Welcome to "AigoraCast", conversations with industry experts on how new technologies are transforming sensory and consumer science!
AigoraCast is available on Apple Podcasts, Spotify, Podcast Republic, Pandora, and Amazon Music. Remember to subscribe, and please leave a positive review if you like what you hear!
Dr. Andreea Botezatu earned her bachelor's and master's degrees in Oenology and Agricultural Management in Romania before working as a commercial winemaker in Europe and Canada. She later obtained her PhD in Oenology from Brock University in Canada, where she also completed a postdoctoral position.
Currently an Associate Professor and Extension Enology Specialist at Texas A&M University, Dr. Botezatu's research focuses on wine aroma, fault remediation, and overall quality, utilizing both chemical and sensory analysis. She also investigates Texas wine consumer attitudes and is exploring the application of emerging technologies like blockchain and LLMs in the wine and sensory fields.
Visit the Texas Viticulture & Enology Facebook Page
Transcript (Semi-automated, forgive typos!)
John: Andreea, welcome to the show.
Andreea: Thank you, John, for the introduction and thanks for having me.
John: Yeah, pleasure to have you on. We met on LinkedIn, and actually it's funny because, you know, I think LinkedIn has its pros and cons, but you definitely meet interesting people on there.
I would say LinkedIn has been good to me for meeting interesting people. And I posted about LLMs and sensory and I had a little sheet and you reached out, you know, because you wanted to kind of get a copy of the guide. So maybe you could, we can start by talking about your background and then how it is that that led you to become interested in large language models
Andreea: Sure, my background. Where do I start? I was born and raised in Romania, educated, did my bachelor's there, my master's, as you mentioned. I grew up surrounded by vines and grapes. We have a long tradition going back thousands of years of winemaking. So I kind of grew up with that. I studied horticulture and then, while at university, I decided to specialize in enology, which is the science of wine and winemaking. Deals a lot with wine chemistry, fermentation, microbiology, wine technology processing, and of course wine aroma, wine quality, all all that is related to that. I specialized in analogy at the time when I did that, it seemed to me that to be a very male dominated field, so I liked a bit of a challenge from that perspective. Now it has changed quite dramatically. A lot of women winemakers and wine scientists. Yeah, and, that's how I started. I worked as a winemaker both in Romania and in Canada, really enjoyed that, and then I shifted towards research and academia with my PhD at Brock and postdoc there.
John: OK. And so you can, I mean, you, I mean you kind of classic intersection here of, you know, sensory science and kind of the industry of winemaking, how is it then that you got interested in large language models, what are the problems you're working on where they started to come up and they started to become interesting to you.
Andreea: I started to get interested when I first learned about ChatGPT. It wasn't necessarily related to work. I was just fascinated by it, to be very honest, I was curious about it, what it can do. At the time, it was a few years ago, 3 or 4 years. I can't remember the exact timeline, but anyways, at the very beginning. I remember being very excited about it in relation to research and kind of finding out data and information and synthesizing them. So I asked to do some of that and I remember being very impressed with the result and also at the same time I was like, this thing is making things up, that this isn't real, start to check the references that it was giving me. Trying to find the papers that it was referencing and I I found out that it was totally making things up. So that was the beginning of my, at the time love-hate relationship with large language models and things of course evolved and I remained curious about its capabilities and I found, I found it I still find it very useful for various, things that, it can help me with, but as, as my research progressed and as I, you know, talk to more people about Blockchain technology, I think it maybe it connected through blockchain technology as well, blockchain technology, how it applies in the wine world, authentication in the wine world, and, and all of that. It kind of all came together in a sense. I am just a very curious person. I like. Stay on top with new developments, new technologies, stay on top of, so it's just, just something that piques my curiosity really, how can we use all these new technologies in research, in science, and since my field is wine and sensory science, how can I use them in my field?
John: Right. Yeah, I still remember that day. I think it was November 2022. It was like, I don't remember the exact day it was, I remember the day very clearly where I woke up and I got on, I think it might have still been Twitter at the time, maybe it was X, and everyone was talking about this new tech, this new thing, chatGPT and of course I downloaded it immediately. The first thing I made was some recipes. That was the first thing I did to test it, you know, and they were OK, but I could tell like this is something, you know, I mean, I knew about language models.I knew they were kind of coming, but I thought, all right, this we finally have done it like I like as a human race has crossed some threshold today. That's what I remember feeling and. Yeah, it's funny because some people don't, didn't get it at all. They still have an opinion of, of large language models based on the performance of the model on that day. And I think every single day it's worth remembering, literally every day is the worst that AI will ever be. Tomorrow it'll be better, you know, Google just pushed an update to 2.5 Pro, and now it's even better. It's always gonna be better tomorrow and I think people don't really understand that we've gone, like, we're going through a revolution, it's happening, and blockchain is part of that and language models.So maybe we can get to blockchain in a bit. Let's talk a little bit more about language models. What are some of the things that you're seeing in terms of you, like just in your day to day work? That's a really interesting question, how are you using large language models and then, you know, the applications that you're seeing to the wine industry.
Andreea: Well, I use it a lot for that very first purpose that I mentioned, kind of looking through literature, seeing what's new, synthesizing information, getting to the, you know, core of it, it's the large language models have evolved, significantly since then and they're much more accurate now. They're much more detailed. They have this reasoning function that I find very helpful to see how it's doing, how it's processing the information, what it's looking for, where it's looking for the information, so I feel that I can trust the output a lot more. I can go back and check the sources and the origins, the articles, the links, so I feel a lot more in control, that way. So I use it for scientific information for, getting, short, you know, focused information from various sources, for ideas, ideas for educational events, for, fact sheets for whatever it is that I need to create for my field for videos, educational videos, contents, things like that. It helps with preparing presentations as well. I, as an educator, I have to do a lot of presentations, so it helps organize the data or structure the data in a way that's logical and makes sense.I really like its ability to use different tones, for various types of publication because you can have very technical, things, you know, articles, and you can have very articles written in layman terms and funny, articles and all kinds of approaches to how you deliver the information and it it's very helpful in that sense. So I use it for that as well. Yeah, I, I use it quite a bit actually, lately, and I could be wrong, I am not a specialist in any of this, but I feel like large language models are more than just that language models. I feel that that's how they started with words and analyzing words and how they come together and creating pieces of text, but now they're much more than that. You can run analysis, you can write code, you can do a lot more than just text. I think that's very useful statistical analysis, for example, it can suggest various types of statistical analysis for various types of data. What's the best type of stats that I can run to analyze the data, how, how I can get the most out of it, and I haven't done a lot with that yet, but I'm starting on that path. I'm curious to how well it works, in that direction and maybe you know more about that than I do, and maybe you can. Chime in here a little bit about that.
John: Oh yeah, I have a lot to say on this topic, but you're the guest, so..
Andreea: I would love to learn more seriously, that's why I got in in touch with you to begin with because I'm very curious and like I said, I want to learn more. I want to know what it can do and how we can help it to do that. Can we as sensory specialists, can I, as a wine scientist contribute in any way to the knowledge of these models and make them better?
John: Well, first off, to answer your kind of first question about coding and analysis, I think that a lot of people don't appreciate the value of the language models as a tutor. And I think that a good habit when working with whenever you have something to do, is go to, go to a very general. Grok is a good model actually for this, you know, the X model, the XAI model, because it's, it's connected to the X database, which is pretty much cutting edge stuff. What people are doing. You know, it's on X, like 2 days ago, 3 days ago, whatever the latest thing is, right. A good habit is to go to Grok and say, how, you know, here's what I'm good at, how can I do this task, you know, and I do think that everybody should try to get comfortable with the basics of kind of lightweight programming that, Google has a very good tool called Google Collab where you can write Python scripts. And you know, I've, I've done a lot of work in R in my life. You probably, I would guess, I'm not sure if you're using R or SAS or what you're using for your analysis, but, you know, I used to be on the statistics side and so I did a lot of statistical programming, but I've moved more towards Python because that's really where the coding developments are happening. And Python is not as scary as it sounds, you know, that like Google Colab notebook is very easy to work with. You can have, if you've got multiple monitors, you have collab open on one screen, you have Grok on another. And whatever questions you have, just say right, Brock, I'm new to Google Collab. I've heard I should get started in Python. Can you teach me how to use this? And they'll say, Sure, happy to help, and it'll help you as much as you want and open this and say, I want to do a binomial analysis. How do I do it? And it'll give you some code and you can put it in and you can run it.
And very quickly, in fact, Vanessa, one of our consultants who is a co-author of a book on R for data science, is wanting to get, you know, she wants to get up to speed on Python. So I gave her a little assignment, make a dashboard to do binomial analysis, and she did it as a kind of weekend assignment. She made a whole dashboard, and she was amazed. She couldn't believe that like.
She went from Grok, and I think that people need to understand that things are a lot more approachable than they used to be.
Andreea: I agree with you. I think a lot of people are intimidated by it and they're like, oh, I don't know what it is. I don't know how to use it. I don't know what to make of it.
And it's like try, just open it and ask it a question and see where it takes you, ask for guidance, for instructions, take it step by step.
It's very User friendly and yeah very intuitive in a way I it's not it's not as intimidating as a lot of people think it is.
John: Yeah, 100% it's very democratic.
It is very democratic actually, yes, I agree with that entirely.
so getting back to kind of the, the wine industry, I know that's like can you talk about some of the problems that you see like you, you work with the wineries there in Texas and you're You know, kind of active with them.
What are some of the problems that they have that you are starting to think that maybe language models might be able to help them?
Andreea: I don't think there are problems, maybe opportunities, I will call them, a lot of the wineries here, Sell most of their wines through wine clubs, for example, they don't export much, they don't sell out of state much.Their customers seem to be located within the state of Texas and a lot of them are wine club members, which means that the wineries will ship, their wines to their members, you know, to the participants, or they will come visit the wineries every once in a while for events.So maybe large language models could be utilized to target the wine club's member wine club members or attract more wine club members, maybe they can be utilized for wine descriptions or feedback from people and the data can be centralized and analyzed how, how customers respond to the wines and what their descriptions of the wines may be, what do they pair them with, what do they use them for, what kind of events, when do they want the most things like that, some feedback from the customers, communicating with customers in a language that is easy to understand by the customer or supplying them with what they want, the type of information they want using the words that they understand. Because word use is very important in the wine world, I find, starting from the names of the varieties that we're selling or we're trying to sell, and sometimes those names are complete unknowns to the customers, especially here in Texas.We grow varieties that are not very well known and a lot of people are intimidated by those, so come up maybe with creative wine names for all of red wine that would attract the customers, what the customers respond to. So consumer understanding consumer wants, needs, trends, things like that, getting feedback from consumers and what faces the consumer from the winery.I think that's an opportunity for large language models. A label creation, front label, back label, again, descriptions, pairings, suggestions, information about the winery, things like that. I think that those could be useful as well. Creating apps for the winery, I mean, wineries could create apps, right, to work with, to guide them to various wine choices or food and wine pairing. Hey, I'm thinking of making risotto tonight. Do you have a wine that would go well with it and what are its characters? I don't know. I'm just, you know, shooting ideas
John: It's quite interesting.There's a couple of points here that like follow up you know, you mentioned tone earlier. You know, that you like that within a large language model, you can write something in the right time. The tone tone is related to what you're talking about, you know, consumer messaging, where you're talking about putting, you've got the winemaker's way of talking about things and now you need to effectively translate it into the consumer language. This is there's kind of this continuum where you got tone and then you've got, you know, the language choices, then you have actual translation and the large language models are good at all of that, that if you want to personalize the messaging, I mean, it could just be generic translating it for, you know, Texas consumers, but eventually, you'll know a lot about these, these people. If they're in the club, they could fill out a short questionnaire and they could get communications that are personalized to them. I, I, and that brings the second point, which I think there's a lot of concern about jobs going away because of AI, but the fact is, I think for every problem, for every like one problem that AI solves, there's two new problems, right, where now you have all these things you can do where you can personalize communications, you can make apps to recommend recipes to go with wines or even the other way around, someone says, hey, I'm having this, which is the right, I mean wine pairing. Here are my ingredients. Can you put together a recipe to go with. But suddenly the, the, the space of problems we can work on has expanded and people need to be adventurous. They need to think, you know, is it possible to make a little chatbot? Like it might be a lot easier than you realize to make a chatbot.
Andreea: I think it is. I mean, it's just playing with chat GPT a few hours ago and it created some code for me to create a chatbot now I don't know what to do with that code. That's a different conversation, but it doesn't seem very complicated to create something like that and to implement it. I think another use just just came to me right now, so, could be adjusting language for different markets, especially for exporting in different countries. different people in different parts of the world use different words for the same characteristics, so, adapting your label to the market where you're exporting so that the consumers understand what you're trying to communicate that that very easily can be done by a large language models or comparing side by side wins to say a temperia from Spain with one from Texas, with one from France and see how they stack up side by side in terms of the scriptures. But from my perspective, so moving on from wine, I don't know if you want to move on from wine yet or not. So from my perspective as a researcher who looks a lot at the chemistry of wines and particularly aroma chemistry, I work with gas chromatography mass spectrometry quite a bit, and that's a technique that allows me to one, identify and to quantify aroma compounds in wines. So what compounds do I have in my wines and how much of them do I have in my wines and, and it's very, this technique is very good at doing that. But then usually in the wine world we try to correlate the data from the chemical data, yeah, all those compounds that the GCMS has identified and quantified with the sensory profile of the wine. And sometimes that's fairly straightforward and sometimes that's not, and there's this whole area where it isn't straightforward yet. Because we have compounds in wine that are below the sensory detection threshold. So a sensory detection threshold for humans is a concentration of a volatile compound where we are able to detect that compound in whatever matrix we have. So, in wine, we need 5 mg of such and such compounds for us to be able to detect it. Yeah, that's an example of a sensory threshold. Well, there are a lot of aroma compounds in wines that are there below the sensory detection threshold, and we know that they interact with each other somehow they influence perception, but we don't know how. It's very hard. It's very hard to figure out the sensory interactions at sub detection threshold levels because there are so many aroma compounds. So maybe we can use these large language models to come up with hypotheses. Maybe if we feed the data, the GCMS data and we feed it as a sensory profile, maybe can tell us, hey, I see a pattern here and if we have this compound at such and such concentration, this compound below this compound above it they seem to interact and create this kind of sensory profile for me that would be super exciting that that would be something from a research perspective that would be very, very interesting to do. Because it is so hard to do without this kind of help. There are so many potential variations and combinations that for humans it's hard to, it's hard to do.
John: Yeah, and that is really exciting. I mean, we've had some success, you know, at Aigora with training neural networks for those types of predictions. I mean, it's a challenging problem for sure, and you do need quite a bit of data. Now, of course, you know, the kind of state of the art models have seen so much data that they have a lot of implicit patterns that are loaded in them. And so it's possible that's a good jumping off point for fine-tuning a model to go this final step and try to, I mean, the thing is, when you take your GCMS data, do you take the images of the gas chromatography, or do you have a readout?
Andreea: We have a readout like we have a report that tells us it's a list of compounds based on retention times. So from the beginning when we start monitoring the retention types a minute. To or whatever to the final minute of our program because GCMSs are programmable, some methods are longer than others. Anyway, based on retention time, all the compounds that it identified are listed with their abundances and if you have a calibration curve with their concentrations listed as well. So based on that you can put that in an Excel or whatever table you're using and feed that data into a model.
John: Because you know, some of the advances in computer vision have, I mean, for example, most of the state of the art models are actually multimodal models and you know, like 2.5 Pro from Gemini or 4o of from ChatGPT. These, these are actually 03 is an excellent vision model. I mean, I think there's just so many problems to work on right now. I can't imagine a better time to be a researcher or to be in business or be at the intersection of the two. Yeah, to me, the idea that we're going to run out of things to do is almost silly. Like there's so much to do. There's more to do than there's ever been.
Andreea: There is. I feel a little bit overwhelmed because I go to it, you know, questions, give me ideas for and then it gives me 20 ideas and then do you want me to write you a script for this or that or do you want to step by step and then I say yes and I look at it and I get very excited, but now I have to make 20 videos, 3 fact sheets, so I don't know how many presentations because it suggested that I do so and 5 different research projects and I go, oh my God, where do I start? I feel a little bit overwhelmed, honestly, and then I have to learn about these other platforms that create the videos or the images or build a chemical formula. It's a cascade of ideas and the projects that come, yeah.
John: We say that a lot at Aigora that like now you really can do anything now but you can't do everything so you still have to make choices. It's very interesting. I have to say no all the time to different ideas, you know, and they're all good ideas. That's the thing that's tricky is you've got tons of good ideas you can just do them, but you have to pick the right ones to work on.
Andreea: Absolutely, yes, I have a library of ideas now and I have to prioritize them because like you said, can do them all.
John: Yeah, and actually I think that's the role of humans in this new era that we're going into that the machines, of course, they're basically tireless and they are very capable, but they have no sense of direction. that they don't have I mean we have we're evolved organisms we have internal drive, desires, intuition, we have all sorts of things going for us that machines, and machines are engineered and they just aren't going to ever have that same inner drive that, you know, we as humans will have.Machine can optimize to a metric, but it doesn't know which metric, even, If it wants to know which metric, it's gonna want to know what's the goal, and then it'll figure out which metric, but only a human can give it a goal. Yeah, and so I feel like we're at this kind of scissors moment where people have to decide they have to go all in with whatever they're doing. They have to say, OK, this is what I want to do. And now there's everything, all the doors are open and people are gonna have to say, what is the thing that I'm called to do and try to find it and and follow it rather than like the days of showing up at a job and being told what to do. And taking your money and going home are pretty much, I think, coming to an end, that you're going to have to think about what it is that you really want to do that provides value.
Andreea: Yes, but you have to be comfortable with that. You have to be open to learning new things. You have to be a little bit more flexible in your thinking. So I think there's a adaptation curve here.
John: yes, I agree,
Andreea: for, people in who, who can be involved in this or who could benefit from this and, to, to accept or to try to understand and try to see the benefits of it and be open minded about it.
John: I totally agree with that
Andreea: Inertia happening, which maybe it's a good thing. Maybe we need the balance, maybe we need those who really go all in and use them to your full capacity and those who stand back a little bit and, keep going in the traditional, path of things and maybe there's, there's value to this, balance right now, but I think overall, as society, as, as things evolve, not but things evolve, we will have to be a little bit more flexible and open minded and see where it takes.
John: I agree, and I do think it's up to the people who can see where things are going to try to educate everybody else and to reassure them, hey, you can write Python code yourself, you know, like I, the, the kid who mows my lawn is doing some work for me here, as an assistant, you know, helping with papers and stuff, and he, today I was doing some Python code. I said, do you want to write Python code? And he said, I can't do Python code. I said, yes, you can. Let me show you. Like it's possible. And of course you need you need oversight. I wouldn't take the code he writes and put it into. You know, anything important, but like it's good for him to understand that these things are not like you get Grok open, you everybody has a tutor, and the tutor can talk to you in the language that you're that you're comfortable in that, you know, connect with the machines is you can make yourself more capable almost immediately by by connecting
Andreea: yeah, I agree. I also see a little bit of suspicion. people are suspicious of, these models and I think, why are you using them? Are you cheating? Is this cheating? Is this cutting corners or I, I, I, you know, I don't know how to feel about that. I don't know.
John: They have a point. I mean, look, like my son's school, they're having this discussion right now about whether or not the children should be allowed to use AI, right? This is an active discussion in education and it doesn't do the children any good if they're allowed to use AI and what they tell the AI to do is write a paper on the French Revolution and it writes a paper and they turn it in. That of course is not doing anyone any good and what does them a lot of good is if they work with the AI and they write some historical fiction and they work together, it's very interactive and they're learning about things and then they present it. If it supplements their learning, then it's great. But if it replaces their learning, then it's you know, it's really pointless. And what's the point of having a machine write a paper and then another machine grades it and then everybody goes out. So I did, yeah, I agree that we have to be careful.
Andreea: Yeah, and I think education in that sense and information about how it works and how it can be used, how it should be used properly, what are the opportunities and what are the pitfalls, possible pitfalls would help, a little bit of transparency or, I'm talking about the people who are reluctant to use it or who are suspicious, and I think, actually everybody would benefit from education, on that. It's coming. It's gonna be here either way. We need to make the best of it and understand how we can use it to our advantage, not disadvantage.
John: That's right. Yeah, sometimes on LinkedIn, I say it's inevitable and people say that that sounds threatening. Well, I'm not trying to threaten somebody. I'm not trying to threaten people. I'm just trying to tell them it's inevitable. It is for sure happening and we have to, and it's a good thing if done properly, but yeah. Yeah, no, that's right. OK, I was hoping we could talk, talk about blockchain, but we're pretty much at the end. Blockchain is a whole other area.
Andreea: That's a whole other area, yeah, but it's so cool in the wine world. I'm very excited about. It's cool everywhere, especially with the AI revolution happening with all the can be created, and we cannot tell if they're authentic or not, and, using blockchain for that, I think it's, it's a must in my opinion but, transfer that to the wine world, I think it has a lot of applicability in the wine world.
John: Yeah, yeah, blockchain and AI are very complimentary technologies for sure. I think, yeah, in some sense, I think that the reason that they're complementary is because, OK, here's, here's what I think is really good about blockchain, is that blockchain allows the digital world to act like the physical world. That the physical world has certain properties, especially non-fungibility. It has scarcity. It has a lot of qualities that the digital world doesn't typically have. And blockchain gives those properties to the digital world. And then AI allows machines to operate in an almost unlimited way in the digital world. That when you think about what happens on a computer, what, what ultimately happens on a computer is always the same thing. There's some goal and there's instructions to the machine to carry out the goal. And if you can figure out what are you trying to accomplish, what are the instructions to get you from where you are to where you want to be, the machine can do it and increasingly, AI can do, can formulate those instructions, you know, you have to supply the goal and oversee it, but at the end of the day, digital activity is It's kind of solved. The digital problem solving is getting close to solved by AI but then blockchain connects that to the physical world. And then I think that that's what allows, you know, with the wines where you're talking about authentication or, you know, you have images, where did they come from? What's the provenance? Is this a real image? These are problems. And of course, when it comes to money and commerce, blockchain is, that's the original use case of blockchain. But I think that blockchain is ultimately going to be the link between AI and the physical world and. And that that's why I think they still go so well together. OK, well, Andrea, this has been Andrea, I should get your name right. You gave me a lot of words that were hard to say in the biography. OK, so you're a professor, maybe you could give some advice to young people starting out here about like at this moment in time, you know, what advice would you give to a young person to try to position themselves well for this, you know, coming time of change.
Andreea: That's a hard question to. I think everybody still needs to learn. You need to learn and know and understand things for yourself. You have to have a strong base to be able to use these tools to their fullest capacity. You have to be firm and confident in your knowledge because otherwise you can get derailed really quickly with these tools if you're not. So while these tools are very useful and they, as you said, they're coming, they will be here and we will need to be able to use them. We ourselves have to have the foundation, a very strong foundation to be able to use these tools at their fullest capacity in a way that helps the people we're trying to help or selves, at the end of the day, depending on what. We're trying to achieve. I think keeping being flexible is important, but don't underestimate the importance of a solid base of knowledge and understanding. I think that can be a tricky path. I think it's very tempting. We have all these tools that we can use. Why do I need to know this or that or the other. If we outsource all the knowledge and we don't hold on to it internally, I think it's a very slippery slope that can be laid, led astray very easily. I mean, information in the digital world can become, can be corrupted, can be misleading, can be faked, can be simply wrong, in so many ways. So if you, if you don't know enough to tell them apart, then you're not in a good position. So I think classical education has its importance and has meaning. With that being said, using all the tools in the toolbox is amazing and such a, such an exciting opportunity, such a ride, honestly, and it opens paths, ideas, doors, visions, that probably we would have had a hard time accessing before, so enjoy the ride.
John: Enjoy the ride. That's a good. I’m going to put that in the show notes. No, that's right, but I think you're totally right that AI needs to supplement education. It should be a tool to help you get educated, but ultimately you have to manage the AIs, and it's. Like any manager,, a manager who knows, I really think that for example, you're the CEO of a small company, you should be able to do at some level everything that's happening inside the company, same inside your lab, I'm sure you should be able to do all of the things and maybe you can't do them great, do them very well, but, but that way when people are doing work, you have some ability to assess it. And if you can't review the work of the AIs and have some idea whether or not it's good or whether or not it's correct, I think you really aren't going to be a good manager of those AIs, so. I do agree that having a solid education is important. It's tricky because we were educated in the before times, so we were lucky.
Andreea: I agree. I feel so lucky in so many ways. I was raised, you know, without smartphones and all of that, barely with phones to think of it, landline and dial up internet and so I have, in terms of comparison I can think of the before and the after. I have that information and, and now look at what I'm, I'm learning now and what I, what is available to me now, it's just super exciting.
John: Yeah, no, it's totally true. Yeah, I do, I do actually worry about the education of young people with these tools, but we will see how it turns out. We got lucky. I think we really got lucky. Like we hit the timing perfectly.
Andreea: Yeah, I was talking to my son about that the other day, and actually it's quite interesting. He's 18 and their generation is already thinking about these things. He was like, Mom, we grew up with this technology. I never knew anything with these technologies. I never knew anything else before that. What would my life look like if I did and what's going to happen in the future that the generation that is born now, what will they have that, you know, we didn't have and how will that affect their lives. They are thinking about these things, they are aware of how quickly things are evolving, so I think it's up to us to figure things out and, and guide them and it's up to them to try and stay grounded and find their way. I think grounding is really important as technology takes over more and more or facilitates things more and more, just taking a step back, taking a breath, thinking things through, you know, I'm getting very philosophical here all of a sudden, but I think it's tempting to fall into the river and, you know, row really fast.
John: Yeah, I agree with all that. Well, it's been a pleasure having you on the show.
Andreea: Thank you very much. This was such a great conversation. Thank you for having me, John.
John: Yeah, it's great. OK. Oh, and how can people get in touch with you? We'll put your information in the show notes. They can reach you on, you have a Facebook page, also on LinkedIn, they can reach out on LinkedIn to connect if they're interested.
Andreea: Absolutely, yes, and I have a YouTube channel and I have an email address. So, yeah, all, all the ways.
John: OK, perfect. Awesome. Thank you very much.
Andreea: My pleasure.
John Ennis is a leading expert in sensory science and consumer research, with extensive experience in statistical analysis and product development methodologies.
