Dr. Ignacio Deza: Full Podcast Transcript
- beyondhorizon965
- Aug 2
- 66 min read
Host: Maximiliano Fabres
Guest: Dr. Ignacio Deza
Location: Bristol, UK
-
Maximiliano: Hello guys, welcome back uh to Beyond the Horizon podcast. Uh the podcast where you expand your horizons of knowledge. Uh I am your host Maximana Faves and with me today uh our guest Ignasio Desa. Uh so Ignasio, who who are you?
Dr. Deza: Well, I'm a senior lecturer at U Bristol. I teach data science and AI. I work with usually bachelor students and master students and I'm the well the deputy program leader for the bachelors in data science at UI. So I I work in I'm also a researcher I work in in different types of data analysis and foundations of AI and different kinds of uh work usually with companies and and with research.
Maximiliano: All right, that's um that's great. That's a lot to talk about.
Dr. Deza: Yeah, it's a lot to talk about. Uh very contemporary uh very contemporary subject. Um so uh okay, let's first uh if you could could you like uh light up the candle?
Maximiliano: Oh yeah, sure.
Dr. Deza: Of course. Thank you very much. Um, so Ignasio, let's bring it a little bit back. Um, so, uh, I, uh, what did you study to become a professor?
Dr. Deza: Well, I initially studied physics. Uh, I'm, um, I studied physics back in the late 90s, early 2000s. Um, I wanted to be a scientist. I wanted to be a researcher. I wanted to learn, I know, maybe win the Nobel Prize. I wanted to to study how the universe worked and all the things that actually still really excite me like how things work.
Maximiliano: Yeah.
Dr. Deza: Okay. And I and there's two that I really crave knowledge and I wanted to understand how all pieces fit together. Okay. So as soon as I start studying it was very very hard. I was um one of the best universities in my country and I it was very competitive and very hard and by the time I was about to graduate I discovered I didn't really want to go into uh string theory and all that. I would I I I wanted to go more into something was more tangible, more something that changed real lives. So understanding the world that surround me, not the world that was like in stars or in the atoms was something more that's why I studied something called statistical physics.
Maximiliano: Okay.
Dr. Deza: Which is about understanding how very complex systems uh are able to work. Even sometimes you understand how one thing works in isolation but you don't know how many of those things like hundreds or millions of things work all together and it's a complete different dynamic.
Maximiliano: Yeah, I feel I feel it seems very similar to understanding okay maybe maybe a bit wrong but I would link it up with like how does the uh body actually works.
Dr. Deza: That's exactly what it is because you got like many different cells in many different organs.
Maximiliano: That's exactly what it is.
Dr. Deza: And you just like so you say, "Oh, we are all atoms." Yeah. But how you put all those atoms together and get a pie or get a human and pies and humans are not the same thing.
Maximiliano: Yeah.
Dr. Deza: Even even if they are made by the same things.
Maximiliano: Yeah.
Dr. Deza: Same composition.
Maximiliano: Yeah.
Dr. Deza: So that's exactly what what I I tried to say. Yeah.
Maximiliano: Okay. Okay. That's good. Okay. Yeah. Got it. Got it. Got it. Right. I gotta say I'm terrible at physics. Uh I actually
Dr. Deza: Everybody is
Maximiliano: I think I did not pass any of my physic class since I started. So I'm not the right person to
Dr. Deza: I mean physics uh in the end is is about understanding the world. uh we lean a lot into mathematics and things because there's a crutch we use to understand the world. So if you need to calculate something to know where something is going to be or what things are going to do then you absolutely need some kind of calculation needs math.
Maximiliano: Math becomes more and more complicated the more complicated things you study.
Dr. Deza: Yeah. Yeah. But uh it's the same idea. Okay. But you don't really need to uh use mathematics to understand. You need it only to being able to use this understanding to calculate something.
Maximiliano: All right. Okay. Yeah. Yeah. Yeah. I think I get it. I think so. For example, if I tell you, oh, gravity like things, I'd know it.
Dr. Deza: You don't need to know the formula for gravity. You just if you understand h how it falls and the principles and you understand that the principle of this pen falling is the same principle of the moon being where it is and the planets and everything is the same idea around which is Newton discovered um you don't need to know the the formula and be but if you know the formula you can calculate things that's more useful but if you don't want to go there you still have the concept Okay.
Maximiliano: And the concept is usually enough for just at least quenching your thirst of knowledge and understand. Oh, now I know how this works.
Maximiliano: Okay. Okay. Okay. Okay. Uh, and why did you not decide to go for the Nobel Prize then?
Dr. Deza: I mean, it's not that I didn't try.
Maximiliano: Oh, first. Okay.
Dr. Deza: But like what made you change to data analysis then? Well, first of all, uh, last Nobel Prize are in physics are related to AI in a way. So, I'm still I'm still running there.
Maximiliano: Oh, you're still you're still you're still in competition.
Dr. Deza: Secondly, uh I mean the the thing about data science is that in my doing my PhD, I was doing climate change uh analysis. It was trying to use physics to understand how the climate was changing and how to get things uh and how to to understand the the the climate of the future and that involve a lot of data analysis.
Maximiliano: Okay.
Dr. Deza: So I was using physics tools like information theory and a lot of entropy and kind of things like that to understand the climate.
Maximiliano: Okay,
Dr. Deza: that sounds complicated as hell.
Maximiliano: It was it was I was a PhD. Yeah.
Dr. Deza: And after that I realized that I was actually doing data analysis mostly and uh physics is was very hard to find a job in physics.
Maximiliano: I could have guessed I don't know why.
Dr. Deza: So I started to work on research more about using physics tools to analyze data.
Maximiliano: All right.
Dr. Deza: And that naturally went into working on research groups about data analysis and then I started to well get different types of jobs doing that kind of thing and then I I got this position as a first I came to the UK as a research fellow so it's like I'm working on in a project
Maximiliano: okay
Dr. Deza: and then I just stayed in university and changed into teaching which is something I always loved.
Maximiliano: Yeah. So you felt the change from physics to uh data analysis was like a smooth transition.
Dr. Deza: Extremely smooth. I I could never tell you when I changed from one to the other
Maximiliano: really.
Dr. Deza: Okay. So there was no barrier basically. Uh no it's the same thing. I mean but it happens a lot.
Maximiliano: Yeah
Dr. Deza: it happens a lot. Like for example, sometimes you don't know where biology ends and chemistry begins.
Maximiliano: Yeah.
Dr. Deza: You don't know when exactly in statistical physics to to data analysis and then data science is more about the technologies I was using. Before I you I used to use some kind of software now I use another software but there's not something I mean also time passes and people just change the tools they use anyways.
Maximiliano: Yeah. I feel I feel that's very uh interesting because um at least for the things that I do well you know me uh we yeah we met before we met in the sauna actually um and um yeah well business is like very restricted I would say like there's one okay yes you can see it as a whole but is very obvious uh what thing is what. So like accounting is one spectrum, sales is another spectrum, sales and marketing you could say is pretty much the same. Uh but like legals is another sector and it's very differentiated I would say.
Dr. Deza: Well I I've worked in in in marketing.
Maximiliano: Yeah.
Dr. Deza: Yeah. So I data analysis.
Maximiliano: Yeah. So I go sometimes the problem with companies is that they don't know where they're going.
Dr. Deza: Yeah.
Maximiliano: So they that's the biggest start issue and sometimes they have to go to the data and understand what's going on.
Dr. Deza: Yeah.
Maximiliano: Um the problem is that sometimes I mean we say that the trees don't let you see the forest. So you
Dr. Deza: oh my god that's a such a big concept. Yeah.
Maximiliano: You see so much detail.
Dr. Deza: Yeah. that you need to but the only way to understand the what's behind that is to go into the data. Sometimes data is is humongous. I'm talking about terabytes of data especially with online and then you need tools. You cannot do an Excel sheet.
Maximiliano: No, no, no. I've tried that.
Dr. Deza: So you need really specialized tools. You need to go there. You need to understand what's going on. Yeah. And usually you need machine learning usually or I mean neuronet networks to go through data and try to find patterns.
Maximiliano: Yeah.
Dr. Deza: Once you find the patterns then you are able to understand what's going on and then if you're lucky you can even predict what's going to happen.
Maximiliano: Yeah.
Dr. Deza: Given the data you have and requires like usually three different things. The first thing is to understand the topic you're talking. So you need to understand marketing. You need to understand the people. So something as a researcher I love because I never stopped studying.
Maximiliano: Yeah.
Dr. Deza: I need to sit down and read a marketing book or read a I know I was working with um construction. I work for the for Highways England.
Maximiliano: Okay.
Dr. Deza: Which now is called National Highways, but I was working for them. Yeah. Yeah. And I had to sit down and understand how high contracts worked. What the basic basically I had to read books about that. I have to sit down. And the second thing you need to learn uh is well you need to know statistics, mathematics. And the third thing you need to know is technology. So how to analyze large amounts of data and how to make it work.
Maximiliano: Yeah. Well, there is um I'm studying business analysis as I told you before and there's uh this graph well this triangle uh called uh puppet people organization something else and it or information technology very very yeah that's so true like
Dr. Deza: that's it that's it once you know the few component in the puppet analysis and in others then you're good to go base view. The thing is we're not in the 20 center anymore. Now you need teams. So you always need somebody who is is a very good person at the data science guy. Then you need the guy who is an expert in the topic guy and somebody who's an expert in like or and the people people. So somebody who has contacts able to to find the things you need. For example, you struggle to find data because the company had you had to navigate all the middle managers until you got the the real data. And then you had death guy too and I there was a quantum surveyor which is a basically an architect uh working with us which was the one who taught me how to things but then he was my reference to go to guy when I didn't understand what was going on. Yeah. And then and so on. So actually um one of the most important part things of data science right now is the social part. There's a part about understanding what you know, understanding what you don't know and understanding where to get what you need to know.
Maximiliano: Yeah.
Dr. Deza: And that means uh socializing means people.
Maximiliano: That is that is such in in business. That's I would say uh thing is I used to I used to be in a sales organization. Uh and there they told you everything. So like they they would tell you the the whole scope. Uh they would tell you look these are the things that you do know currently and you would actually have like a checklist so to say. Uh but it's mainly because it's a very small it's a very specific business. So it's you can you can actually know this. I'm guessing as a researcher that's that's a major problem. You don't know what you don't know, especially if you're like at the tip or like at the edge of knowledge in general.
Dr. Deza: Sometimes, sometimes it isn't.
Maximiliano: Oh, yeah. Well, it depends, I guess.
Dr. Deza: But yeah, they would tell you basically this is what you do know, this is what you don't know, and these is how you can get uh to the people
Maximiliano: Exactly.
Dr. Deza: who you who know what you don't.
Maximiliano: Exactly. And it's always this continuous uh uh thing. And that's what I really loved about that job that it was very structured and they would they would be very open to to teach you and everybody would be like very helpful towards it. Yeah,
Dr. Deza: I would say um it's a little bit more complicated uh outside of that uh business uh so to say uh mainly because if you are like an entrepreneur or running a team the basically you have nobody to look up to as especially like you're the head of an organization very very difficult I mean I think there's a there are pros and Yeah, of course. Pros is that nobody you don't need to deal with other people's with the clients but not with your bosses. You have no boss which is great.
Maximiliano: But you have clients which can be as picky and as uh problematic as many bosses. Your client is your boss basically.
Dr. Deza: But uh so you decide who to you you can make your own decisions.
Maximiliano: Yeah. and also teaches you that making your own decisions uh have consequences.
Dr. Deza: Yeah. Which is something some people who work in very large especially large institutions they don't have to deal with consequence of the actions because they they're isolated. they're isolated and they can it's a blame game more about um facing your the consequence of your actions. That's something we don't have in academia. Academia we we are actually very uh consequent to what we do if we are not if we're not able to finish our research
Maximiliano: we have to give the money back.
Dr. Deza: Damn. Really?
Maximiliano: Yeah. Rarely happens. I've I've I've heard that academia as a general thing is very entrepreneurial in in that sense is it's like you are doing your own thing.
Dr. Deza: I I'm my own boss. I I'm not my own boss in the sense that I can earn more or less money. So my my salary is fixed.
Maximiliano: Yeah.
Dr. Deza: And that's very good because I don't need to worry about that part which can be so good.
Maximiliano: Yeah.
Dr. Deza: But that's so good. I mean I mean also it works the other side. So even if I do more things, I'm not going to earn more.
Maximiliano: Yeah. Yeah. Yeah. That's that's
Dr. Deza: But uh apart from that, I I'm the leader of my teaching. I can decide what to teach. I can create my own uh teaching uh program for my students. technology changes. uh I can change uh a topic in my I was before for example we were teaching something about something in big data and now we're changing it because uh now is everything's in the cloud so you don't need to teach that anymore and then you just can adapt your own curriculum and you are uh your boss you know as an employee you wouldn't be able to that you would need some permission from some manager or something
Maximiliano: oh my god I hate that
Dr. Deza: but you don't have that and as a researcher something that I truly hate is that asking for permission to do the obvious thing.
Maximiliano: Well, that's that's how big organisms work.
Dr. Deza: Yeah. Well, that's what they that's what I will kick their asses, I'm sure. Then on the other side, as a researcher, you have uh freedom.
Maximiliano: Yeah.
Dr. Deza: And then you have to write your own proposals. you need to apply for money for grants or from things from collaboration with companies or something and once which is very similar to basically a sale.
Maximiliano: Yeah. Like you're pitching basically.
Dr. Deza: I can help you out with that.
Maximiliano: Sure.
Dr. Deza: So once once you do your sale and you get your money
Maximiliano: Yeah.
Dr. Deza: then you you get a a job or a collaboration which again is not it's not really going to go into your bottom line but it goes into your CV.
Maximiliano: Okay.
Dr. Deza: Yeah. So you are able to uh do then your um your your research in some uh topic. Okay. Some companies really want to be with researchers because usually researchers are very they have better education than workers
Maximiliano: than any Yeah.
Dr. Deza: And usually they're quite cheap actually. I mean compared to paying uh data scientist for example
Maximiliano: yeah full-time salary data science
Dr. Deza: um collaborating with university is usually cheaper and you get maybe one or two like PhD top researchers to work on your problem.
Maximiliano: All right.
Dr. Deza: The difference is that we require to publish that.
Maximiliano: Okay.
Dr. Deza: And they don't like that much.
Maximiliano: Not always.
Dr. Deza: Yeah. Sometimes they love it is a free PR. Sometimes they don't love it because they have to uh give up some like information about their companies. They don't want to. So that's uh it's part of the sales pitch.
Maximiliano: Yeah, I guess so.
Dr. Deza: Yeah.
Maximiliano: Well, yeah, it's targeted people and so on. So I until now just to give you a taste of what I've been doing I work with highways England and some construction companies on highways motorways.
Dr. Deza: What do you do there?
Maximiliano: I was working uh it would take a bit to explain but basically was uh making um AI enabled system to be able to translate costs from the bottom like the guys like breaking earth all the way to a because they don't have like a a easy system to to um basically account for all the work. They use something called bills of quantities which are open text where they explain this we did a hole in the highway and we did this and installed this and then and that's the money we we charge and from for there to go up to a high level where you need to account for all the work and everything that happens.
Dr. Deza: Yeah.
Maximiliano: It's not easy. It's not like it's not like buying potatoes. you get uh you cannot do accounting with just an a description of the work. So we use AI
Dr. Deza: because you don't know the terrain. You don't know that. Okay. Okay. Okay. Okay.
Maximiliano: So we use AI in order to try to find similar kinds of jobs.
Dr. Deza: Oh all right. Okay. Actually quite interesting.
Maximiliano: It's very interesting. And then the way to do do this up and then we we use a code which called ICMS which basically is able to uh account for similar things and then we all the way up it was actually pay for the innovate UK which is a department of transport.
Dr. Deza: Fair enough.
Maximiliano: Yeah very good.
Dr. Deza: Yeah that's and other things I did I worked for marketing for trying to find churn in customers.
Maximiliano: Okay.
Dr. Deza: for um an academic pollution. Then we so trying to find if people are likely to to delete their accounts so to stop paying.
Maximiliano: Okay.
Dr. Deza: And we discovered that because they they wanted to know how much time in advance
Maximiliano: Mhm.
Dr. Deza: you could know if somebody was going to was going to pay stop paying the subscription.
Maximiliano: Yeah.
Dr. Deza: And I did uh
Maximiliano: and if it was possible to change it. I'm guessing.
Dr. Deza: I mean, if you know with enough time,
Maximiliano: yeah,
Dr. Deza: you can email them or call them, offer them a discount or something.
Maximiliano: That's great.
Dr. Deza: But if you don't know in advance,
Maximiliano: you win. I am I am getting scared on how good companies are getting on keeping customers on subscriptions.
Dr. Deza: Yeah, it's I am getting very scared on that.
Maximiliano: Yeah, the problem is um people like me are part of the problem. I mean people like me will be part of the problem as well you know like I'll I'll make everything to keep people buying my products and stuff like that.
Dr. Deza: Um but but yeah but sometimes I mean there are like a in the vending machine company for some people like us use this dark patterns.
Maximiliano: Okay.
Dr. Deza: To make people forget they have a subscription and then that that's not what they do. Okay. So some people like basically that's that's not good. That's not not
Maximiliano: Yeah.
Dr. Deza: Some people, for example, you download an app, you have to subscribe, they they they ask you for, you know, 50p. You pay 50p, but then actually it was a subscription for 50p.
Maximiliano: Yeah.
Dr. Deza: And then you don't realize uh only two years later that you were being paying 50p a month for to those guys for and that things are are are usually they rely on people not wanting to spend 20 minutes trying to unsubscribe and just keep paying.
Maximiliano: Yeah, that's that's not what I do. But I what I do is try to understand the people, try to understand what move people and try to provide a better service so they don't want to to subscribe.
Dr. Deza: That's what I mean when I say like I want if I am part of the problem in that sense
Maximiliano: I want people to keep buying my products.
Dr. Deza: Exactly.
Maximiliano: Uh just so just so you know guys I don't have any products being sold within this podcast. Okay. relax from that. Uh it's just I got a small vending machine uh business and I would I would do anything to make people keep buying basically mainly because I want to make sure I want I have the right service for them and is helpful for them and so on. So it's not something that like
Dr. Deza: absolutely
Maximiliano: yeah I wouldn't do like I don't know some weird stuff. The thing is uh those kind of behaviors don't uh scale. You can scam people for a bit but then you'll be known as a scammer and then you cannot do any any time. You have to be like a twisting uh changing names, changing things.
Dr. Deza: I mean some people are always going to do it
Maximiliano: but that's not uh the way at least I see uh growth.
Dr. Deza: That's I mean at least that those are my values. Yeah. Well, if if we iterate it infinite infinitely, this is how I see it. If if my behavior today uh infinitely iterated, so infinitely either getting better or adjusting depending on the market would be good in the future. So in the infinite future so to say, then I'm going to keep that behavior. If it's not then you know and lying, cheating, stealing etc.
Maximiliano: doesn't work on the long run.
Dr. Deza: So well for me it was more about research this about the churn analysis. I liked it. I I learned a lot. I learned that for example you could not not only one month before I could know if they were were going to turn a year before
Maximiliano: really
Dr. Deza: yeah because it wasn't about and there wasn't the main it's a paper we wrote it wasn't about the behavior they had before turning it was about the behavior overall how they were using the product
Maximiliano: oh Sorry, what company in particular?
Dr. Deza: It was um um basically to subscription to books uh was institutional. So they had this books uh online and then they gave the access to for example you ask a university or a company you pay them monthly and then you could download all the books you want and depending how many books they were they were downloading and how many people were doing it and how in which how
Maximiliano: you could basically see uh a pattern and then we found that people who basically um they had their loads were in in a certain way.
Dr. Deza: Yeah.
Maximiliano: Uh where basically it wasn't worth it for them.
Dr. Deza: All right. Okay. That's that's actually quite interesting. Do you reckon uh we could see that in like gym memberships and other
Maximiliano: Well, of course, somebody who doesn't use the gym is going to stop the subscription at some point.
Dr. Deza: Well, most gyms actually survive thanks to people who don't go. So we need to be thankful to those guys because u otherwise Jim's description would be much more expensive. Thank you very much for all people who decided in November that this year was going to be their year. Thank you very much because our gym memberships are cheaper. Thank you.
Maximiliano: Okay. Well then about research I can I can keep talking but basically it's a a nongon of project small projects project with um some project with students some students come up with some problem they have and then we sit down together and then we we publish a paper or we do something usually papers are help them to get a job or to further their careers and I'm involved in that and I help them sometimes there are big projects I work with another teaching staff or companies. Uh, one more I'm talk I'm working on diagnosing cancer.
Dr. Deza: Damn. Okay. I I remember you telling me about it. I thought it was incredible because I I've worked in the space of charity and it's it's just amazing. Every time I had to research for a particular charity, specifically the cancer ones, I was amazed by the type of studies that the there were, it's it's just amazing.
Maximiliano: There's a big project. um, and I'm basically the statistician, the guy who does the runs numbers. um, there's medical staff, there's people who do the analysis of the of the samples.
Dr. Deza: Yeah,
Maximiliano: there's a lot of different moving pieces. I don't even understand all the moving pieces. I don't know who the project manager for that is, but it's
Dr. Deza: praise to him.
Maximiliano: And then
Dr. Deza: or her. Yeah.
Maximiliano: You get to the I I get basically a data set.
Dr. Deza: Yeah.
Maximiliano: And I need to analyze it and it's like a nightmare data set. is the worst kind of medicine you want
Dr. Deza: oh
Maximiliano: which is I only have 200 people because they cannot
Dr. Deza: oh
Maximiliano: they in medical studies they cannot do a million people analysis you cannot like basically play with lives you need volunteers so finding people is very hard and then uh once you analyze all the the samples you get hundreds of different variables that can go long. So it's like the one of the worst cases. Uh took me a long time like a months to learn enough about the problem to be able to tackle it and then study it and then implement my things. I I I tried many types of machine learning and AI and things into it. Finally after months of work I was able to get um almost 90% of accuracy on detection of of the samples and now we are
Dr. Deza: almost almost 87
Maximiliano: this is on diagnose of cancer
Dr. Deza: uh prostate cancer yeah
Maximiliano: prostate cancer
Dr. Deza: actually um I don't know well for for the viewers out there uh the two toughest cancers mainly I mean the two cancers that kill the most people are pro prostate
Maximiliano: and uh is it uh
Dr. Deza: breast cancer?
Maximiliano: No idea. Maybe I
Dr. Deza: think uh those two are the the most common and the ones that kill the most people. So basically that's the that that's a job I it's a lot of mathematics of love of computer science a lot of trying different algorithms trying different things trying beta data driven so trying to very very pragmatic so I can try all these things and see what works and then try to understand why it works and then try to make it better and then try to like a very
Maximiliano: iterations
Dr. Deza: iterations and iterations and iterations um so like a business maybe but uh trying to to get the goal of getting the best possible and then I I was there's also a dose of luck to be and I I was able to
Maximiliano: there's a lot of luck involved
Dr. Deza: I was able to
Maximiliano: I wouldn't call it luck that's the thing because you have to study a lot I wouldn't call it luck I would call it uh maybe uh I don't know like instinct maybe you see there's I I like this quote from Thomas Edison in light bulb guy who said that inspiration. No, he said luck luck exists but he it has to find you working.
Dr. Deza: Yeah. Yeah. What a great Yeah. Yeah. Yeah. I've I've always Yeah.
Maximiliano: So it was that. So I
Dr. Deza: Well, if if luck exists, it better catch me working. Yeah. So I I I spent many many uh days and months working in something not not getting results and after I was expecting something like 80% 80 something 80 and then when my results like really went over my expectations and I I'm very happy that we're going to publish very soon and so this is like unpublished results. M. Oh, okay. You got the information for free, you little bastard.
Maximiliano: Just so you know, I tried to put exactly the same information before him and published before him.
Dr. Deza: No, no, I'm kidding. So, that's uh if we're So, yeah, I'm I'm very happy it went over my expectations and now uh this hopefully can get into something that we can be used to predict. uh cancer. Uh that's that's so great. That's so great. You know, something that I really would like to do. So the I would call it um uh Jim Ron I think it is is a speaker like a motivational speaker and so on. He said um you should um you should try to get rich the first half of your life and try to spend it all the last half. That's that's the idea. I think I would do that. I think I would do that. And the reason why I will do that is because of the charity's goal. This is what I love about charity as a general thing. Instead of having a uh revenue the or a profit,
Maximiliano: you don't work for a profit. You work for a result within society.
Dr. Deza: Exactly. So the profit is the improvement of society. So in the case of uh charity water, shout out to them. They're great. If you if you want to research about them, that they're great. They actually uh put all the money that you uh give to them into uh the their projects.
Maximiliano: Uh of course what they do is it's always related with water. normally in uh Africa in in
Dr. Deza: they're the ones who dig holes right in the in the desert.
Maximiliano: Yeah. Yeah. Well, they do a lot of stuff. Not not only that, but like they they secure the water, they uh put irrigation systems, etc., etc. Um anyways, so um they them for example, their profit would be how many people have they actually helped
Dr. Deza: which is great. or how many wells have they digged or whatever is which is amazing. That's what I love about charity.
Maximiliano: The thing is you don't really need to put always money on top of your priorities.
Dr. Deza: No.
Maximiliano: For example, as a scientist I I don't
Dr. Deza: Yeah. Well, as a scientist you don't. Yeah.
Maximiliano: I I measure myself as how much I help, how much I learn. uh can be routine as publications, it can be done as collaborations and can be done many ways but usually and there there are many other ways to do it.
Dr. Deza: Um and I think it works because it creates collaboration. So if uh the only thing I want is money and you want money, we're going to be in competing.
Maximiliano: Yeah. Yeah. So, not not always because I can be your your you know uh you can work I I've been giving you some things for your business and so on and so forth. It can be a a supply line or something supply chain but
Dr. Deza: uh if the only thing I want is money sometimes you cannot create meaningful collaborations the way if uh you have a charity for example that wants money for that and then you get for example a politician who wants um um who's a able to harness money but wants I don't know votes and somebody who's um and so forth and gets more collaborative.
Maximiliano: Yeah, I I do know I do know what you mean. Uh I feel okay so this is the problem. Legally businesses are required to think about profit
Dr. Deza: of course
Maximiliano: and that as a major goal of the comp a company is basically made to make money money making and that's that's all the that's it can it can be not it's allowed to not be the only focus but it has to be the main focus.
Dr. Deza: Yeah. But but you you see what happens when when companies focus on that that they start extreme competition degrading the the environment uh the workers rights and because I mean it can't be the only thing I mean I'm not I'm not against profits I'm not
Maximiliano: I know I I know actually in my opinion uh thing is I would say that would be the government's It's the government's job
Dr. Deza: job to limit you
Maximiliano: to limit to put the rules.
Dr. Deza: Yeah.
Maximiliano: Nothing else. That's that's the whole point of the government. The government should be there to
Dr. Deza: But then if if the government was was there was a job then the government should be overwhelmingly more powerful.
Maximiliano: Yeah.
Dr. Deza: Than the companies.
Maximiliano: Yeah.
Dr. Deza: Otherwise companies will be able to overstep and So that will mean a lot of things are not happening. So because right now companies are not overwhelmingly less powerful than companies. Governments and companies are at the same level. And also uh something that does happen quite a lot is that uh the benefit of the people is not visually like that when the government puts that policy. So for example the CO2 uh tax in Europe People don't like it because when they put the CO2 tax, all their expenses go up.
Maximiliano: All of them. Literally every single one of them goes up because every company will emit CO2. So everything goes up but the government technically is making sure that future generations so maybe not uh the population now but all the future generations live in a society with less CO2
Dr. Deza: or not
Maximiliano: oh yeah well
Dr. Deza: but but that's that's a whole that's the whole point that's a problem on regulation and that's
Maximiliano: but that's the problem of the government and that's why I think government should always regulate And it's good to have the competition.
Dr. Deza: Yeah. But the problem is that
Maximiliano: mainly towards that
Dr. Deza: because then you know what what's going on. The problem is that from a physics point of view companies and our complex system.
Maximiliano: Yeah.
Dr. Deza: So they just imagine a plant but ramifications and things. It's very complex.
Maximiliano: Yeah.
Dr. Deza: And governments uh are simple. They just give simple rules.
Maximiliano: Yeah.
Dr. Deza: they cannot go case by case basis because that will be impossible and unfair. So basically uh if you have a plant and uh the only thing you can do to a plant is you can put it in a warm place in a cold place in a sunny place in a in a shady place. You can water it more. You can water it less. You cannot do many other things to your plants.
Maximiliano: Okay. You can maybe put some some something in the earth, something in the water, but it there are simple things.
Dr. Deza: Yeah.
Maximiliano: You cannot go uh at the cellar level to and just fix one cell.
Dr. Deza: Exactly. You can't you can't fix one tower at a time. You can fix the whole thing and then the plant which is an extremely complex system like uh is going to adapt to that.
Maximiliano: Yeah.
Dr. Deza: So uh and And and that's a big problem with with governments that they are not able so you cannot uh explain a complex system in simple terms. You cannot have simple solutions to complex systems. Those are complex problems.
Maximiliano: So this is this is where my view of the government uh goes a little bit in into your side. I would say there should be a difference between uh laws and policies than regulations. I really like federal states mainly because you got basically two or maybe even three like in Germany uh because then you got the European Parliament that gives you uh the policies the overall policies.
Dr. Deza: So where is the future kind of going? Then you got the governmental which is laws.
Maximiliano: And then you got uh federal law which is regulations technically. So basically it goes deep into a system and you can actually make those tweaks. Um now I'm not saying it's is perfect and so on but and it will take time and that's the whole point.
Dr. Deza: I I hope you're you're right. I what I see now is that
Maximiliano: but the more policy makers that we do have in different categories uh the better.
Dr. Deza: Yeah. Well, I I think that a bit of of self-regulation would be would go a long way. So, but that's my opinion. I am not an expert.
Maximiliano: Yeah. Well, me neither. Me neither. just I just think we've been talking a lot about our data analysis and then we went into politics a little bit but um if you would put it down to let's say a sentence or maybe two or three what is data analysis and why is it helpful? uh data or data science.
Dr. Deza: Let's go for data science first. Data science is more than data analysis. Now this is something everybody does. Everybody does. You go um you see your expenses in your shopping list and you say oh actually I'm buying too much milk so I'm going to cut it off because I'm not drinking it. So that's late analysis.
Maximiliano: Okay.
Dr. Deza: Okay. Okay.
Maximiliano: So uh can be done in a very basic level. can be done on a huge level.
Dr. Deza: I love that you're explaining it simply because I yeah it is what it is. Data analysis is analyzing data and everything is data. Okay. So when you're I mean you're crossing the street and then you see a car and then you're analyzing in your mind are I able to cross the street before the car comes or I need to wait until car passes. But that's it's
Maximiliano: and what's saying science then?
Dr. Deza: Science is uh basically um comes from three different things.
Maximiliano: Okay.
Dr. Deza: Okay. So the first thing it comes is basically statistics. So understanding big data sets we're using statistical tools like um I know all all these tools statistics. Then the other part is uh machine learning and AI. So using uh new tools to understand um get more insight from the data.
Maximiliano: Yeah.
Dr. Deza: And the third uh is uh being able to deliver value and that means usually human skills vis visualizations, dashboards, presentations and being able to sit down with somebody who's not able to understand the data because of lack of statistics or lack of time and being able to deliver value. So I'm going to tell you three things about your business and I'm going to make you understand why you are underperforming. this this and this
Maximiliano: and then the solution to the
Dr. Deza: the solution is probably depends but maybe is something I don't know the solution maybe I
Maximiliano: okay
Dr. Deza: okay sometimes it's obvious
Maximiliano: and sometimes complex as hell yeah
Dr. Deza: as a data scientist I need to be able to uh get the technical part well the statistics then I need to understand all the new technologies and uh which keep changing And I need to be able to have um enough understanding of human to be able to sit down and tell somebody um what's something very complicated into something simple
Maximiliano: and a way of delivering value,
Dr. Deza: right?
Maximiliano: Which is uh which is the ultimate goal of this how to extract. So go from
Dr. Deza: information and deliver
Maximiliano: from noise to signal.
Dr. Deza: Yeah, I really like that. I really like that expression. Noise to signal ratio. That's what I I we all do as designers from noise to signal. So you get all the the data I imagine I know you have an online uh marketplace or something and you have people buying people selling but whatever and you don't know if you are doing well or not.
Maximiliano: You money is coming in money is coming out.
Dr. Deza: You don't know if you're profitable. You know, maybe you you know you're profitable because you're I mean money is ramping up in the bank, but you don't know if you are as profitable as you could be. You don't know what to do in order to be more profitable. You don't know what to change. You don't know are people happy with my system or or are people slowly uh bleeding into my competition.
Maximiliano: Basically, it's how to look in a little bit in the future.
Dr. Deza: It's uh there there are three steps basically on data science. The first is trying to understand what's going on. The second is trying to uh predict.
Maximiliano: Okay.
Dr. Deza: And the third is uh being able to take action to make decisions.
Maximiliano: That that is actually great that okay. So it's basically taking a picture of the current problem whatever it is. Uh then looking at future possibilities and then the actionable step
Dr. Deza: and being able to make action
Maximiliano: making
Dr. Deza: uh about action depends on on many different things
Maximiliano: on many different things and the
Dr. Deza: the complex final say what's going to be management
Maximiliano: management needs to be informed and that's basically the job of a data scientist in the long run I usually don't work in uh at that level of companies that usually is employees employees do that
Dr. Deza: but I I advise I consult
Maximiliano: and I help them to understand their problems or their solutions.
Dr. Deza: Yeah.
Maximiliano: Yeah. Yeah. Yeah.
Dr. Deza: I feel both are very good. So all right. Um and why did you decide to become a senior lecturer in uh within all this?
Maximiliano: Well, I always like love not to teach. Teaching is uh
Dr. Deza: like what what made you actually say, "Okay, you know what? Physics, let's go into
Maximiliano: But it's not it's not static."
Dr. Deza: Oh, sorry.
Maximiliano: I'm I'm doing I'm doing physics all the time. So, it's basically there's not a big difference between what I'm doing and statistical physics. It's the same thing.
Dr. Deza: Well, yeah. Right now, for example, right now as a researcher, as a consultant, I'm I specialize on using physics tools to uh tackle non-physics problems which is something I I can do because I I learn physics, I know a lot of tools, I know the mathematics, I know all the things and other like data scientists come from management, come from business, come from mathematics, come from other things. So, I have tools they don't have. I mean, of course, they can learn them, but at least it's my edge.
Maximiliano: Yeah.
Dr. Deza: Well, that the learning of that, I'm guessing it'll take a couple years.
Maximiliano: Well, it's my edge. So, I I usually go for very complex problem. So, a company comes with a very very difficult problem or researcher or collaborations and they say, "Okay, we have this problem and we don't know how to fix it. uh can you take a look and take a look and see what happens and then I I need to to study I need to learn I need to to go into a topic and finally uh when it's done uh it's a it's a really really good
Dr. Deza: sensation
Maximiliano: but then why did you not uh keep I don't know started as a consultant or or like keep going as in a consulting career why do you decide to
Dr. Deza: Well, I I am I am a consultant.
Maximiliano: Oh, okay.
Dr. Deza: As a consultant. Yeah.
Maximiliano: Um
Dr. Deza: but but then why did you decide to teach? It still doesn't
Maximiliano: because teaching gives you
Dr. Deza: what is it about teaching?
Maximiliano: Teaching helps you a lot to first of all there's let me put it this way. You you gain from two sides. Okay. First of all, you are giving back. You are giving teaching people next generation how to think, how to do and you're giving an good impact to the world. Okay.
Dr. Deza: Second, you are actually meeting the next generation of professionals.
Maximiliano: So, you're you're networking.
Dr. Deza: Yeah. Technically, yeah.
Maximiliano: Okay.
Dr. Deza: Third, you are able to influence them. You're able to give them the tools they want. you're able to be uh somebody to able to to okay and finally uh the fact of teaching makes you study. So, it's the best way to keep uh
Maximiliano: on top of everything updated
Dr. Deza: because you're going to fall into conversations like this when people are going to be oh why have you done this? I'm using this technology and you never learn about this technology because everything the world is changing sultaneously and rapidly and
Maximiliano: every time it's also way to to be there but also because I like it. It's nice. It's you really learn from people. I learn from them as much as they learn from me. And at the end of the day, I I enjoy it.
Dr. Deza: Yeah.
Maximiliano: So, but but yeah, I I work on I I do research for for scientific purposes. I work with companies and I work on uh with um university projects with other researchers.
Dr. Deza: A lot of things.
Maximiliano: Yeah. Yeah. Well, you do a lot of things at once basically.
Dr. Deza: Well, less than you.
Maximiliano: Sorry.
Dr. Deza: Less than you.
Maximiliano: N it feels like you you're overwhelmed with all this.
Dr. Deza: No, it's like more like uh it's a very social thing. You need to talk to people. Sometimes it's just a very tiny conversation, but it's a spark. And then six months later, that spark generated a fire. And then maybe you get a collaboration.
Maximiliano: Yeah. You know, you know what's funny? Uh I know I know it seems like I do uh quite a lot, but in my opinion, I don't do much because I mainly think and I outsource most of the problems that I have. So for example, for the podcast, I don't edit. I don't do the social media. I don't even uh now book the actual uh podcasts.
Dr. Deza: Yeah. But I do the same. For example, I I work with students.
Maximiliano: My thing is just to think how the whole thing actually goes.
Dr. Deza: I I work with students. I collaborate with other researchers. I work with um workers at companies where I consult. I'm I'm not doing everything. I'm not
Maximiliano: Yeah.
Dr. Deza: driving the bus and like
Maximiliano: Yeah.
Dr. Deza: clean cleaning the toilets. I
Maximiliano: Yeah. Sometimes sometimes in in uh in doing a startup you actually have to do it which is funny. I love those moments though. You're like stressed cleaning the toilet or whatever.
Dr. Deza: But but but yeah that's that's life. It's life. Yeah.
Maximiliano: So it's very similar to being an entrepreneur.
Dr. Deza: Yeah. Yeah. I guess I guess
Maximiliano: and in some cases it is literally being an entrepreneur. But uh in other cases also helps you a lot to meet people and it gives you um helps a lot for for understanding. So you go to people ask you what what what do you work at? You say I'm a senior lecture in data science.
Dr. Deza: Already defines you a bit.
Maximiliano: Yeah.
Dr. Deza: It's not like oh I'm doing data science know what to do and also helps you a lot. So the university job gives you a lot of um uh it's not how you call it like uh something everybody can relate. So they can tell you okay it's there it's an expert.
Maximiliano: Yeah. I feel that's the big thing that I that I have a struggle with to define what I do. Every time somebody asks me, oh, so uh what do you do?
Dr. Deza: I uh sometimes I say I'm unemployed. Sometime I say I'm an M andd of a company. Sometime I say I got a podcast. Sometime I just say I work in a pop, you know. And it's it's very interesting how to just like
Maximiliano: because I'm not defined because of the entrepreneur. The problem is this. People are going to define you no matter what.
Dr. Deza: Yeah.
Maximiliano: Unless so the easiest way is to give them a definition of yourself.
Dr. Deza: Otherwise, they're going to find a definition of you which may not suit you.
Maximiliano: Yeah.
Dr. Deza: So, it happens all the time. So,
Maximiliano: yeah.
Dr. Deza: So, that that's that if you're a CEO of a company, which you are, uh you can you can say that.
Maximiliano: Sure. Yeah. If you are are a bartender, which you are, you can say that. So it's it's it's not lying. It's more
Dr. Deza: Yeah. Know just say whatever thing
Maximiliano: or just define yourself.
Dr. Deza: Yeah.
Maximiliano: Because people the world's very complex.
Dr. Deza: Yeah.
Maximiliano: And people and this is going to come as a theme in a moment.
Dr. Deza: Uh people need um to feel they understand things. People need uh clarity
Maximiliano: and if you don't give them clarity they are going to just
Dr. Deza: get the easiest way out of the of the complexity.
Maximiliano: Yeah.
Dr. Deza: So I if if if I refuse to to to give you what you want from me,
Maximiliano: you're gonna invent it or u assume it
Dr. Deza: or something. If we ever I mean keep talking and become friends and everything, you all gonna know me in all my complexity.
Maximiliano: Yeah.
Dr. Deza: But that's something that takes time.
Maximiliano: Yeah. But but that's not the first.
Dr. Deza: But the real reason you're going to want to spend time of your life to know somebody else is that that person gives you something that you feel is interesting or it's nice. because of the first impression, you will unlock, so to say, the other ones.
Maximiliano: Yeah. Well, the good thing that that I've Yeah. Yeah. It's it's it's a it's a good point. It's a good point.
Dr. Deza: I mean, it's not my point.
Maximiliano: Yeah. No, no,
Dr. Deza: at least I
Maximiliano: I feel it's a very analytical point. It's it's very related to data an analysis because
Dr. Deza: I mean, it's a it's a physicist point of view anyway.
Maximiliano: Yeah. Absolutely. like you got to either define or
Dr. Deza: but that that's about
Maximiliano: it will be defined for you.
Dr. Deza: It's about complex systems and complex systems are society is a complex system.
Maximiliano: Yeah. Well, people are complex system
Dr. Deza: people society is a complex system of complex systems.
Maximiliano: Yeah.
Dr. Deza: And maybe three or four more of those in the chain.
Maximiliano: Yeah.
Dr. Deza: Until we get to the atom.
Maximiliano: Yeah. Yeah. It's a good point. Well, um so uh I just want to uh go back to to the student and your relationship with uh them.
Dr. Deza: So how how do you actually uh work with them?
Maximiliano: What is your depends relationship with
Dr. Deza: I teach a lot of master students.
Maximiliano: Okay.
Dr. Deza: So are grown-ups? Yeah.
Maximiliano: So you don't teach 18 year olds?
Dr. Deza: I I do. I do teach also but but I
Maximiliano: Okay. No, but I I teach uh people who are grown-ups.
Dr. Deza: Yeah.
Maximiliano: Uh and I teach them uh data science which is not I'm not teaching like to say ABC or
Dr. Deza: Yeah. It's not like
Maximiliano: so it's not a like a
Dr. Deza: basic
Maximiliano: basic things they are and then of course you need to respect them as people. You need to respect them as professionals. Some of them have professions.
Dr. Deza: So maybe they are ch changing careers or they want to add data science to their curriculum. Maybe they are developers or skills or programmers or
Maximiliano: Okay. Some of them have skills I don't have.
Dr. Deza: Okay.
Maximiliano: So I you need to be very respectful of them
Dr. Deza: and uh basically what they do is to engage in conversations.
Maximiliano: Uh where I try to teach what I want what I need have to teach to teach them but at the same time I need to be able to to adapt it to what they do. So it's like a I have my curriculum, I need to teach this, but sometimes they need to learn it in some way.
Dr. Deza: Yeah.
Maximiliano: Because of and sometimes it really changes from person to person. So needs a bit of a
Dr. Deza: That's what I was going to ask. Have you seen like a different like a big difference in ways that new generations are acquiring knowledge
Maximiliano: or how you have to treat them or so on so?
Dr. Deza: Well, the thing is the world is changing. That's basically the main thing. So, well, I was telling you before that when I was a a student, there was barely not internet. Internet was some curiosity. A very small thing like
Maximiliano: Yeah. You had to go and like read.
Dr. Deza: I I Yeah, exactly. You had to go to library.
Maximiliano: You to learn. You have to go to a book. go to library or use encyclopedia or in carta CD or something like that.
Dr. Deza: Uh I mean yeah
Maximiliano: rent rent CDs rent documentaries to to understand the topic.
Dr. Deza: No or wait for them where wait for them to come at the hour exactly as the cable TV.
Maximiliano: Yeah.
Dr. Deza: So you couldn't go on demand like now on Netflix oh I want to watch it now.
Maximiliano: So you had all the those things. So information was precious was something you people if if you were able to know something if you learned something
Dr. Deza: was very elitist.
Maximiliano: No no no elitist but it wasn't so is it wasn't like you didn't have to pay for it but you need to spend time. Yeah, you would have to go more into the
Dr. Deza: so libraries were free. You you my my schools books were available mostly and well depending on the topic but mostly available
Maximiliano: but you need to go there and spend hours reading until you understand the things. So information was uh valuable you if you go oh do you know this and this and this and this it meant that person had spent time learning this.
Dr. Deza: Yeah.
Maximiliano: uh then when I was basically finishing my my BSE my bachelors
Dr. Deza: internet became more common place
Maximiliano: and by the time I was
Dr. Deza: how was the switch like though like I know that the information was the major point so
Maximiliano: at one point you couldn't access to information at all points in your life
Dr. Deza: and now
Maximiliano: it was very very gradual
Dr. Deza: because internet at the beginning was very bad
Maximiliano: it Okay. Yeah. Yeah. It
Dr. Deza: was very bad. I mean I I was there from the from the beginning. I I remember exactly when when
Maximiliano: I was there from um things we didn't have phones. So I remember like my mom getting iPhone one. Well, I'm I remember when internet came to be.
Dr. Deza: I mean 19 I remember
Maximiliano: we I knew there was something called internet
Dr. Deza: but but by the time internet became something you could buy. You could have
Maximiliano: uh you to it was dial up on your phone. You had to phone to the company to get dialup
Dr. Deza: was extremely slow. I had my my first
Maximiliano: Windows Windows 1990 199
Dr. Deza: yeah 3.1 I had my my first email account was uh 1995 uh
Maximiliano: all right
Dr. Deza: yeah the first and I was one of the first guys who had email account was like I have an email account
Maximiliano: right
Dr. Deza: that was fantastic and it was so then maybe 10 years
Maximiliano: MySpace and all Yeah. Yeah. And all all the other ones who died.
Dr. Deza: Yeah.
Maximiliano: But that was that was years later.
Dr. Deza: Great great pictures at MySpace though. Great pictures.
Maximiliano: Many others by the time I I maybe 10 years later by 2005 more or less.
Dr. Deza: Um well 2008 I I got my first my Facebook account.
Maximiliano: YouTube started 2008. Okay. Yeah.
Dr. Deza: Uh YouTube started. I remember that.
Maximiliano: Yeah.
Dr. Deza: And
Maximiliano: it was an app. It was a Yeah.
Dr. Deza: Yeah. It wasn't It was in Google.
Maximiliano: Yeah. It wasn't in Google. It wasn't an app. Yeah. Yeah. Yeah.
Dr. Deza: And there wasn't There was no apps. There was
Maximiliano: You had to install it in your computer. Yeah.
Dr. Deza: No, it was a web page.
Maximiliano: Yeah. Yeah. Yeah.
Dr. Deza: And you had to this and then when I was doing my PhD, finishing my PhD or
Maximiliano: uh there was uh finally smartphones.
Dr. Deza: Okay.
Maximiliano: Uh I had a phone since 2003
Dr. Deza: like a like a dumb phone like Okay.
Maximiliano: Alcatel thing and
Dr. Deza: Nokia.
Maximiliano: Yeah. No fail. Yeah. And but then my first smartphone was 200 10 maybe. I I didn't want to have a smartphone back then.
Dr. Deza: Okay.
Maximiliano: So, and iPhones were in 2008, I think.
Dr. Deza: I don't know. Yeah.
Maximiliano: Well, but but the the fact is that the world changed.
Dr. Deza: Yeah. right now. After that, if you want with a bit of trivia, oh, do you know that this people say, "Oh, let me check it out." Oh, yeah, you're right. Oh, no, you're you're not right. You could verify or you could find Google things in seconds. uh, something that would have taken you hours before.
Maximiliano: Yeah. Or even weeks.
Dr. Deza: Well, it depends. I mean, you can go to the encyclopedia and probably it's there. Or there were books of trivia. You could buy
Maximiliano: books of trivia. Oh my goodness. Damn.
Dr. Deza: Okay.
Maximiliano: And
Dr. Deza: so information changed. The access to information changed. Yeah. And therefore the value of the information also changed.
Maximiliano: Yeah. It went down. It went down. Was less important to know.
Dr. Deza: So if you if in the 90s you say, "Okay, I read this book." uh you could have a conversation with somebody who had read the book and enjoy conversation and have and that gave you a bit of a status about you know and reading a book intellectual or some something like that.
Maximiliano: Now um
Dr. Deza: well now nobody thinks you as an intellectual if you know things.
Maximiliano: Exactly. Because everybody can go and especially now with AI, you can go to chat GPT and say, "Okay, please tell me the summary of this book."
Dr. Deza: And and then you read that summary.
Maximiliano: Read the summary.
Dr. Deza: And then you can say, "Oh, yeah. I read the book, too."
Maximiliano: And how do you recognize the it has because now we're tilting basically in in in AI. What do you reckon this change is going to bring? Because we just had uh well as we just agreed the information was the biggest change now.
Dr. Deza: So
Maximiliano: it's the same. So AI is a continuation of the same
Dr. Deza: and
Maximiliano: is it taught to information?
Dr. Deza: So what we call AI
Maximiliano: is not real AI.
Dr. Deza: Yeah, I know. AI would be artificial intelligence. Uh what we would assume is artificial human intelligence
Maximiliano: which doesn't exist.
Dr. Deza: Okay. So
Maximiliano: and what do we actually have now? Then
Dr. Deza: imagine this. Imagine you have uh you're at school back at school. Okay. And teacher comes and gives you two pages to to read. Okay.
Maximiliano: Then there's one student uh who goes and reads the whole thing 10 times.
Dr. Deza: Yeah. and is able to cite it, is able to summarize it, is able to make everything but has no idea what it says. So has not hasn't doesn't understand the text. Okay?
Maximiliano: So maybe I know some history text. The king this had this war and this happened, this happened and this happened
Dr. Deza: and just can repeat exactly the whole text.
Maximiliano: Okay?
Dr. Deza: or can can make slight changes but it's basically that and then have you know a student who actually reads it a couple of times and then he understands the motivations understands why this king invaded the this happened
Maximiliano: what effects did it could it have
Dr. Deza: yeah so it has a more deeper understanding of thing but it's not able to to summarize it is not able to site it verbally vervatin it's not able to to do the
Maximiliano: the the thing
Dr. Deza: word by word thing Yeah, you can't. Uh, so the AI would be the first
Maximiliano: Aon current AI is able to get the text, is able to manipulate text to make it shorter, make it longer, but it doesn't understand
Dr. Deza: the full
Maximiliano: the the text
Dr. Deza: in depth.
Maximiliano: It doesn't understand the text.
Dr. Deza: It's just a it's a the thing is
Maximiliano: and what do you reckon? What? Okay. Yeah. No. So basically humans we have two modes.
Dr. Deza: The first mode which is the principal mode
Maximiliano: okay
Dr. Deza: is what we would call uh
Maximiliano: information
Dr. Deza: it's um reacting to stimuli.
Maximiliano: So if and the second mode would be formal education which is basically this kind of learning memorizing wrote
Dr. Deza: all right
Maximiliano: and then bring it back. But if you get somebody who never had any type of education like maybe 100 years ago was totally normal to not not to go to school.
Dr. Deza: Those guys weren't weren't dumb at all. No,
Maximiliano: they were extremely intelligent. They were able to live their whole lives, marry, get kids and do things and make businesses
Dr. Deza: and create build houses. uh and never engaged in anything similar to memorization or learning.
Maximiliano: I actually want to put uh a small story there. One of Warren Buffett's greatest business partner uh was a ex-s Soviet uh woman uh sorry a woman from the Soviet Union who fled the Soviet Union or came came to the America uh in uh well started a a business for like furniture or something like that. She made the most profitable company within uh Warren's Buffett's portfolio. Incredible. And she did not know how to read and write at the beginning. And she didn't know anything about accounting, which is the opposite of what you
Dr. Deza: you need to have to.
Maximiliano: Yeah. That that would be like the logical step. And that's what they teach you now that you should learn to read, right? Of course, but also know accounting. And this woman did the most one of the most successful companies uh within Warren Buffett's portfolio. So yeah,
Dr. Deza: so and computers cannot do anything of that. They cannot learn from errors. They cannot find their way. They don't have any motivation. They don't
Maximiliano: okay
Dr. Deza: know things. Okay. and just basically get the information and like vomit it out.
Maximiliano: Exactly. So it's not it's not really intelligence in the sense even a an animal has more intelligence. Animals a fly is able to go from one place to another looking for food.
Dr. Deza: Damn. Really? You would actually say that fly is more intelligent.
Maximiliano: Yeah. Flies have motivation. They want food. They can fly. They can find the food. They can find a place to to lay their eggs. they can't find a mate. They You see what I mean? They
Dr. Deza: even now with the new developments in the thing is I've read a couple of articles saying that um 80% of the time in an experiment where basically AI uh was asked to turn itself uh out or like off. uh 80% of the time they said like they try to keep uh be kept alive so to say.
Maximiliano: Yeah. But that that's that's not surprising. Uh because
Dr. Deza: would you wouldn't you reckon that's motivation?
Maximiliano: No. No. No. Because it it it's not it's not in the training.
Dr. Deza: Okay.
Maximiliano: Okay. So um humans all all the the data they they're fed with or most of the data they're fed with is created by humans and we
Dr. Deza: hate we hate to die.
Maximiliano: Okay. Good point. Good point.
Dr. Deza: Okay. So it's basically repeating patterns.
Maximiliano: You see what I mean?
Dr. Deza: Okay. I see I see I see what you're doing. I see what you're doing.
Maximiliano: Oh, and probably Yeah, it it was in within a company the the experiment. So, I'm guessing turning it off would be the same as
Dr. Deza: either killing it or well, in the case of a human firing it.
Maximiliano: So, yeah. Okay. I see. I see what you mean. I see what you mean. So, uh there's a lot of hype. There's a lot of because I know saying uh AI has gone rogue and putting a picture of Terminator is going to warranty views and clicks on on your blog.
Dr. Deza: Yeah.
Maximiliano: And that's true. So there's a lot of hype around around that. But basically um computers as they are working now don't even have the mechanism to to think.
Dr. Deza: Okay. That that gives me a little bit of an ease. Uh it does it does real mechanism. We call it reinforcement learning.
Maximiliano: Okay.
Dr. Deza: And uh it's used mostly in robotics. So in robotics you basically instead of
Maximiliano: CH GPT you're just giving them all the text of humanity
Dr. Deza: and uh it's basically
Maximiliano: concentrated and makes a database of all that is able to learn semantics or the meaning of things able to catalog everything. So when you ask it something, it's going to be able to very very fast find um relevant information and give it back to you in a uh very articulate.
Dr. Deza: And why does it invent information? Why does it create information out of their name? Basically
Maximiliano: because it's never been programmed to say I don't know.
Dr. Deza: It's it's unable to tell you I don't know.
Maximiliano: Oh my god. Oh, those little bastards in Chad GBT, stop start stop actually doing that, guys. They they can't they can't because they they not they're unable. So if you ask
Dr. Deza: what do you mean they're unable
Maximiliano: they are incapable of telling you I don't know.
Dr. Deza: But why don't they program it to
Maximiliano: because there would be bad business.
Dr. Deza: That ising crazy.
Maximiliano: Okay.
Dr. Deza: That is crazy.
Maximiliano: So
Dr. Deza: Oh my god. There's going to be a clip for sure, man. What the ask it ask JPT anything anything uh which it doesn't know and it's going to invent it. So wait, wait a second. You're telling me that Chad GBT is not programmed to say I don't know only because it's bad business to say I don't know.
Maximiliano: I mean, what other thing would be?
Dr. Deza: Oh my god, you're I'm so sorry if you get killed because of saying that. But
Maximiliano: I mean I mean it's me it's not me saying it. is like there's a lot of of other AI experts saying
Dr. Deza: okay
Maximiliano: it's not it's not something I mean if you ask it something is going to make it up.
Dr. Deza: Yeah. Yeah. Yeah. But but I I just never un quite understood why but yeah. Okay. Oh my god that's terrible.
Maximiliano: It's it's a it's a business.
Dr. Deza: I know it is but it's just crazy. Okay.
Maximiliano: I mean the thing is people are treating AI like some kind of oracle. This is
Dr. Deza: and it's not
Maximiliano: it's a tool.
Dr. Deza: It's a tool. Okay.
Maximiliano: Okay. So, uh reinforcement learning
Dr. Deza: is basically I touch a fire, I get burned, I get a reinforcement about touching fire is not very good for me.
Maximiliano: Yeah.
Dr. Deza: So, even if I'm absolute idiot, I touch a fire 55 times and then I finally learn I shouldn't touch a fire. Okay. Even if I'm like a turtle learner,
Maximiliano: I'm I'm going to
Dr. Deza: you're going to pick it up at some point.
Maximiliano: Yeah. And that's how uh animals and humans learn.
Dr. Deza: Yeah.
Maximiliano: We just get reinforcements.
Dr. Deza: Does AI do the same?
Maximiliano: Uh no, not at all. It's an absolute different mechanism. uh reinforcement learning is okay is is used mainly in robotics and is not uh as developed as um transformer technology which is the one behind um LMS
Dr. Deza: okay
Maximiliano: which is basically uh trying to predict the next word
Dr. Deza: oh there's no LLMs are literally uh the predictive on your phone on steroids. That's what it is. There's nothing else.
Maximiliano: But then why wouldn't people try to do this type of learning where you touch the fire?
Dr. Deza: Because it doesn't work as well.
Maximiliano: It's not for lack of trying.
Dr. Deza: Is it because of the amount of data required?
Maximiliano: Think of this. If you gave a GPT the same amount of data you hold in your mind like all the knowledge you have
Dr. Deza: it will be terrible you will be a
Maximiliano: because it's not intelligence the usefulness of all these LMS
Dr. Deza: is it because we have
Maximiliano: sto we store data maybe on DNA as well
Dr. Deza: no no it's because you're far more intelligent than any computer the thing is uh and the more intelligent you are
Maximiliano: the less d you need.
Dr. Deza: Yeah, that's that's a that's a well that's
Maximiliano: so if you are extremely intelligent, you don't probably uh you feel this is warm and then you need to touch the fire.
Dr. Deza: Yeah,
Maximiliano: you don't need I mean if you're dumb, you need to touch it 10 times until you realize that you shouldn't be touching it. that. And if you expand that and to logical conclusion, you go to um the more intelligent you are, the less data you need. And as humans, uh you show a baby an apple. And then maybe you show a baby another apple, a green apple and a red apple, and that's it. It's magic. After that, one month later, the baby or the baby or toddler going to see the apple. That says if you want to teach a computer with Napolis, you need to show it thousands.
Dr. Deza: Okay. Yeah.
Maximiliano: And thousands of images
Dr. Deza: of different types of apples until those tense. All of that is
Maximiliano: okay. Okay. So, it's not int let's say
Dr. Deza: it's not intelligent. It's the intelligence
Maximiliano: of one
Dr. Deza: IQ of an insect.
Maximiliano: Okay.
Dr. Deza: Okay. That's what it is.
Maximiliano: IQ of an insect. Okay. Yeah. Yeah.
Dr. Deza: Okay. What's the difference? It's extremely fast. Our neurons work uh at basically more or less hundreds to thousand times a second. Okay. That our
Maximiliano: Okay. Let's say a thousand times a second.
Dr. Deza: Let's say that. I'm not sure. Don't do the quote.
Maximiliano: Yeah. But but yeah,
Dr. Deza: but it's just for
Maximiliano: computers work at gigahertz, which is a million times more.
Dr. Deza: No. Uh uh billion times more. Okay. Okay. So,
Maximiliano: so it's a very fast
Dr. Deza: they do a billion iterations basically. So and uh is
Maximiliano: every time I think so
Dr. Deza: so a computer is able to do a gigahertz is uh million million
Maximiliano: okay
Dr. Deza: so it's a trillion times a second the amount of uh it's so fast
Maximiliano: that they cannot make them faster and they cannot make it faster because of the the speed of Right. Because so wait, but you see that computers are stuck at 4 GHz forever.
Dr. Deza: Yeah.
Maximiliano: There's no getting faster. Why don't they have computers of 10 GHz or 50 GHz?
Dr. Deza: Yeah, I've always asked myself that actually when
Maximiliano: because uh uh in one cycle of the clock
Dr. Deza: Yeah.
Maximiliano: is the time it would take light to go from here to here inside the chip. So if you make it faster you will start to have relativistic effects inside the computer. So information wouldn't be so it's a physical limit.
Dr. Deza: So basically computers are
Maximiliano: can can get faster uh this way
Dr. Deza: they they are at the physical limit.
Maximiliano: Yeah you can change the architecture you can make them more efficient. You can make a lot of things. But what now what we're doing
Dr. Deza: information
Maximiliano: right now what we're doing is we're just cramming more of them together.
Dr. Deza: Yeah.
Maximiliano: Putting more chips because we cannot make them faster.
Dr. Deza: Okay. Okay. Okay. Okay.
Maximiliano: They're so fast and at that speed and with the incredible consumption of of like a know of energy. Okay. You get something which is about the intelligence of a fly. mosquito.
Dr. Deza: Oh my god. Okay,
Maximiliano: so we are not there.
Dr. Deza: Oh,
Maximiliano: we are not there. We are absolutely not there. Okay, it's definitely not there. AGI is not happening.
Dr. Deza: How how long? Okay. Okay. So, can you put that in a sentence? Agi is not happening
Maximiliano: because the Let let me formulate this. Um right now with all the the GPUs and all the data centers and everything we have um and all the best LLMs we have the only reason all this models are so effective is because they are able to concentrate all the information of human race into one database which is what they are. So you can go and you can ask the CH GPT whatever and it's going to give you information very fast and very reliable. Not reliable fast. Okay. But they are not intelligent things. They're not thinking machines and they are going to give you information from what combinations of what they know. But they weren't able to generate new um uh anything new from what they unless it's just a combination of what they know.
Dr. Deza: Okay. So I'm I'm getting calmer and calmer as as as as this is getting farther and farther. So okay, let me let me kind of understand what what what you've just said. Okay. So first of all there's a big difference between AI and AGI let's call it
Maximiliano: AGI should be something working around reinforcement learning
Dr. Deza: okay
Maximiliano: which doesn't exist
Dr. Deza: okay so AGI doesn't exist the AI that we currently have
Maximiliano: it's basically a database
Dr. Deza: check is a is a database
Maximiliano: and they cannot learn as we learn
Dr. Deza: they cannot even learn interacting with you, they need to be specifically trained
Maximiliano: to Okay.
Dr. Deza: So, if you have a conversation with the system and you allow them to use your data to train them, they're going to see humans are going to see your data and they're going to decide to put that into the model later. It's not learning. It's not like you and me I'm learning from my interaction with you at the same time we're talking. This doesn't happen.
Maximiliano: They go learning mode.
Dr. Deza: Oh, okay. Okay, they have to like switch it
Maximiliano: basically
Dr. Deza: and then then they go non-learning mode.
Maximiliano: So it's okay they they're not complex at all then I mean they are complex in the quantitive of iterations.
Dr. Deza: The reason they are so effective is because they concentrate all the human all the information we have and they're not getting better because there's no more information. They have everything. They have all the books that were written all the blogs. So basically we've kept that.
Maximiliano: You're going to keep getting better at writing. You're going to be keep better at doing stuff maybe agentic stuff like I know moving things in your desktop. Maybe in the future you're going to be able to say to the you're going to have some kind of Alexa which actually works
Dr. Deza: instead of being useless. So it's going to be able to understand you and do things.
Maximiliano: Okay.
Dr. Deza: Maybe in the future you're going to be able to to do the things you want. like I know you but that's it.
Maximiliano: It's not going to get a lot better because all the information we gave it is all the information we have all the books all the music all the videos all the things they know everything.
Dr. Deza: So basically we've kept AI
Maximiliano: now it's going to be a very I mean and that's what people don't want to know. People don't want to know that because that will mean uh investors will say okay now why why am I investing billions into this if it's not going to get better I want it to get better very soon
Dr. Deza: it's not going to happen
Maximiliano: AI bubble is here then
Dr. Deza: it is it always was we all know it
Maximiliano: oh my god
Dr. Deza: I I think I finally get it in oh my god that's So terrible. I'm invested in AI myself. So, oh my god. AI has capped basically currently and we need to find another basic formula. You need to find another way hopefully some something related to kind of reinforcement learning something like that which is able to do things even
Maximiliano: animals can do like learning from their mistakes or being able to uh look for a reward and go away from from from problems. And what do you reckon? What do you reckon is like the major component separating it? Thing is what I thought was that because uh this was just a a guess. maybe is completely wrong, but yeah,
Dr. Deza: I'm thinking because AI doesn't have an actual body,
Maximiliano: it can gather well can gather some information that is relevant to situations.
Dr. Deza: Always sensors and things. It wouldn't be so hard to get that. So that's what a robot does.
Maximiliano: Yeah, exactly. So I thought that the problem with robotics was a major cap. Uh mainly because if an AI only has information,
Dr. Deza: it's like explaining a blind person what a what res is, you know, it's impossible
Maximiliano: more or less. So
Dr. Deza: So that's that was what I was thinking.
Maximiliano: I mean, for example, you you could let's do an experiment. Okay,
Dr. Deza: imagine you get uh a robot.
Maximiliano: You you get camera, a motor can go.
Dr. Deza: The thing is that's also information for that.
Maximiliano: Let me let me give you an example. You get you get a camera. Okay. You get a camera and then the robot can go in any direction you want. Okay. This is like a you can buy it for 100 100 quid.
Dr. Deza: Okay.
Maximiliano: Yeah. Yeah.
Dr. Deza: It's like a like a control remote control. Mhm.
Maximiliano: And then you not a robot but it's a remote control thing with a camera and
Dr. Deza: Yeah.
Maximiliano: and or a drone if you want. Uh then you connect that to chat GPT
Dr. Deza: the camera is connected to takes a picture
Maximiliano: and then sends that picture to chat GPT uh and says please tell me what's here.
Dr. Deza: Okay.
Maximiliano: Yeah. And then the very very easily the the picture made by the camera is going to go to some kind of LLM which is going to describe the what is scene. Okay.
Dr. Deza: Okay.
Maximiliano: And then you can feed that into GPT.
Dr. Deza: Yeah.
Maximiliano: And say okay I'm here. What should I do?
Dr. Deza: Yeah.
Maximiliano: Okay. And then CH GPT is going to tell you I don't know what you want to do.
Dr. Deza: Yeah.
Maximiliano: Because there's no motivation. Now imagine you add a motivation. Okay, I want
Dr. Deza: describe this.
Maximiliano: Describe this or I want to charge myself. I want to to I know do things.
Dr. Deza: Uh you could in principle use AI to see
Maximiliano: describe what you see. Yeah.
Dr. Deza: Feed that in language into the LM.
Maximiliano: Yeah.
Dr. Deza: Then take action.
Maximiliano: Okay.
Dr. Deza: Okay.
Maximiliano: You that could work.
Dr. Deza: It will be extremely slow. because it takes minutes to do other thing. But
Maximiliano: it it would work but you wouldn't be able to it would be basically remote control anyway because you need it would work for example for something like a Mars rover where you can give them overall things. I need you to go to this this address and then it will be able to take pictures and say oh I need to go around this whole
Dr. Deza: Yeah. And then and then
Maximiliano: but eventually
Dr. Deza: you would have to like basically give a set of instructions
Maximiliano: that that would basically be better than just remote controlling a thing when you okay
Dr. Deza: it would work hopefully there probably somebody doing that uh but it it wouldn't be a
Maximiliano: thing is I'm learning so much like my brain is over like
Dr. Deza: oh but it has no motivations
Maximiliano: okay
Dr. Deza: so without motivations it wouldn't know what to do.
Maximiliano: Okay.
Dr. Deza: So, because it
Maximiliano: Oh my god.
Dr. Deza: Okay. But you get a even a small animal, a rat,
Maximiliano: it's going to have more
Dr. Deza: it's going to be motivated in
Maximiliano: it's going to be hungry. It's going to be cold. It's going to be hot. It's going to be
Dr. Deza: But what if you would teach that? how just saying like I don't know when your battery is low go and make it your main motivation to go and charge yourself and then when you're not doing anything make the main motivation of doing something uh that would improve this other system and stuff like that.
Maximiliano: Try it.
Dr. Deza: Try it.
Maximiliano: I mean it would be like a set of incredible complex rules. Do yourself a favor, go to CHPT and try to make some kind of a role game like a Dungeons and Dragons kind of game with the LM,
Dr. Deza: okay?
Maximiliano: And say you are a soldier and you want to go there and try to do whatever the LM says. You're going to see it's not you're not getting anywhere. It's going to be so frustrating.
Dr. Deza: Okay. Okay.
Maximiliano: Okay. Uh he's laughing at me, man. Laughing at me. Hey, you mother.
Dr. Deza: Let's um let's finish with um something probably everybody wants to know. It's about okay. LMS are not getting human intelligence. We don't we're not going to have terminators.
Maximiliano: The thing is I didn't know that. Okay. But but yeah. Okay. So
Dr. Deza: basically what you're saying is we're not getting any any
Maximiliano: terminators right now.
Dr. Deza: Uh but
Maximiliano: but
Dr. Deza: right now or in the future as well? Uh I mean you never know what's going to happen in the future.
Maximiliano: Uh we know intelligence is possible because we exist.
Dr. Deza: Yeah.
Maximiliano: So we
Dr. Deza: it's just a question of
Maximiliano: it's a question of time and and learning how to do it. Okay.
Dr. Deza: Yeah.
Maximiliano: But so it's not I'm not saying AGI is impossible. It is possible. We exist.
Dr. Deza: Yeah.
Maximiliano: Uh but I'm not seeing it happening anytime soon.
Dr. Deza: Okay.
Maximiliano: Okay.
Dr. Deza: Perfect. So
Maximiliano: anytime soon as here to 10 years
Dr. Deza: yeah I don't know 20
Maximiliano: I don't know because there are mechanism where uh computers could start um designing themselves.
Dr. Deza: Yeah the thing is I feel it's like a flip is it won't be like a
Maximiliano: it's going to be probably a process.
Dr. Deza: You reckon it's a process?
Maximiliano: I feel it would be like a switch and then it would just take time to Yeah. Well, it's the same base.
Dr. Deza: Uh And it's going to be the same process which is happening now. So if you have 100 years ago
Maximiliano: increasingly more intelligent than perfect. So you will see it coming basically.
Dr. Deza: Of course you probably wouldn't be able to do anything about it. But the same with what we're talking about information. I saw it coming back then but uh when smartphones came I was like oh now I can see I can Google everything from my phone the streets. It's so fantastic. Oh my god, this financial crisis is going to be so bad.
Maximiliano: Well, just don't go there. Go there.
Dr. Deza: The thing is, um,
Maximiliano: don't go there.
Dr. Deza: I'm just thinking about the
Maximiliano: just sell sell everything.
Dr. Deza: May I'm I'm literally thinking about it. So, it's
Maximiliano: the thing is, um, the major companies in the world currently, top three, let's say. Well, it's it's more than the top three. The thing is within the top 10, eight of the companies are AI or related to all of them are going to go down like a mother and those are like
Dr. Deza: not really not really. It's not that they are not the thing is AI is useful.
Maximiliano: Yeah, that's the thing. But they're not as useful as they
Dr. Deza: they their projections are o I mean they're overstated
Maximiliano: but they are but the the product works.
Dr. Deza: Yeah.
Maximiliano: And the product it is literally making people more efficient. So you can get any programmer and you are able to
Dr. Deza: do seven times
Maximiliano: do seven times more code than you used to do. Okay. Of course, that's what I remember hearing in in my economics class.
Dr. Deza: Your generation, this was my teacher. He's 60 50 something.
Maximiliano: He was about he was about to retire in 10 years or something like that.
Dr. Deza: So, he told us um your generation is going to be um statistically we're expecting your generation to be seven times more productive. seven times as seven times more food, seven times more information, seven times more.
Maximiliano: But but look, in the 60s and I wasn't alive in the 60s,
Dr. Deza: uh if you were you've had a company,
Maximiliano: you had to go and everything was paper.
Dr. Deza: Okay.
Maximiliano: So, you had to clients.
Dr. Deza: Yeah. So you had to go and find the the folder, the paper folder where all the client things were. You went to the first floor, find the clients. Then you go to the second floor, you got the the products, the inventory. You go on to the third floor and there were people basically paid people who you who went to all the floors, got all the the folders and gave it to you.
Dr. Deza: Yeah. And then you made the sale and then you wrote something, you put it back into your folders and then this guy went back and put everything back.
Maximiliano: I I was I was one of them for a legal company and it Yes.
Dr. Deza: So that was was no computers. Then the main frames came. So the the the first computers were and every single system had it own computer. So you couldn't just integrate everything. You need still needed to to uh you made a sale but say your sale was in sales it wasn't in accounting.
Maximiliano: Yeah.
Dr. Deza: At night some guy just related.
Maximiliano: Yeah. Related everything. Yeah. Yeah. Yeah.
Dr. Deza: So everything was
Maximiliano: which was
Dr. Deza: one day after
Maximiliano: and that was like the main problem for accounting
Dr. Deza: that went all the way until the 80s.
Maximiliano: Okay. And in the 80s they finally got this big data set data basis and the SQL and all that and then they started to
Dr. Deza: implement
Maximiliano: implement things more or less like we have and then the web came.
Dr. Deza: Yeah.
Maximiliano: And everything went online and now we have big data.
Dr. Deza: So and now we can process millions of times more. Not only not only your company but all the companies from all around the world
Maximiliano: and now for for doing something now it takes one person uh only 60 years ago in the 60s will have taken 100 people.
Dr. Deza: Okay.
Maximiliano: Okay. And that's going to continue.
Dr. Deza: That's that's
Maximiliano: it's going to continue. One person in 20 years is going to be able to do more work. Okay. And then that then comes the question okay if more people is people are going to be more productive what's going to happen with the other people and that's uh what I wanted to talk that
Dr. Deza: where are you going
Maximiliano: exactly so
Dr. Deza: where the where the are you going so there are two ways one way is productivity overall is going to go up
Maximiliano: so people are still going to have jobs only everybody's going to be insanely productive for our standards
Dr. Deza: which is what what's what's going been going on on the last 100 years
Maximiliano: and that's basically uh the work hour week let's say
Dr. Deza: value cutting let's call it value so value is like a production of objects services or whatever
Maximiliano: yeah but then
Dr. Deza: value but then uh There's a devaluation of the product and the service.
Maximiliano: Yeah.
Dr. Deza: It's cheaper.
Maximiliano: Yeah.
Dr. Deza: Because
Maximiliano: it's cheaper and cheaper and cheaper.
Dr. Deza: So, for example, in the middle ages, you want the a shirt
Maximiliano: and it costed
Dr. Deza: and some guy came
Maximiliano: and some guy came and had to um to make it for you.
Dr. Deza: Yeah.
Maximiliano: And they may had to thread all the
Dr. Deza: Yeah. everything
Maximiliano: and get the sheep and everything from scratch
Dr. Deza: and
Maximiliano: and that would cost
Dr. Deza: it was costly.
Maximiliano: That was a week's uh I remember this it was a week's salary was uh one t-shirt.
Dr. Deza: Exactly.
Maximiliano: I remember exact the exact amount I think I that was one of my questions to one of my economics teachers
Dr. Deza: after the industrial revolution
Maximiliano: you had uh productivity was over the roof compared to the middle ages.
Dr. Deza: Yeah. And therefore that's when we got
Maximiliano: exactly and now we have so much productivity that we can do something called fast fashion. We can really
Dr. Deza: just change every what was 40 how many seasons were
Maximiliano: I think it was 40 something seasons per year
Dr. Deza: that's that's fast fashion
Maximiliano: which is ridiculously fast and uh but we can handle it we can manage it production side okay not environmentally side but
Dr. Deza: yeah but but but the
Maximiliano: production side we we can we can handle that because everything's everybody's so productive everything's so cheap
Dr. Deza: that you can't do that
Maximiliano: u like a t-shirt now is worth productive
Dr. Deza: but I feel that's why and this is what I like about the the market the market has like a very weird curvature where like at the beginning it's extremely expensive
Maximiliano: then cost goes down
Dr. Deza: but then uh it goes it goes more and more expensive again
Maximiliano: so what happens for example with water because water has been in the market for a long time.
Dr. Deza: Uh at the beginning, so let's say African
Maximiliano: Yeah.
Dr. Deza: was expensive.
Maximiliano: It's is uh yeah, extremely expensive. Then it goes extremely cheap uh day-to-day water that we have now, but then people want quality water
Dr. Deza: and then there's a market for that
Maximiliano: and then it goes up again. So that's that's how it works. Uh so our economy now is basically services
Dr. Deza: and services are going to going to go up and people are going to require more services.
Maximiliano: Yeah.
Dr. Deza: Which are going to be uh because we have all of our material needs more or less or at least developed work
Maximiliano: um fulfilled.
Dr. Deza: Yeah.
Maximiliano: So we have food, we have shelter, we have
Dr. Deza: Yeah. The basket of goods in England I think uh also uh is adjusted to also have a vacation which is incredible. So now if if we talk about our needs,
Maximiliano: yeah,
Dr. Deza: if you think about even 10 years ago,
Maximiliano: the amount of services you have now, you can have a LM which is able to give you advice and things can give you able and those things costed a lot of money only 10 years ago.
Dr. Deza: Yeah. or psychologist psychiatrist
Maximiliano: psychologist or people who so that and that's going to to be worth less and less less
Dr. Deza: skin skin treatments I' I've seen that there's a there's an app now that
Maximiliano: smooths or you can use a yeah like there's a Instagram thing you can basically tweak your face and make it better and
Dr. Deza: so all the things are going to keep coming
Maximiliano: productivity is going to go probably over the roof.
Dr. Deza: But anyway, some jobs are going to stop being profitable.
Maximiliano: Yeah.
Dr. Deza: For example, even five years ago, uh being a a writer, uh I mean not an artistic writer, more like a being a copywriter or working for for for a company just writing um publicities and messages.
Maximiliano: There was that was very profitable.
Dr. Deza: Yeah.
Maximiliano: Now, only a couple years ago, I remember. Yeah. Yeah, I literally remember watching videos on YouTube saying, "Oh, you should learn copyrightiting and so on."
Dr. Deza: And now it's it's it's not unprofitable. It's gone.
Maximiliano: It's just unuseful to
Dr. Deza: It's gone.
Maximiliano: Yeah.
Dr. Deza: Completely.
Maximiliano: As a researcher, I need to write papers. I write
Dr. Deza: and I was very proud of my skill. I was able to write English, good English, like very formal scientific writing. That was also now done
Maximiliano: and now it's useless. I still write my papers because if I know I'm oldfashioned, but then
Dr. Deza: some of my collaborators are obviously going to use some kind of AI to check my writing.
Maximiliano: I wouldn't if I would be a professor. I wouldn't because my time is worth more than writing paper. Sometimes it takes more time to to actually check what the thing wrote
Dr. Deza: than to
Maximiliano: but I I use it but I don't use it like a write my paper.
Dr. Deza: Yeah.
Maximiliano: It's like a
Dr. Deza: is a tool.
Maximiliano: Yeah. So and and the programmer same thing before even five years ago even when I was doing my PhD I thought about becoming a programmer because salaries were much better than than yet but then now a programmer is either needs to program 10 times more than it used to do for the same job or they some of some of them are just out of the job
Dr. Deza: and it's happening it's happening so that's one of the things how you keep a job and how you keep true to your own calling
Maximiliano: on the time of AI what do you do and the only and I'm
Dr. Deza: what do you reckon is the answer because I know it's a difficult question
Maximiliano: it is it's a very difficult question and I had a lot of conversations with students and of mine who come and say look I'm studying computer science because I'm I was a programmer And now I want to do something else because I see my my job is squeezing me or I'm getting uh I don't know if I'm going to have a job. I'm not I'm paying lots of money for this education, but I know if I graduate my job is going to exist.
Dr. Deza: And and it's true many of the data science things I'm doing can be automated. Okay, it's true. So the thing is this is happening. It's been happening for a lot of time and it's going to keep happening and you have no control over that. I didn't have control about um
Maximiliano: internet, smartphones and anything else was going to happen. I I didn't have control. So but how did I adapted and how people are going to adapt in the future is by uh realizing that uh there are things going down like it's a devaluation of knowledge information and not knowledge but more information of information devolution of other things and but in the same time there is
Dr. Deza: people craving for other things and those things are for example uh connection uh human connection usually and trying to understand uh trying to many jobs that are really um are going to people need to understand uh things from noise. As we said before presenting yourself, talking yourself with people, people are going to follow some kind of leader or some kind of expert and I'm going to trust them a human more than they trust a computer. So if I go and I want to use uh somebody's gonna oh you you're good and tell me which is the tool I should use or what should I do and and they're going to reach people they know they're going to reach other other people.
Maximiliano: Yeah.
Dr. Deza: So because and that's going to be very valuable. It wasn't valuable like only 30 years ago people but it's going to become valuable now. At the same time things which were expensive before are are becoming cheap. So is simply changing.
Maximiliano: So everything related towards information uh will go basically down or everything
Dr. Deza: be worthless.
Maximiliano: Worth okay. So you reckon things like um okay I'm going to put like big big names here doctors.
Dr. Deza: No not only that not only that even if you are uh any any even a technical job
Maximiliano: your job is going to be less about doing the actual typing.
Dr. Deza: It's going to be more about human things. It's going to be more about talking to people and making your clients understand what their problems are.
Maximiliano: That is actually incredibly interesting because we have seen a slowly decline on how people relate.
Dr. Deza: Uh basically people are having less friends, younger people have
Maximiliano: but that's why that's why it's becoming more and more valuable. is why people are craving it.
Dr. Deza: Oh my god, there's going to be a huge Yeah, there's going to be a huge huge change.
Maximiliano: But people who are able now, people who are studying now, people who are becoming like people in the 20s are studying something need to understand that their job is going to be less technical and because that's going to be very
Dr. Deza: it's going to be more about experiences. It's going to be more about understanding people, more empathy, more that even if you are an introvert person doesn't matter. Introverts
Maximiliano: have to learn.
Dr. Deza: Introverts are extremely good at understanding people.
Maximiliano: Yeah.
Dr. Deza: Once they get to know them,
Maximiliano: I feel introverts have this special thing where they fully understand somebody
Dr. Deza: incredibly fast and that's why they're they become Yeah. they they become self-absorbed I would say because of the amount of information coming in
Maximiliano: but now
Dr. Deza: incredibly sensitive
Maximiliano: but now many but that that's good and many people are going to um get this type of jobs
Dr. Deza: who weren't valuable at all only 30 years ago like you could go to your neighbor and and have a lunchon with them and and then but that wasn't even thought as something valuable but Now it's going to become that.
Maximiliano: Oh my god. I think I've I I think I've starting to see the glimpse. I didn't I Oh my god. This this I think interview has been the best I've I've done until now because of how much I've learned. My major problem was uh before uh was uh I did not like I accepted okay Chv is coming in, AI is coming in. I did not know the implications of that. I didn't know what chat GBT was and how it was actually going where was it going to cap because I knew it was going to cap as any technology would is going to cap somewhere. Now I know exactly where it's going to cut. I mean
Dr. Deza: I'm not Oracle
Maximiliano: what the future
Dr. Deza: they may be able to crack some some things.
Maximiliano: I know. Yeah, I know. But you kind I feel I kind of understand now where where we're going at and oh my god it's been so so helpful.
Dr. Deza: Oh my pleasure
Maximiliano: Ignasio like I oh my god I would recommend anybody to watch this podcast.
Dr. Deza: They already watch
Maximiliano: till the end.
Dr. Deza: No no no this is for the clip before.
Maximiliano: So Ignasio thank you very much.
Dr. Deza: Thank you for inviting us. Uh, oh my god, I can't really uh thank you enough for how helpful it was.
Maximiliano: Pleasure.
Dr. Deza: Um, guys, uh, thank you very much again, uh, for watching another episode, uh, of, uh, Beyond the Horizon. Uh, and, uh, just as always, uh, please follow, uh, please follow, uh, like and subscribe if you can.
Maximiliano: Uh, and we are currently on YouTube, uh, Instagram, Soundcloud. Anyways, thank you very much and see you, uh, next Tuesday.
Comments