know, resource that could be combining in these kinds of, in these assemblages. So yeah, and, it was kind of quite a theoretical paper. But then subsequently, in the project that I’m working on at the moment with the, with the people at cork that you were, that you were a reviewer for, we were looking at sort of trying to sort of use that in a more kind of concrete, more, more kind of, use kind of case situation, looking at, say the, you know, value co creation between users and developers, say, of, of services like revolute. And looking particularly at, say, you know, forums like, you know, Appstore comments that people leave, and sort of trying to sort of see how the apps kind of evolved, you know, it in some way related to the way that their users kind of comment on these apps. So I don’t know if that sort of really helps very much. But we’re sort of trying to build the kind of more concrete kind of empirically based sort of view of this kind of idea of recombination, and generativity that we seen digital innovation to sort of try and actually now migrate to analysing actual cases. And we’ve got a paper in review now a second round with Mr. askew, which is exactly on the case that we talked to you about. So I think it’s a very nice work, because he’s sort of seeing how something which probably, you know, was very theoretical, even in, in the Schumpeter view, but actually, in the digital economy, is something that becomes quite concrete. And I think the challenge to us is to try and see, you know, what conceptual tools we need, whether the kind of quantitative or qualitative or combination to kind of actually try and try and understand this kind of digital environment in this way through this kind of combination recombination. And the thing that sort of made us go for that view was this sort of, you know, seeing how users can actually combine and recombine kind of digital resources themselves. But also, you know, increasingly, you know, smart contracts is one subset, but you know, digital agents, you know, can actually be combining digital resources to provide, you know, maybe sort of new services that are sort of almost sort of Robo designed, you know, so this is the sort of reason why we think that’s a long term kind of interesting area to kind of investigate. So this was sort of working information systems, but I’m sure there’s a lot of crossover with a lot of the work other people are doing. And you know, even with the, you know, just to go to the last paper we had about the business news, you know, we’re actually sort of developing from the initial paper, now, we’re sort of looking at sentiment analysis of those comments, but in a kind of machine learning kind of way, and seeing if we can actually sort of find some kind of relationship between the sentiment that we pick up in the user comments and kind of tendencies to download, or our uninstall kind of particular apps and so on. So it’s a very kind of interesting area, I think.
Sheri Markose
Thanks, Niki, that will open up the discussion now with Sudip, Sudip let me put it to you about quantum computing of what I understand, I think it’s on a pragmatic level it is going to happen because with all the super positioning helps people do is it to my mind, it is like a sort of Uber, Uber parallel path, you know, introducing parallel computing in a very Uber way meaning to say, at the Super positioning, whereas we would do things sequentially, or in a less parallel where you can conceive multiple scenarios at one of the same time, right, you get the example of somebody having an apple and having a banana, obviously cannot happen at the same time in real time. But offline in simulations and in cap calculations, I think this is going to be the case. But there is a very interesting issue to do with blockchain and the double spend problem as I keep telling students, the physicality obviously obviates the double spend problem with cash. Because if I give you $1, then it’s no longer with me and it cannot be in two places at the same time. So, in one sense, the entire problem of double spend is obviating or trying to get around this thing that a particle can be in the same in two places at the same time, but on the other hand, in the offline environment, so you know, so we got this thing that in real time and in on online environments, these things cannot happen, you know, shouldn’t happen in proper financial accounting. But offline, the pragmatic aspect of superposition in multiple scenarios holding one of the same time is something that can happen and is going to be the future of computing because it speeds up this parallel process, you know, in a, in a sort of very intense way. So, I’d like your, you know, comments on that.
Sudip Patra
Yeah. Yeah. Thank you very much. Yes, I think that a lot of exciting stuff is happening in, you know, trying to adapt to quantum computation. And as far as I see that two aspects, one is certainly, you know, a more pragmatic mathematical modelling that, you know, we are doing, it’s really not going into really using the resources, for example, entanglement and superposition, how to use them for kind of future financial technologies that would be right really, the second revolution one can say in that area, for example, recently, as I was just mentioning, there is a very interesting sub branch of game theory, which has come up with is quantum games. Now, quantum games is not simply like mathematical modelling one is certainly there, you know, when you observe real people play a game and then try to describe their behaviour with the help of this framework. But the other one is like when real people are basically sharing certain kind of subjects or sharing certain kind of physical systems and within the physical system that can be certain interesting features in grade like superposition, as you have rightly mentioned, or entanglement. And it has
been trained
Sheri Markose
can you can you tell us what’s the difference between the two?
Sudip Patra
What is the superposition is certainly the linear combinations of say zeros and ones or Yes, and No’s are different words one can say and when you are still not measuring a particular state, the state is in a linear superposition of possibilities. So that is the superposition principle say, but entanglement is entirely different entanglement though it you know, it kind of takes its power from superposition principle, but entanglement would actually mean say for example, that two agents Alice and Bob, and they are spatially separated, and it may be even light years, but they are sharing some kind of system, which had generated at the same origin and then the system has been separated for some time. Now, what happens is that if you can really design that system carefully, then whatever update you do on Alice’s end, that is immediately going to inform Bob or inform Alice about Bob state. So that is like the entanglement and now entanglement is very much like blocked out and it is used as a resource
Ankur Sinha
and when you say at the same time instantaneously, are you expecting some sort of a gap of certain amount of time for light to travel from one system to the other?
Sudip Patra
Yes, yes, actually, you see that we have to be careful here. One is certainly the superluminal travel which is not allowed, but suddenly you can use entanglement as a resource to increase the computation power of your system that can be done without breaking the speed limit of light that is, because that is what is known as to be no signalling condition in physics, but, but very interestingly, these technologies are really coming up and people are working on these technologies for financial markets and so on. So, it is quite promising to see that maybe. And also, another thing I just forgot to mention that in game theory, there is another interesting concept which has come up, which is quantum Bayesian networks. So one was Bayesian networks, where, you know, you are trying to calculate the conditional probabilities at each and every node of interaction. But now, if you think that humans, human agents or even machine learning, etc, they use a different kind of framework for abduction of beliefs and probabilities, then quantum Bayesian networks gives more accurate results. And there’s a huge amount of research that’s happening in that direction
Sheri Markose
there’s one more question, Sudip, that I want to ask you and that has to do with you know, either the point that you made about agreeing to disagree in common knowledge models, right. So there is another solution to that problem. I mean, you know, the the idea that this is sort of fun, I suggested that at a fixed point you know, when you have a contrarian agent, right, somebody is putting in negation on some some outcome that can be predicted. And then when you look at, then you try to find the fixed point of that negation, that becomes non computable. Right. So, of course, since that’s not computable, you have to agree to disagree those points when you have these contrarian agents in the system, so the fact that you saying so that could be another explanation over and above the one that you suggested.
Sudip Patra
Yeah, yeah, yeah, yeah, that’s true. And I suppose I have seen that quite a few people I, you know, research that has happened in that direction. One particular research that I was pointing to, was done by Andre clinic, AVI is like one of the most noted mathematician in this area, and he was the first one to show that if agents do have this quantum like framework for belief updating, then what would happen is that there would have been more general solution to amens theorem or in other words, in most general cases, even the priors are known to everyone and posteriors are common knowledge, you know, the probabilities, even then people might, will be happy to disagree there will be like chances where people may disagree, but that doesn’t mean that the classical solutions will not occur, they will also occur, but they will all occur as a like kind of like more special cases of a general solution. So that, but but I found that that is an interesting direction, because you see that in economics, for example, costly signalling literature is like a very central piece. And in costly signalling literature, there may be some very interesting connection with the, you know, modifications of amens theorem in that way.
Sheri Markose
So it’s from Alex Anthony. What’s the motivation behind the use of quantum like modelling considering that there isn’t much of an argument to be made towards human brain being quantum mechanical in nature timescales under concentration of millisecond sizes and the concentration of millimetres or micro metres temperature is room temperature, the energy scales are too small to consider quantities.
Sudip Patra
Yeah, yeah, I am fully aware of that, you see that is what we always refer to that. One is like this quantum physical brain, which is not our research area. So that is certainly not what is the universe Sir Roger Penrose and Stuart Hameroff are our model
So that that we are not suggesting at all. So our point of view is more into the mathematical modelling of decision making. And also, then certainly, while designing games, then you can harness the real power of entanglement and superposition. So that is like designing the game. So in these two areas, we can contribute.
Sheri Markose
Okay, so thanks, Alex. He’s one of my students actually, double as the first degree in physics and a second degree in physics, and now he’s doing data science at Essex. So let me move to Dan. Dan, I know you know, the theory that you see the role of algos and systemic stability, right. That, you know, the jury’s still out on that, isn’t it? You know, that do algorithm trading, and so on destabilise and stabilise the system. Do you want to say some more about that?
Daniel Ladley
Yes, I can do. Yes. As you said, I think the jury’s still out on it. I think there’s quite many questions or trading behaviour and creating continuity this understanding the effect of algorithmic traders on market instability is very difficult. I think the perhaps a dominant position at the moment is that during good time, so beneficial, they improve liquidity they reduce spread they reduce costs in general. However, during crises or crashes, the opposite unfortunately happens. they’re the first people out to the market, and what do look to be illiquid market suddenly isn’t and evaporates. I think this is particularly interesting as well when you have algos traders who are trading multiple assets have been some very interesting work over the last decade looking at how pairs trading and similar strategies can result in crashes in one market spreading to other markets due to, essentially liquidity spill overs. The presence of algorithmic traders potentially makes up well, post trading is a very attractive strategy for algorithmic traders as all sorts of work, fascinating bits of work looking at how changing communications technology changing lines of wire between Chicago and New York, for instance, are affecting the algorithmic trading, high frequency trading between them. But yeah, so the rise of algorithmic traders has really seen an increase in pairs trading between different stocks. And that is a real mechanism for the spread of liquidity crisis. I have some very recent work looking at that which mirrors some findings in the banking literature that suggests that high frequency traders doing that type of work who are effectively connecting together different assets actually provides stability, or for the time that they IO shocks to diffuse relatively harmlessly. However, again, during severe events, the opposite can be true, that they actually provide a new mechanism for shocks to spread between different asset classes and exacerbate the the crisis.
Sheri Markose
Okay, I think I’ve done my job if you guys want to ask another questions, spontaneously, please, we have one another piece.
Harald Stieber
I had a question to Daniel and a question to, just an observation, when I listened to Ankur and Sudip style lock on fuzzy logic, I remember reading sady original contribution 95 in a library, and I thought it was really interesting, but I can kind of grasp that the formalisation that you proposed today is quite matching very real world use cases, because the continuous transition function of physiologic, it works super well with electrical motors, which closely match this continuous change from zero to one, but human decision making is very rarely a matching a continuous transition function. So we have with much more use cases where we have very discrete you know, and I also see why Bayesian logic is really coming into its own with the addition of the quantum state, I never liked it a lot patient without before quantum but it seems to make sense. Like cryptography makes much more sense when we get to quantum cryptography. I think it was it’s like a missing piece that we were looking for, for quite a while. And it’s super interesting, but I wanted to ask a question to Daniel that it has been haunting me for forever since I was working on financial markets. So for quite some years, I never saw how I mean, I can find many good reasons why you want to have dark pools and other you know, informal markets, non lit market, no matter how you call them, adding flexibility in to existing markets and regulated markets. I think there are many good examples in other parts of the economy, and what I never understood about the financial market efficiency argument is on the vast majority of companies, there is almost no news for the whole year. Okay. There are a few companies that generate a lot of news and the vast majority of companies has no news outside the financial statement. And when I mean, you know, what can you learn when there is no news you know, and the scar city of news is really a compelling feature of markets and the extreme distribution of new so a few players are basically concentrating at 80% of the news items. And the 20% is for all the rest, and that always escaped me apart from the fact that the work of Andrew at Imperial College I have trouble with family names today showed that what really crashes is the link between the mathematical and the physical implementation the computer it’s not so much the algorithm that is the problem. But when the algorithm needs to be implemented by machine, and in the machine, you have waiting times you have discrete jumps, whereas the algorithm works on the fiction of, again, continuous functions, which in the machine implementation do not exist. There’s no such thing as a continuous implementation in a physical machine.maybe Daniel has an intuition that helps me with that struggle?
Daniel Ladley
Yeah, no, I think it’s a really interesting question. I think Well, I think it’s a set of connected questions. Yes, you’re right. Yeah. In many cases, the actual news originating from firms is very minimal. I think there is more news in the broader macroeconomic environment, you will see individual firm line will be changing the environment is, I think, from my point of view off to it, though, is liquidity, training, trading, you know, individuals trading off the actions of other individuals. But still, I think there’s a very valid argument. And this applies to patient care in general is who’s really benefiting from, you know, price efficiency, if we call that call it that out the the microsecond level? It’s not clear that there’s an economic benefit for that, with regards to bringing in algorithmic trading in this? Yeah, I think that’s a fascinating question. I wrote a paper a couple of years ago, looking at the the trade off between algorithmic speed and sophistication, essentially, the future the economics of the financial, economic literature, the theoretical side of it assumes, essentially perfect rationality of algorithmic traders. And as Sheri and others have shown, also, the decisions that these traders are making are into the hardest computational classes. So trying to make any assumptions around algorithmic traders, being perfectly rational at the microsecond level, is ridiculous in a really meaningful sense. Yeah, you’d speak to the people implementing some of these algorithms, and they talk about 10-15 instructions, one conditional instruction, very, very, very simple. Literally, instructions cost money, it takes a very small but positive amount of time to execute each instruction. So each instruction you can shave off, speed up your algorithm and speed up the potential execution speed. So I have a paper from literally a year or two now looking at that trade off, where should you be in the spectrum of speed versus sophistication? My findings are, yes, there is a trade off the multiple equilibria from an economic point of view. But you do start to lose, you’ll lose opportunities, you’ll lose money as you get faster and faster. But you do get through the first two trading opportunities. The next step I’d like to take on that journey of understanding that I think is getting to the one that you’re hinting at that. As well as trying to understand the decisions made by algorithms, it’d be really good. behavioural finance, has told taught us a lot about how individuals make investment decisions and how they trade off risk versus return versus things like anchoring effects, and so on. How do programmers who are implanting algorithmic trading strategies? How do those same sorts of trade offs affect their programming style? You know, do the do the behavioural biases of programmers then filtered through to the algorithms they’re writing? Does the programmer’s understanding of discrete versus continuous, but in your theory versus practice? How does that then come through? To my knowledge, there’s very little work looking at that. Yeah, too. I think it’s a fascinating topic, but it’s not yet been fully addressed.
Nikiforos Panourgias
Yeah, just a quick riff of Daniel’s answer about liquidity trading. I mean, this, this is an area that sort of quite it’s of interest to me, this sort of social studies of finance area where, you know, you’re looking at the sell side, and you know, how much work the sell-side does to kind of fill in those news gaps, that say, the business news isn’t covering and sort of generate liquidity trading, you know, because that’s how they, you know, the sell side makes Commission’s, you know, by encouraging people to trade even if there isn’t that much news. So, in a sense, you know, the other side of the sort of actual, you know, the business news out, that’s out there is also the sort of, you know, the news that is within the sort of markets, you know, so it’s interesting to sort of keep an eye also on the sort of market microstructure a little bit as well, as well as the sort of news element of information going into the trading. So yeah, just a quick sort of add on that sort of came to my mind as Dan was talking.
Ankur Sinha
interestingly, I would just like to add one particular thing to it about the value of business news for trading. So we did have branches cordiality test on does news impact the sentiments or the sentiments impact the news? How exactly does it work? Now, of course, the response that we got was that it works both ways. So the Grangers partiality turned out to be positive both ways. And that kind of explains the analysis that we were working on, when we were taking just those companies into account, and trying to take into account their sentiments and predict the stock prices, price and generate useful traits out of it. It turned out to not lead to anything positive. So there were companies that were generating maybe one news item in two weeks. And then there were companies that were generating hundreds of news items in particular, when we were talking about companies generating hundreds of news items in a particular day, it was much easier to account for the sentiment because a lot of noise got cancelled. And we were getting very good measures of the sentiment, and then you were placing the trades. Whereas for the companies that were in the context of which there was just one news item coming out, after two weeks or so it was not possible to create a tradable strategy primarily because maybe the news came out much later. And that information was already there in the market. And just because it’s such an unknown company, that people who were kind of closely following the company, they had that information. And much later, one of the reporters found out that okay, there’s something I need to report, and then the news item came out in the media, so it just doesn’t have any tradable value. So these are some of the outcomes that we got out of this massive study, which lasted for almost three years or so while we were analysing data from 20 years and so on. But very quickly to take up this question from Singh as well. So, just like to state that, when it comes to understanding the news itself, from the machine learning point of view, that is already a solved problem, as I just showed you that figuring out the sentiments and accuracies and stuff, we are working at a very good accuracy of 95%. So, in case there is one particular statement given to you and you want to understand what is contained in that statement, is it positive is it negative, which companies are talking about which sector it is talking about? That is very much I would say a solved problem. The next challenging problem, of course, that we are facing is that how do we differentiate from reliable sources and fake sources. And that is where most of the research is identifying fake news, identify fake speeches, identifying fake people, fake videos, and stuff like that, that definitely remains a challenge.
Sheri Markose
Thank you very much unlisted a very pressing questions, I have to say, this is exactly the sort of discussions we were hoping to stimulate, in the original format of this conference, we were all going to be in a single room and sort of, but i think, you know, we have communicated and form links, and, you know, sort of got into one another’s space and heads. And I think that is exactly what we want. And hopefully this communication will continue online, you know, we can have further smaller groups that we can have specific areas that we can investigate, and let’s say half an hour video, and so on, so forth, right? To come up with the ideas because we’ve got the funds, it’s the only thing that we’re spending money on, we’re not spending it on booze, we’re not spending on travel, or on, you know, accommodation, things like that, so that we can continue with these conversations in that way. So all you got to do is drop me a line, we’ll try and you know, accommodate that Twimbit is there to help us create podcasts blogs, little further discussions. You know, I can come up with more specific things that let’s say Sudip then he or he answers it. Also Sudip himself puts further blogs on his little thumbnail you notice that all of us have little thumbnail boxes on our Aidigecon website into which we can funnel not just the talks that we’ve had the YouTube videos and all that we can put other stuff into it. So we’ve got just one more thing. I think we’ve exhausted people’s capability of sitting in one place for what is it four hours five hours?
Harald Stieber
if you admit I would have a last curveball to participants going into tomorrow’s discussion on AI and quite philosophical topics. And it’s another of these questions. So I was not very successful maybe in presenting that the, in our study, we had confirmation that we may have, the digital economy maybe showing us actually at the end of this trend to commodification of everything. So we had a secular trend, that everything becomes a commodity, I think digital goes in reverse. Yeah, because people get the values from being part of a particular network. And if you try to rip them out of the concept, you destroy much of the value, you know, you either buy a whole team or not, but you cannot buy an individual member and get the value of the team. It’s not possible. And actually, my, you know, I think the human being in the long term wants to create the economy and the universe in his picture. You know, I mean, this is not a religious statement, I really think we try to create our environment in our in a way that relates to us, because from psychology, I think it is long known that, that if you rip things out of context, people, people go crazy, you know, I mean, people need to relate to their environment in a way that they understand, you know what to think they understand because we tell us. So I wanted to give this to going into tomorrow, when we talk about AI and more human economy in this sense, where this everything can become a commodity was actually not really relating to us. That’s not us. It’s not, it’s not how we function as human beings and in society. So as a question to going into tomorrow.
Sheri Markose
No, I think the point that you made about how price is not everything. And secondly, that profit maximisation may not be the Schumpeter pointed out, we don’t really know all those bits. And what really, what we knew was perhaps only our cost, the duality doesn’t always hold. That’s a very good point you made. And so internet and the session tomorrow afternoon would be just as intense because we brought some very big brains to talk about this embodiment is embodied cognition is very high upon Karl Friston and list of things and you know, active inference, and so on. So it’ll all come up again, tomorrow. So join us again tomorrow and I think we completely lose lost a lot of our external audience. Because it’s, really quite long session, quite intense one. So it only remains for me to thank all of you. That is wonderful. I really think we fulfilled many of our ambitions of being of banging heads together, and hope to see you tomorrow. We have a session starting at 9:30 on the application of FinTech and so on to overcome issues such as financial exclusion, sustainable development, so on and so hope to see you there then. It’s not late in the afternoon. Thank you very much again, and goodbye for now.
