🐙 Lunch & Learn: Future of AI In Data Analytics w/ @AlexTheAnalyst | Tina Huang

0:00All right, go live. Okay. All right, friends. And we are live. Is it working? I think so. All right. Okay. Okay. Wait. I'm going to. It should be okay. Okay. Yes. Yeah, it is working. Wait. I don't know where the audio is coming from. So, okay. I see. I always have this problem. You know, it's just call it part of the. We literally test. Like 10 minutes before. So I think it should be good. Yes. It's good now. Hello. Hello, friends. Bonjour, live it. Legends colliding. Oh, yes. Nian Alex finally, huh? Christian Lewis. Hello, Tina. Hello, Tina. How's it going? We're doing good. We're doing good. Thank you guys so much for joining. So I'm sure I don't need to give that much of introduction

1:03to the amazing person that you see here. But we will wait for maybe a couple of minutes or so before we jump straight into it. Hi from Turkey. Where are you guys coming in from? So Alex, you are an East Coast time. Remind me again. I literally, it's not Virginia. North Carolina. North Carolina. Charles and South Carolina. I was close. Very close. Very close. I lived in North Carolina for like 10 years. I lived right outside Charlotte and then went to school at East Carolina University. So I'm very familiar with North Carolina. The Carolinas. I actually was going to go to Duke at some point, which was in North Carolina, I think. Yeah. Great school. Yeah, I had two friends go there. Really good school. Basketball. Apparently. Very, very big basketball school. And right in that triangle is also UNC and C state wake forest. All really big competitive basketball programs. Sports people. Yeah. Yeah. That was smart. That was for smart people. That's why I didn't go there.

2:03Mm. That's true. Apparently people would go literally go there just for sports. Lima, Peru. Hi from Peru. Hello. The creators. I have learned a lot from pleasure to see you both. Well, thank you. Audio is good. Thank you. We're all good. Everything is good for checking. Everything that they literally made no sense. Everything is good to go. Hello from East Tennessee. Hi from Toronto. Ah, yes. University of Toronto, my undergraduate. I'm from La Marazgana, Kansas. They long I got school. Oh, yeah, true. It's Friday. Thank you for popping in though. Greetings from Indonesia. Texas. You are very perfect. We are very perfect, Alex. Woo. I love Tennessee and Texas and Canada. Can't play school. Azerbaijan. Indonesia. Texas. Kenya. West Virginia. Hello from Australia. Oh, thank you for making it. Australia. It is pretty late in Australia. Vancouver here. That's where I keep going to. Sweden. Habiba from Agana. South Florida.

3:03Cape Town, South Africa. I have a couple of Texas people here. Greetings from Qatar. Thank you guys so much for joining. We are so excited for this. And Alex and I, you know, we purposely made this. I don't know what's going to be on the public. I'm going to be on the public. I'm going to be on the public. I'm going to be on the public. I'm going to be on the public. I'm going to be on the public. I'm going to be on the public. I'm going to be on the public. And we are so excited for this. And Alex and I, you know, we purposely made this. I don't know what's going to be on the presentations. I'm really interested in learning the future of AI and data analytics. Because Alex and I, this is the best person to talk about this. So, yeah, I don't have any clue as to what he's going to be saying as well. So I'm genuinely very, very interested in knowing about this. Especially, Alex, I'm sure, like, you're like in the midst of it. There's so many people talking about is AI going to die? Like, is data analytics going to die? Because AI, like, is this the worth? Is this the career that's worth going through? All of these questions we will be answering. So I'm really excited to hear about that. Yeah, me too. I'm excited to see what I say. I'm excited to see what I say. Okay, good to be on the event.

4:04All right. Okay, so before we're going to get started, I just want to say that Alex has a platform called Adelaide's Builder, which is going to be launching very soon. If I'm correct in that, Alex. So if you have not checked it out, I'm going to be pinning that in the pinned comment. So you should definitely check up on that. Sign up for the wait list so you get notified. And for, sorry, I just literally woke up. If I don't make that much sense. Yeah, is that correct, Alex? It's going to be coming out pretty soon. And then people on the wait list are going to be able to get notified. Yeah, you go inside, you go to analyst builder.com, sign up for the wait list. You'll be a beta user. You'll get everything really discounted. And you get to like beta test the a new learning platform, which will be really cool. So I'm going to pin that. And also, I just want to say lonely octopus applications are now open. So if you're interested in learning about data and AI skills in order to work on real freelance projects from companies,

5:07you can also apply on lonely octopus. I'm going to pin both of these. So please do check it out. Let me just pin that. So pin comments and. So I'm going to pin this message. And yeah, should we get started, Alex? You good? Well, let's do. I'm ready. All right. I will share the screen for Alex's presentation. I guess started as always. You're just going to be you're just going to be sitting there. It's like watching this. Yeah, you let me know when I should move the sides. Yeah, I'll, I'll let you know. Yeah, you know, when you're going to be doing that. And also, as per usual for these live streams, we have any questions in the chat.

6:09Any of these I'll be moderating looking at those as well. So feel free to drop your questions here. Also, there will be a Q&A afterwards. So make it a chance. I'm not going to be looking at the chat at all. Like, I don't even have it up on my screen. So I can't see it. So you just tell me if there's a good question out there. But you go ahead. Yeah, this is the the future of AI in analytics, lunch and learn with Alex animals. I'll be there. But you can go ahead and go to the next one. Or go to the first one. All right. Oh, yeah. You can skip the you can go you you do it. You want to do I'll just say you're for a second. Well, here's the agenda. So we're going to do an introduction. Alex is going to talk about the state of data analytics, present to feature and looking beyond and ending with a Q&A, a spicy Q&A. So. All right. Alex, give your introduction. All right. I'll tell you a little bit about me and why I'm super, super, super interested in this topic in general, just because. I've just been asked by it from a thousand people. And that's probably a very low understatement because there's probably a lot more than that.

7:12And I've just I've had to dive into it like crazy in depth. But if you don't know who I am, my name is Alex freeberg. Better known as Alex analyst on YouTube. I create content on data analytics. That's that's really what I do on YouTube. But besides that, I was a data analyst for many years. I was an analytics manager. Now I run my own consulting company for analytics, consulting with companies to help them with their analytics, go to market strategies, all these other things. And so that's who I am. And A.I. is honestly, it's just like completely changed the game. It's everyone he's talking about. It's the next big thing. It is amazing. I love it personally. I think it's amazing. And there's a lot, lot, lot that is going to change over the years. And I have talked to a lot of people. So everything that I'm going to be talking about isn't only my viewpoint. It's also a lot of other people that I've talked to who I believe are a lot smarter than me in certain areas. And that's why I go to them. I'd be like, hey, listen, you're, you know, the director at this company in analytics.

8:13Tell me about how you see A.I. in fact, impacting your company. So I would go and talk to them and just kind of get their insights on it. All of this is off camera, just like friends and people I've met through YouTube and people I've met through like my work. And so I don't really, that's not the kind of content I make on YouTube. And so I didn't, you know, post that stuff, but maybe one day. So I have a lot of firsthand experience as well as I'm getting a lot of information from people who I think are probably even in some areas more equipped than I am. And so you can go to the next one. So I'm going to talk like the agenda said earlier. I'm going to talk a little bit about what it is now. So how we currently do things in analytics, which is going to be pretty simple. And then I'm going to talk about how we're going to be doing things going forward. I'm kind of talking about the next like 10 years. And then we're going to talk about the future future, which is 10 years and beyond. At least that's like my scale. So whenever you're ready, you can go to the next one. There we go.

9:13So this is a state of data analytics. And if at any point, Tina, you have any questions or you think something's interesting. You stop me because I will just I can go. I'm going to start these like talking wheels and I'll just I will never stop if you don't stop me. So the first thing is that right now, if you look at analytics across the entire, I'm just going to talk about the US, but it really applies to you can apply it to almost anywhere in the world. But in the US right now, you'll see analytics teams or data analytics or data teams that have someone as a data analyst on their team. You see these mostly at medium and large size companies. You don't see a ton of them at a lot of smaller companies. Or if you do, it's a single data analyst, right? You just have one person. And so at these medium and large companies, they have some type of data team for their data needs. Now, I'm going to take one tiny step back when I'm talking about companies I'm talking about companies that use technology as their main.

10:17You know, that's what they do. They need technology to provide insurance to provide, you know, whatever I'm not talking about like a mom and pop shop who. You know, goes out and cleans homes. They're probably aren't going to be using data a lot, right? And so this is mostly for established companies medium to large size. That's where a lot of these data analytics teams are really established. So they have, you know, dashboards and reports and flows of data that they have as part of their process. They use that data for an intended output or goal for companies or they sell it or whatever they want to use it for. Now, this is important because in future slides, you'll see that this is going to change. And so that's that's where it is currently. And so you see large tech companies, Tesla, Twitter or sorry, X, Amazon, all the Netflix, all these large companies use massive amounts of data. They have huge, huge, huge infrastructures and teams dedicated solely to data.

11:20That is all they do. Now, what do data analysts do specifically on these teams? This is a very content simplified version of what they do. And I just wanted to keep it super simple. But on like a super simple level, our end goal or what we output is often dashboards and reports for stakeholders or insights as well. But that's usually in the form of a dashboard or reports and Excel file, you know, an email. However, you want to write this. So we produce these insights. We produce these dashboards and reports for people to then utilize on their teams. So that's what data analysts do right now. A lot of the work that goes in before that output is we clean and we analyze this data. So we're using tools like SQL Excel and Python. And those are three of hundreds. And I say that and I, these are like those are the three that I teach everyone or I teach a lot of my channel to learn because it's very transferable. You can use Excel in many different variations and different tools that are Excel-like.

12:24Same for SQL. Every database you have to be able to get in the data and use it. If you know SQL, it's pretty transferable to a ton of other proprietary tools that companies may have. So these are some of the tools that you might use, but there's literally hundreds. And then at the end, you know, a lot of what data analysts should be trying to do is also automate those repeatable times. So especially with reports, dashboards, you know, CSV files, all these things that you're sending out to people every month, every week. Should be trying to automate these things. Now, this is where we're currently at. And I'm saying current as in. AI hasn't been fully integrated yet. And it will in the future. So look at it right now. Even today, I consult with a lot of different companies. I go and I'll consult with them and I'm like, okay, you know, show me what you got. Almost none of them have AI anywhere in their tech stack, except maybe if they're like asking it simple questions here and there for code snippets, things like that.

13:26It is not integrated. What I'm talking about is when I say it's integrated, I think more like Microsoft Copilot, you know, Copilot is going to be launched sometime soon. They've been teasing it for the past like six months. It's going to be launched sometime soon. And Copilot is very integrated into outlook into word, all Microsoft products all the way up into Azure. That is integration. And so when you're able to use these LLMs and AI at every single level, that's when it's going to be integrated. So we're not there yet. We have not gotten to the point where AI is actually fully integrated. We can use AI at a very simple level. But it's not fully integrated. So that's where we are currently. Now I'm going to go to these next, the next one is just a second, but the next two, I'm going to talk a lot more about because this is where I think most of the conversation needs to be held. At least on this topic. So I'm just kind of brushing on like here's what here's where we're at. This is where most people are comfortable with. This is what most people know. And then you can go to the next one.

14:27And I'll talk about this one for quite a while. And then the future one I think is even more interesting than this one. Any, am I talking too much? Or can I keep going? Please keep going. Yeah, no, it's good. I think it's a good foundation for everybody. So yeah, it's good. All right. You interrupt me at any point. You just, you just stop me. So now we're going to look at present to future. We're going to look at present to future. So we're going to be taking a look at from kind of now to the next 10 years. That's where this timeframe is. I didn't specifically put that. But that's where this timeframe is. It could even be a little bit shorter. Just depends on how fast things move. But I'm going to kind of talk why I'm saying 10 years instead of think something like three to five years. And why I think a lot of people are a little bit too positive about the integration of AI. In systems. And how that will actually look.

15:28Now, earlier I said that we have data teams and data analyst traditionally at medium to large size companies, not a ton of small companies hire a lot of data analysts, maybe just one. In the future, starting now. And I, I have talked about this actually for the past six months. I think in the next two to three years that this first bullet point is going to be very true. But then even more so in the long run, which is data teams are going to be a staple even small companies all the way up to large companies. And we didn't include small companies or I didn't include small companies because it's the again, most small companies don't have very large data teams. They maybe have one like database developer or they have one day engineer, one data analyst. It's pretty small. And then they have other parts of their company that are more important at the moment. These are things like sales people. These are people. These are things like IT support all these different areas where they need people in those spots so they don't have the budget or they don't invest in a data team in the future.

16:29Three years, two to three years is kind of my estimation is a lot of more small companies are going to be popping up not only popping up, but they're going smaller companies that are already established today are going to need to be growing and investing in AI. And because of that, most of these people are not AI experts. Most people out there and especially for people in the data world like you and I, it seems very obvious. It's like, why don't people just do it themselves. But for other people, layman, I talked to my father-in-law. I talked to other people who I would consider in if my father's father's father-in-law was watching this. It was not an insult. Other people who are kind of layman in the area, they do not understand it in the slightest. Now, I think I am very, I understand a lot and I could totally see why people are like, well, you're not going to need data engineers in the future. You're not going to need data analysts in the future. Data scientists, AI will be able to do it all. But that's not what people want to do and they don't want to be the person to have to do it.

17:33So how are they going to integrate AI into their systems? They're going to hire people. Specifically, in my opinion, I didn't write this down, I should have. I think a lot of people are going to be growing their teams, starting with consultants and freelancers, who want to start small and then grow their team over time. So, for example, I was just working with a client who was like, hey, I really want to know how we can integrate AI into our current tech stock. So I came in and I was like, okay, let me take a look at what you've got. Let me take a look at what you're trying to do with your data and all these things. And so I looked at I was like, okay, you have your infrastructure is fairly good. You have good, a few pipelines that are moving, but it's not crazy advanced. I was like, and I kind of pinpointed some of the areas that I think needed improvements in order to start utilizing AI, especially in the coming months or in the future. And so I said specifically, I really think you need to hire on a data engineer who knows AI for this use case.

18:33And I've recommended data analysts for other use cases and data scientists for other use cases, but this one I was like, you need a data engineer who also knows like AI and the coming technologies. And so right now they're looking for a consultant who they can bring in, not somebody who they're going to hire on full time. So they're looking at freelancers and consultants. This I think is going to be a huge trend in the future where companies are going to see the value of AI and they're going to be like, wait a second. Our competitors are doing that. Right down the road, this other company is doing it and they're seeing this impact. We need to have this, but we don't currently have the budget. We need to start small, start with a freelancer, start with somebody to come in. That's kind of what I mean by it's going to spread to those small companies quickly. And then they'll have to grow over time that infrastructure. Now, if you look at, I'm not going to go crazy technical about it, although I could because I love this stuff. Current data architecture is really great for current data products for power BI and for all these different things.

19:38When you think about what you need it for in the future with AI, it's going to be different. We have different tech stacks, we have different tools, we'll have different data architecture. And the companies right now, most companies right now are not even at current technology. They're five years behind, they're 10 years behind because they don't invest in their infrastructure as much as they probably should. For example, I've worked with many companies over the past year now and I'm like, okay, you guys are still using on-prem servers, which is fine. I have nothing and there's actually great use cases for on-prem servers. I was like, why haven't you moved to the cloud yet? It's too expensive or it's too complicated or costs a lot of money to migrate our data over there. Or we have these dependencies on our on-premise servers. And so we can't, we don't feel like we can do it. And I'm like, okay, I hear you on those things. But what's going to happen in two, three, four years when you need to do X, Y, Z. And they're like, well, we haven't thought about that yet. I'm like, well, you need to start thinking about that because it's going to happen.

20:39Like it's coming. And so it's not just a plug-and-play with AI. It's an investment in the future of AI and so there's a lot that goes on with that. Most laymen, most people, just people who have an MBA, people who have went to, you know, like a business school or went to school for, you know, really almost anything else. If they don't understand this or don't have experience using data, creating pipelines, connecting data sources, all these things that AI have to have, AI is not just going to be able to magically do it for you. It's going to take a lot of work and a lot of investment in that infrastructure. And AI soon might be like, you might be able to use it with your current data architecture. But especially as it gets more advanced, it's going to take a lot more than what you currently have. And you're going to have to upgrade. So all that to say, I'm going on a tangent on just that one point. It's a lot more complicated than what most people are going to be able to handle.

21:39So they're going to have to hire on consultants, freelancers to come in, especially at smaller companies who haven't really used them before to invest in that. Some medium-sized large, they already have data teams. They'll just bring in someone other hire someone on full time for their AI initiatives to help them plan for the future. That's in my opinion, for a lot of people that talk to who are analytic specific directors, managers, who I know personally, that's essentially what they're saying to you. They're like, yeah, we're not 100% sure, but I'm already talking with our CTO about investing in something AI related. We just want to bring in like a consultant to get us up to speed and how we can do this. So that's that point. I wanted to, sorry Alex, I wanted to comment on that and there's a few questions from the chat as well. So first of all, everything you've said is I was just nodding my head in the entire time because 100% seeing this and it's not just seeing this in a way because we're building our own products as well. And what is it's like, it makes a lot of sense, especially in this day and age, hiring on people as a freelancer, people coming with a certain skill set, potentially turning it to full time later.

22:49It makes a lot more sense, especially in this economy as well as opposed to just hiring people full on and then just hoping that they have like the correct skills that are trying to train them in doing that. So I've been seeing this everywhere from the people I talked to with Lonely Octopus, the companies that we work with. That's where their their head is as well. So they incorporate AI in some fashion into their companies, but they're not ready to just like straight out investing it, right besides AI is so new like who are you going to hire that you can just say like very confidently this person is going to come in and build an AI team. Like you first got to figure out what that even means, like what's that specific use case for a specific company. I agree with you on the sense that that data teams are going to be a lot more stable because it's not really like a one size fit all kind of deal. I think people have this misconception who haven't used AI extensively. It's like, oh, like AI is just going to come in and do everything for you. No, like it has to be adjusted to a specific use case like an example that I like to give a lot is say you're a health care clinic, right.

23:53Go and just generate content, for example, we're like, you know, do things without considering privacy matters. So those privacy matters are going to be very specific to your specific company. And similarly to other companies, the other things that they care about AI in some ways is actually like a very small component. It is that part may be static. You could be fine tuning up, but even if you're just directly like using that part of AI, it's the things the architecture surrounding that all the things that the checks that are built in how it is that you're monitoring AI, making sure it doesn't just go and do things that you don't want to do those data infrastructures engineering infrastructures is I think where a lot of that investment is happening right now. So kind of just wanted to like add on to that point. So absolutely, I 100% really like freelancing is just huge right now. And anybody that has that has like AI skills like honestly anybody that has data and AI skills right now.

24:53Let's just say like the bar is low right now. So learning these skill sets is so important. And just like putting yourself out there talking about these topics, people get reach out about it because there's just not the many people who have the specific skill set yet. So if it's something that you're interested in doing and you're like, is it too late to learn about this? Absolutely not like this is the time to learn these skills and you'll be at the cutting edge of where it is that companies are looking for. So 20 years you'll be like, I was one of the first people using AI at my company and they're going to be like. Their mind's going to be blown. It's crazy, crazy early. Yeah, there's absolutely no no there should be no thought in your mind that you're too late. There's hardly any people in the actual workplace right now they're utilizing AI because the products aren't there yet. And so if you start using it now, start getting familiar with it when these kind of enterprise level AI products come out, you can be one of the first people to start using it. Yeah, you'll be you'll be set. You'll be doing really good.

25:54I agree. So there's a few questions from the chat. It's what would be your advice who wants to change his career path to data science after 35 years with no experience. So do you would you recommend for example Alex learning about like going down I think we're relying on that like still learning about this but then also learning about the infrastructure how it is that data. And AI there's a crossover there is that what you would recommend after 35 years. What was is that what the question was. Yeah, so they've been someone who's in the beginner who's in the beginning wants to change their his record path to data science after 35 years with no experience with someone who's trying to get into that field right now. What would your recommendation be in terms of learning those data skills and how like what kind of AI skills should they be learning. Gotcha. Okay. So I just super I did a live stream. The other day I had a super similar question and so I'll give you essentially the same answer, which is you know you've been doing something for 35 years.

26:58Age on some level does play a factor when you get to a certain age there are expectations that you know things right that you're experiencing you know things. And so what I've always my recommendation especially for a little bit older is that hopefully you've been in a field that has some type of domain experience that you can utilize whether it's healthcare finance construction agriculture could be anything. So if you know in-depth knowledge about that you're trying to move to the data side learning skills like you know well if you're doing data science maybe a little bit different but for data analytics you know sequel excel Python tableau AI and how you can use that. That this technical skills are probably not going to be what sells you as much as your domain knowledge for example when I was working back in my healthcare company or my health care and the legs company. We would hire on nurses who were nurses for 30 years but knew the data inside it out because they were on the front lines or doctors who were you know just doctors for 25 years and they were switching over to the data side of things.

28:03And because of their domain knowledge they were super super smart and what they did they just had to learn some of the data skills they just had to learn how the data sits and how it's stored and what it actually means being able to interpret it that domain experiences what really sets you apart from other people and can help you make those transitions. Oh, you're muted. Oh, sorry. Thank you for that. Should we move on? I'll also be trying my best answer questions in the chat so please be patient with me. I'm trying to multitask a little bit here. Yeah, I'll be gone. Oh, it's only 12 30. I got so much time. I'm booking it but I'm also got plenty of time. Okay. Next, the next thing is in the future going forward there's going to be a lot of AI augmented analytics tools. Now I'm just going to read this and I'll kind of talk about that as that's going to help provide real time suggestions and automations to analysts as they work.

29:04This is referring to AI integrated systems. Now I just went out to the SAS conference out in Las Vegas. They literally had a booth on AI augmented analytics which I had already kind of been somewhat familiar with but they just kind of talked about how this works or what it looks like. And there's not it's not just SAS right it's all these big companies are creating tools like this that are going to be integrated integrated the one I'm always going to keep referencing is Microsoft co pilot although there are tons out there but co pilots are one that most people are familiar with and that's where you're going to have Excel but you'll have AI some little sidebar and when you need it you just go in and you type it. It's going to offer suggestions it's going to offer you know you're going to say hey how can I automate this with a macro and it's going to kind of explain it to you and maybe even do it for you in the future who knows. But you'll be able to be working and using AI in the same system so as you are trying to dive into your client client comes to you says hey you know here's what we're trying to solve with our data we know we're having these gaps and we don't know how to solve it you know we don't know what to do.

30:12You know can you give us some insight into this now you having the domain knowledge and the analytics experience that's why they're asking you you go in and you start looking at everything you know doing everything you start asking questions and you're like okay I'm seeing this and you ask AI hey how do how does this line up because this is what I'm seeing it'll give you suggestions or things that should prompt you. And people can already do this a little bit with chat you be T but it's just not as into it's not as super helpful because it's not integrated it's not there looking at your data it's a copy over the data doesn't have all the context and all the information. And so it gets some things wrong but it gives good suggestions imagine that but just even better that's what this is going to look like it's going to be you know really kind of like a little assistant at your side at all times. In my opinion because I've been actually using chat GVT or different you know I've been using a little bit of bar a little bit of claw I've been trying these out with my workflow and they're very helpful but I have seen so many downsides I'm just like oh geez it does not do this well like I it just is but I'm hoping it'll be even better in the future because for me personally I do not see it as it coming for my job especially as how much I've been using it I'm like okay this is not close yet but how it could be used in the future.

31:27I absolutely see it being really really helpful with our work so giving a lot of suggestions automations. So something I mentioned that we currently do now or in the previous slide was we build dashboards and reports that's a lot of what data analyst do for their end product. Now we do a lot more before that with data cleaning and exploring the data and then working with data engineers and data scientists to you know help create those transformations in the ETL process. This is a lot that data analyst do but for that last piece the last mile of creating the dashboards and reports we finally did all that pre-work or about to send it off to the client we build these things and I think that not yet but very soon in the next year or two creating dashboards and reports will probably be one of the easier parts of being a data analyst. I don't I don't think you're going to need to invest as much time into learning things like Tableau because Tableau isn't going to be coming out with some AI stuff. Power BI is going to be coming out with some AI stuff. I don't think you're going to have to invest at as much time learning those skills as you will more of the concepts of data modeling data cleaning data exploration ETL process.

32:39Those more complex things that have a lot more nuance. Dashboards and reports I think very soon and I've already seen demos demos from different companies you're just going to type in hey I need a chart displaying this information it's going to pop it up for you now it's going to be perfect probably not. But it will get you 70% of the way there and then what you need to go in and do is just do that last 30% or maybe the last 20% or in the future 10% where you just need to change some things send it over the client they're like okay yeah that's good but we actually want this and you go and you revise that you finalize it you're good to go. So I think a lot of the dashboards and reports in the future aren't going to be as heavy of what a data analyst may do although it still will be something that we're probably in charge of and so that's just that is a lot of a lot of upfront help with the initial development but then you know we still need to be there to finish it. And then the next thing companies will have to upgrade current architecture for more advanced AI now this is what I was just talking about on the last slide is the exact same thing and so just highlight this again is that our current architectures I want to say good for some companies is not good but AI requires data they require a lot of information with even things like oh geez it's on the top.

34:07So I think I'm going to do my time data documentation a lot of companies don't document anything you would be surprised how many companies I've come into or works for in the past like as a data analyst that had absolutely no architecture no systems in place for what they actually do they just it was all in their head everybody knew everything but none of it had been documented so even when I say architecture that's usually referring to database schemas how your database pipelines are flowing this can also refer to just your processes of your company which are going to be more important in the future because AI is going to take those things into consideration you'll give outlines hey here's how all of our data comes in here are the sources here's the oh geez I'm blanking on words today it's a data dictionary so you have this data dictionary that tells you exactly what these columns are exactly what the data is it can understand if the data is messy or if there's something wrong with it. So I think I'm going to do my time data but it needs a lot of context and a lot of really good building of infrastructure which most companies I would say the majority of companies 70 to 80% plus don't have a great data infrastructure for even today's purposes but especially not for AI.

35:25So let's just again that's kind of what I'm talking about with they're going to need people to come in they're like what I I know this is already happening but it's going to happen more which is a company is me like we need to use AI let's start using AI all right guys let's start this AI initiative and they start trying to do it and then they don't get the results they wanted at all like they're getting their results are like 50% of what they want like 50% is spitting out somewhat good content 50% is giving out nonsense that makes no sense and they don't know why they're going to have to hire people on who know more than that. So I think I'm going to have a lot of people who can come in and be like oh that's because of this your data is sitting like this it needs to be sitting like this so it's there's a lot of nuance that the vast majority of people don't know or don't understand yet so that's where we are in the next 10 years now let me explain actually what last thing on this slide that let me explain why I'm talking about the next 10 years and why I'm not talking about five four through three years which a lot of people think the reason for this is is

36:25actually somewhat what I was talking about a little bit ago is that the assumption for people who are saying it's going to be in two years will be completely different have never worked in depth at an actual company before in my opinion because I've worked at a lot of companies I've consulted for a lot of companies and goodness gracious most are not anywhere near being where they need to be like I was talking about early with on-prem versus the cloud there has to be an investment in the cloud. There's an investment in money talent and there has to be a purpose driven company that's actually making those efforts to upgrade their technology over the years now I worked in healthcare and it is semi laughable how behind healthcare is with their tech stock in fact it's kind of like a running joke in the healthcare data world that the healthcare is always 15 years behind with technology and it's not a joke it is real so I have to imagine just the end of the day

37:25the entire healthcare sector which is like hundreds of billions of dollars per year just in I'm probably more like let's say a trillion dollars a year in revenue just the United States for healthcare companies someone go get that for me because it's probably well you have to imagine these companies are very behind even the company that I used to work for that was a fortune 10 company we were just when I was getting there transitioning to the cloud like just making the first steps I helped with that transition then I got hired as an analytics manager to help migrate our entire IT department to Azure and so weird like you have to imagine we're behind right so think about it three years five years ten years I don't know who Jeff Dean is sounds like a smart guy though I've never heard that name sorry but he's the big shot at Google share his link in his like whatever do you have to post for this because I had no idea like who I had no idea he shared it is it the same one that I'm thinking of because that's Google deep mind chief

38:36chief so yeah sure that with me that I've never heard of him yeah but if anybody is here for Jeff I apologize that we really had no idea he's not going to be speaking that I'm aware of at least that Alex is aware of that's why I just asked him so sorry to disappoint here but we got to thank him for sharing it yeah he's the top Google fellow awesome sorry didn't interrupt no that's fine yeah so we just one clarification so he was saying very similar things is that or he was saying just that on that very topic I don't know so yeah like can anybody are you guys able to share links you might not be able to share links but if you could we just migrated though I'm like I mean I'll check is like wait I'm gonna just is it this guy yeah I had no idea so is it this guy I'm not sure you guys can share links yeah he's the top Google fellow yeah apparently a bunch of you guys came from Jeff Dean all right Alex we better better put on the good show

39:50Oh wait people were here to steam Jeff Dean yeah apparently he just shared a link to the stream and people thought that Jeff Dean is here but apparently he just shared it like well I don't care if you're asking Jeff Dean I appreciate you man if you're out there thank you I appreciate it I appreciate as well yeah yeah so let's so yeah I don't I'm not gonna keep harping on that but that that is my thoughts on the next 10 years that's why I don't think it's gonna happen in three to five years I'm talking about 10 years span because a lot of companies are gonna need to spend a lot of money and like you were talking about earlier especially in this economy it's hard to invest it's hard to invest a lot of money if your business is already working but in two to three years in that five years span when AI is really ramping up and becoming a lot more integrated with a lot more companies there's gonna be not a lot of a reason to there's not gonna be a great reason to not invest in it it's actually gonna make a lot of companies behind and they're gonna see that so in like three to five years I think companies are really gonna start investing a lot more into it and then in about the 10 years span

41:00you'll see most companies having some AI integrated with them it's just it takes a long time for companies to catch up it really does but you can go to the next one now because the next one I think is the most interesting I this is like this is a long term guess and this is not just my own again I want to reiterate I've talked to a lot of people on this and so when some of these things are not my original ideas but then we would chat about it and I we go have some back and forth so these are some of the ideas of my community that I've talked with about this so this is looking beyond this is like 10 years plus I keep hitting my wait Alex I think I don't know if you died where I died where I died I think you died

42:04I think you died yeah Alex your mic let's see on Riverside you hit your mic and something happened a little break well in that eyes as I Alex figures out his mic situation I will continue to look at some of the questions that are here also thanks to everyone for asking so many questions I apologize I am trying my best to answer all of them you hear me yep very good now so here's a question thank you Lloyd for the donation I've got VBA tableau sequel and some python I work in staffing industry but I explore stock data my free time what's that was must I take as someone who wants to incorporate AI into their life slash career

43:04we can hear you Alex you want to take this one? yeah it's really interesting right now because everyone wants to know how they can use AI there's not a lot out there right oh I say that I say that I'm like there's not a lot of integrated solutions right now there's a lot of tools so you can go and use chat GPT or bar or quad those are simple there are other ones as well that are being built and you can just google it like say look look up AI tool for healthcare there is one out there I've looked at it so you can look at all these tools and start messing around with them for me how I've been using it as I'm like okay I you know I know I want to write something for sequel and I will plug in you know kind of like my my columns or something or like a little sample the data and I'll ask it to generate it it'll try to generate it it'll get you like 90% of the way there and then it revise a little bit that's all I have to do right now but if I'm writing Python I use there's a I already integrated with Python for a lot of tools like visual studio code I use get hub co pilot so there's a lot of different things that you can use just

44:14get a fine one out there for you but there's lots out there but it's not as like crazy integrated as what I'm talking about like on the last live where it's like in there it's integrated it's like working perfectly that's that's kind of this looking beyond section right here yeah I'm gonna kind of comment on that a little bit so from my just just to kind of clarify so I'm coming from more of a like data science and even like some engineering data engineering engineering based perspective right so I would agree and her is integrated solution not the many I think people are just trying to figure out what to do right now it's like there's a lot of startups that still like this is like the solution for this and then if you actually use their thing like that is not the solution for that maybe it's like a single solution however with that being said though I think these are a fundamental like skill sets that you should learn about AI right now as a data analyst but I think again I'm coming from more to perspective from an engineering perspective as well as a data science perspective so I talked about this before but in terms of I think I

45:15I answered this in the chat so prompt engineering right if you're going to do anything learn prompt engineering and just understand how to communicate with AI this is like a very basic skill and it's like almost a different mindset right because you're doing prompt engineering when you're communicating with AI there are certain ways of communicating with it mindset shifts and communicating with it which I'm not going to go into much detail maybe in the Q&A that you need to learn how to interact so with that unlocks a lot of potential even if you're just chatting with chat TVT for example if you know how to prompt engineer and how to interact with AI know what it is that it can do for you it can do a lot now from an engineering perspective this is something that we see a lot in the freelance projects for a lonely octopus there are like a lot of companies trying to integrate AI into their companies right now and the fundamental like skill set that I think people should have if they want to learn how to use these new tools is not necessarily going to these integrated solutions which as we were saying it's not it's like not really very yet but go to the basics right the basics of prompt engineering how do you use these large language models through API's like for example

46:23the API's for open AI right that's how you can interface with some of these models and then on top of that how do you build like a UI around that there's certain so that you're able to display certain responses that you're getting a certain interactions how do you go through and play around so that you're setting a different like system settings for how you're communicating with the AI as well as databases and language I would say these are like fundamental more engineering based things are really important to learn because these are the solutions that companies are looking for right now and as you play around with these with these technologies you will find that so much opens up to you it is so much more than just interacting with chat you be tea directly the ability yeah the ability to build certain products the way that you can use these solutions I use it in my own life as well like I have some like silly scripts that I have where you know I'm able to sell like a certain tone for for the large language model to communicate with me a certain like interfaces they're very simple UIs for this I'm incorporating my own personal data by using

47:30chain all of these things I really recommend just go into basics that was my round I agree with everything in fact that is something that I'm going to mention a little bit about the future of data analysis 10 years of beyond a little bit at the very bottom is what I'm going to talk about that that I completely agree that everything you just said very very good all right you want me to go yes please keep going it's 1250 we may run over I'm okay with running over if you are yeah let's like let's say the maximum we run over is maybe like 20 minutes we're so just so I know it's a Friday as well so I don't want to make people choose my mind for hours I got a I'll speed through this section but this is the most interesting one to me now previously or currently medium to large companies in the next 10 years small medium large companies will have it then looking beyond to start every company is going to be somewhat reliant for some processes in their company on AI that's my prediction

48:38um AI is going to increase speed accuracy and overall product or whatever you're building in my I really think it's going to be that big but with that becomes a lot of things right it's not as we talked about in the last one that's a lot of the what we talk about just a little bit ago is going to be even more so in the future um so think about it with you know uh just a data animals we're creating dashboards reports um might be helping with data pipelines insights in the future a lot of this can be done with the assistance of AI now even right now in you I'm sure you keep up with this stuff but even right now you look at things like AI agents if you if you're privy or if you whoever's listening if you've heard about that if you haven't go on um go on twitter there's some uh great people who talk about AI agents not only that but just being able to generate really good code um in the future you are going to have once that infrastructure is laid over the next 10 years

49:43people are really investing in this infrastructure they're investing in documentation creating great data dictionaries create processes and procedures all these things that AI is going to ingest when we are trying to then build dashboards reports data pipelines insights all these things it's going to have a pretty good idea now will it get a hundred percent right probably not right that's not going to be most likely the purpose of AI it's going to be getting your job done lot faster so you can do more work but you are still going to be needed to interpret what AI is saying making sure that you're prompting it correctly making sure that you're asking the right questions making sure that what it's giving is actually accurate and validating those results these are all things that data professionals whether it's in data engineering making sure the data is actually getting improperly setting up you know monitoring systems to make sure that AI is doing what's supposed to be doing all the way down to data scientists and data analysts using the data for whatever they're using it for these are things that you are still going to need to be monitored tested qAid and have a final say on

50:45I don't see and I say looking beyond it's kind of indefinite right but I'm thinking like 20 years down the road even 20 years down the road I still imagine and still I'm quite confident that humans are still going to be kind of the final say where we're not just going to ask AI it's going to be like oh that is the answer we can know it without a shadow of a doubt most likely we're still going to need quite a bit of interaction and checking in qAid all these things now the next part I talked about quite a bit so I won't go crazy in depth but last one is actually the one I'm most excited to talk about and hinted I kind of talked a little bit about what you were talking about earlier but this one says the infrastructure needed for AI will be crucial and will need data engineers analyst project managers and more for good results it's it's the human condition for it for anything data related data is only as good as or the AI that's going to be producing any results is only as good as the humans that are producing the data I think about healthcare all the time because I'm in healthcare but I know this is similar for a lot of other people

51:49but when I worked really closely with healthcare we had the messiest data you could ever imagine that's because of the systems where the doctor was inputting something freehand or he was having someone translate it after he wrote it on paper his assistant was translating it and putting it into their system these are things that I would consider the human condition right we're always going to mess something up we're always going to you know cause issues and so it's going to be really important that data engineers analyst project managers work together even just as they do now but even more so in the future to really make sure that the entire pipeline from the creation of the data all the way into the database data warehouse data lake wherever you're storing it that it is seamless and that is very very very very very difficult to do even today even at big companies that have thousands of engineers there's still bad data it just happens and so you're still going to need a lot of these people to ensure that the data is good so that your output from

52:51these AI systems are really good you're going to have to check them double check them and then you know will go on from there so that's that's kind of in the future like 15 20 years down the road I see AI being really really integrated but still needing a lot of work so a lot of hands-on people doing these things but just AI helping out a lot now when I talk about AI I've been kind of in in this whole conversation I'm talking about almost general AI general you know chat gbt type use cases I don't think and I this is something that I've actually done a lot of research on so this isn't my only my thought but I've been thinking about it for a long time which is we're not going to have a one AI system or we're not going to have one AI system that does it all and not everyone every company is going to plug in chat gbt which is what they're doing now for a lot of things and it works to an extent and it's really impressive but when you get down to the nitty-gritty when you get down to the really so many specifics and you need like an expert of an

53:54expert on in this domain we're going to have tailored AI's for your company or your domain so I keep using healthcare so I'm going to keep going um you we need a company that I work that I can imagine using an AI that was trained specifically on our data that knows it inside and out very very well and then is already pre-trained maybe you know we buy it from a company that pre-trained it to be basically like as smart as the smartest doctor in the world or the smartest nurses in the world and we use that as like that's who we're able to work with and this is going to be really really really important in the future because we want it to be for us general that those general AI's are great I use them a lot myself I already see a lot of limitations with them that I've used and I've personally encountered and I'm like oh geez here there are a lot of limitations but when you only need it for a specific use case let's say um you're predicting outcomes

55:00for a patient that was a company that I used to work with when I was first starting out we would create these this logic and we would predict okay um you know what percentage likely is this person based off all these factors to have a heart attack then the doctor would say okay that's over you know a certain percentage he's going to then test for some cardiovascular disease or he's going to do these extra tests that stuff is going to become a lot more popular especially in healthcare and you can imagine I don't want a general AI doing that that kind of knows it I want an AI that's only trained on just that that's going to get it right 99% of the times that of 94% of the time and 93 that accuracy has to be there because it's you know that's life and death when you're dealing with healthcare in some instances now for this the data analyst based positions are going to become a lot more niche now there's already ones out there healthcare analyst financial analyst marketing analyst I think you're going to see a huge rise in those because most

56:00companies aren't going to need just a general data analyst to do the technical skills they'll need someone to know the technical skills I don't think that's going away anytime soon but I do see the need that they're going to need somebody who's a lot more domain ex uh knows a lot more of a specified domain really well and also knows the technical skills because AI is going to help with the technical skills I think a lot more than it's going to help with the domain knowledge and so in the future this is looking beyond this is 15 20 years down the road I think that will see a lot a huge uptick in domain specific AI analysts who are going to be using AI in their workflow and they're not going to probably call it AI analyst don't just be an expectation at that point they'll just call it healthcare analyst need to know AI in the job description right but I think you'll see a huge uptick in that and that'll be kind of an expectation that you are going to be using AI for a lot of the technical stuff for building things and it's going to be integrated into the workflow but we need

57:00you to not only use AI with the technical things because that's going to be not crazy difficult at that point but we need you for your brain and your in-depth knowledge of that domain that's what we really need you for and so that's my that's to me that piece right there especially for data analysts that's probably one of the biggest takeaways that I take out of all of this is technical skills you need them you have to have them they're going to be an expectation and AI is going to help a lot with that but if you have expertise in a domain if you go to school for you know nursing or you go to school for finance or you go to school for construction or agriculture or whatever it is that domain knowledge combined with AI makes you way more important than just someone who knows the technical skills which AI is going to help a lot more with anyways so I know I've been talking about this on my channel for probably the past year or two

58:01is that you know technical skills are always going to be important try to find a domain that you'd like and you enjoy and that you can learn really in-depth find a mentor that can help you you know who can teach you the ins and outs of the industry find somebody who's you know who just you can chat with on LinkedIn or on Twitter or somebody who inspires you and reach out to them that's how you can find these mentors and you can kind of get connected to an industry that can really set you apart from other people so that is the look and beyond section I'm trying to keep this somewhat succinct because we were we were short on time I know we have a QA section but that is my that that is my current short-term future long-term future with AI and analytics and a lot's going to change but it's not changing as fast as I think most people thinking it's going to change quick but people will be able to keep up with it my last note before I handed it past the buyer review and this is just a really bad analogy but it makes sense in my

59:06head so I'm going to share it anyways is you know before the pandemic if you bought a house you're sitting pretty because that home value has increased and now it's really tough to buy a home right now right I see it the same way with AI right now we're at pre-pandemic we're at just the beginning you buy you buy into this AI stuff now you start investing in it learning it three years down the work down the road it's going to be amazing for you but for the people who didn't learn it who weren't investing in and didn't care about it three years ago that it's going to be really tough for them to break in and start using it in three years and be like okay I'm just now learning it whereas you could have had years of expertise and actually digging into it and figuring it out so don't get left behind I myself and not trying to get left behind you know don't get left behind start you know using AI start learning AI looking into it you don't have to be the biggest expert in the world but start learning that's my that's my last piece thank you Alex I'm going to add on

1:00:06to that a little bit because I cannot resist myself in doing so like yeah I want to say from a business perspective like looking at enterprises working with other companies that are again trying to incorporate AI into their work stream that this is where I'm coming from so 100% agree with you on all of these things one of the biggest factors is you know how like a lot of companies just ban chat gbt when they come out understandably so understandably be so because they don't understand it and they're afraid of issues are coming up but I think this is a a showcase is the human reflex very much so I don't understand something I am scared of it so I'm going to run away from it instead you need to bypass that instinct I don't understand it this is something that's clearly very powerful I need to learn more about it it's like instead of being afraid of something and running away you need to approach it I understand how do I incorporate this into my work stream and that is what sets apart someone who really catches on and takes advantage of new technologies and creates opportunities for

1:01:08themselves and other people as opposed to someone who is just kind of like following the crowd I genuinely think this and I also wanted to add like in terms of the infrastructure needed for AI I think from a data science perspective and from an engineering perspective I think there's a much bigger requirement now for people to have technical skills just like Alex was mentioning from an analytics perspective especially the use of development like how do you use APIs APIs are huge right now like before say as a data scientist sometimes you're just like tinkering with things by yourself and you make a model you launch out of these things but so much of it now is how do I use these open a eyes these like large sorry open a API for example where like these APIs that are outward there the large language models that are out there how do I use those things instead of just trying to build something myself like nobody is going to go build their large language model in most cases it's about how do I use it how do I incorporate technologies like language again this is from like engineering these are traditionally engineering things but data scientists are very much required now especially

1:02:09going into the future to understand how to use these engineering things so you can't get away with oh like I'm a data scientist I'm going to make my own models and just like not just work in isolation it doesn't work like that anymore especially if you're working at a medium size to smaller company this is where a lot of your job is going to be like full stack there's going to be things that are coming up where you're expected to deal with engineering problems so if anybody's out there are interested in a data domain and if they the biggest question is like how do I change the trajectory of when I'm learning for the future that's what I recommend learn about these technologies but more importantly learn about these fundamental traditionally engineering things because that is where a lot of these companies are headed now that was yeah that's kind of how I see things so with that being said so I'm I like to say you're okay staying for a little bit longer great oh I got I got three more hours now I know where hours okay in that case let's do a little bit more of Q&A so let's do

1:03:15some Q&A there are some questions that were popping up did you know that Tina in Spanish means bucket no I I didn't I didn't know that that's valuable that's inside right there also I know I talk really loud so I'm gonna turn my game down and move I have this thing where I just talk really loud so I'm gonna try to require myself no I didn't know that thank you for letting me know that okay so um let's see actually let's start with some of the questions from Instagram and because I did do a poll on there and then also on LinkedIn and YouTube there are some great questions sit and you pull that up the analytics some questions for Alex the analyst he's better begin only a 15 minutes he's got to be high quality questions okay but he's never a little bit longer you want to see like 30 minutes max I'm fine I can keep going for three hours I know um okay some of these how can I ask his favorite moments of the SAS AI conference

1:04:22SAS AI conference was actually really good and there was one thing that I I almost added it to this because I really loved it they were talking about um oh what's it called synthetic data now I loved this because they used healthcare as an as an example and it was literally the exact data that I used to work with which was kidney disease so I used to work really a lot with kidney disease um heart attacks on college you which is cancer hematologist's blood so I used to work with a lot of that do a lot of um studies with that they literally talked right to me they were they were talking to me they're like what something that you can do with synthetic data is you can say hey AI create this data for this example that hasn't happened yet this use case that is never happened but we want to know what our data would look like and what would happen if it looks like that AI generates that data you can test your systems on that data and see how your product how your outcomes how everything is going to react to that fake data I that to me almost blew me away

1:05:24I was like oh my gosh I've never thought about that before but using it on enterprise level like SAS I was like that's a really great use case that one that was probably the biggest thing that I was kind of blown away with with AI in SAS specifically but I know other companies are going to start doing that because I that is like that's super super super super great use case for AI really really love that that was my my favorite thing that's super interesting huh yeah it's not just an help here I can imagine that for so many like imagine financial models you're using financial models you know like let's predict what's going to happen if these things defaults or these loans defaults or these things happen let's run these you what if scenarios on this fake data and in imagine now we can try to mitigate that because we're starting to see this trend or for downward let's see what happens if it continues to go downward then you can kind of predict for that and mitigate for those financial risks it was really fascinating in terms of like the

1:06:27quality of the data itself I'm sure you can we can talk about this forever I was safeguards that they have yeah so they gave like one demonstration and again it wasn't crazy good what they all they would do is the demonstration itself was good I just didn't get to see like I couldn't see the data you can't see anything they just talked about it so the demonstration itself was really cool but I can't validate any of it essentially what they did was they said okay hey AI that's integrated this I want to create some data to see what if you know this enzyme continues to increase with this patient and then it produced that and then they could run it through their systems and it could predict you know mortality rates or kidney disease failure or all these things that was like the example now in terms of what it was generating they talked a lot one of their main folks is was on the reliability and accuracy of the data that's what they were really focused on that whole time they talked a ton about it which was great because they were like we know it's not

1:07:29super trustworthy right now AI systems out there we are trying to create the most trustworthy systems we possibly can so when you ask for something it gives it to you and you and it shows you what we're doing on the back end so you can validate that it's accurate so they're putting all these processes in place to make sure that the AI whatever AI you're using you can validate what it's giving you and if it's wrong you can correct it and so I gave an example about so to the accuracy of the synthetic data I can't I don't know but it was really just the idea of it which I can totally see it being used in so many use cases that it and that alone is just a great idea for AI and generating data that was awesome yeah I think that is really really powerful use case it's actually kind of similar to like when people are creating models and they don't have enough data especially in healthcare where they're generating a lot of like data based upon data they do have to train models so this is actually good I know which thank you for this question so it actually

1:08:31feeds into this how about the data regulation that's GDPR how will companies handle that part so to the data regulations I actually really want to touch on this point because remember how we were talking about earlier about building the infrastructure around using these AI models it's like not necessarily I go build my AI model and like launch it and do things it's like how do I build infrastructure around it and there's another question earlier that I address this too it's like why are companies afraid right why are companies afraid like that's a big and I touched upon this tangentially they're afraid because you can essentially have a model come in and be hallucinating and tell you like incredible things that you think are real but if you're putting this into reports these things are going to be wrong and think about how scary it is if you have like one wrong thing everybody goes like wow like this is a very interesting right thing and then it starts propagating throughout and that's why having like information that's incorrect can be really really detrimental and one of the primary reasons why a lot of companies are just like straight out batting it until they can figure out. It's a tough problem because there's no ethical things that can be safeguarded

1:09:32for specific for specific use cases like chat to BT I'm like you know open AI all these different companies large language models they're trying their best to avoid these hallucinations but even not just hallucinations that can be an issue like for example um wasn't only like this we were building something for a company that wanted to have like an AI chatbot right it was like essentially to give walkthroughs to give hints these kind of things how do you get this chatbot to not stray off topic like these are very specific things that you need to do like how do you make sure that it doesn't just stray off topic and have a conversation that's inappropriate for the platform that you're using how do you make sure that like when it is communicating like everything it's like what is it if you just if you just ask it questions um that are outside of the scope of what it usually answers what is its behavior these are things that need to be tested empirically as well a lot of the data scientist's job is to run these experiments as well so you're not only just like trying to fiddle around with it like through uh purely from like you know fiddling around with

1:10:38what's happening there's also a lot of application in industry where you're testing out different models like seeing making tweaks to it making tweaks to the prompt making tweaks to the infrastructure of how it is that AI is being incorporated and seeing um is this good like if I change this is this good so a lot of these traditional data science things are still there they're just being applied in a different way and you can see large language models as a new uh type of experimentation that's happening i hope that made sense it's a bit of a rambulant yeah that also touches a little uh just the ethics and adoption that also touches a little bit on why i think and i didn't talk about this although i've talked about this on a lot of other uh my livestreams which is companies are companies specifically were banning chat gvt because chat gpt was training on that data they knew that and they're like wait we can't get have our data going out there now chat gvt or open a i is releasing an enterprise level but companies that people trust

1:11:39like Microsoft and AWS they're going to have these things in the future and they may use that or they may wait to see how feedback is but that that inhibits adoption because of that inherent distrust where they're like our data is leaking out there trade secrets personal data hipa compliance all these things are bit need to be taken to account and it's very serious especially in like you know the united states in europe and canada like you can go the companies can go bankrupt very quickly by leaking data selling data that they shouldn't be out there um they're all there all sorts of regulations and compliance things um and they can get i mean my company we used to work we had a data leak and it cost us like five hundred million dollars um it was it was like crazy high that's what happens and that's a lot of the concerns around AI is around data leaks personal data that should not be out there compliance with regulations and then in places like financial institutions banks places like health care it is regulated to death so you have a lot

1:12:43of hurdles that you need to jump before you start using anything AI and i think you know there was a healthy fear especially with chat gbt i don't think there's going to be as big of fear in the future especially when when large enterprises like microsoft and amazon um start releasing more secure systems um but that also inhibits adoption right so instead of starting now they may have to wait a year or two to get all do all their testing make sure everything's right then start building infrastructure that just delays adoption for some companies um especially that are really heavily regulated thank you so much Alex and on the chat as well a lot of agreement that side there's just a lot of things that essentially people are figuring out and you view or someone who's helping figuring out these things that can be a really good opportunity for you um i also did want to mention like one of my favorites um i just think there's a lot of like amazing free resources resources out there um that are just i'm just like blown away by the quality so deep learning.AI

1:13:46is one of my absolute like favorite go to who's i recommend that to pretty much everyone if you want to get started um learning about AI and then learning how to incorporate it and build projects office so my phone it's like by far my favorite course from them because it's like simple and it's so powerful is this course over here if i just link it in the chat so it's a chat gbt prompt engineering for developers and it's only an hour long so not only does it talk about the process of thinking about prompt engineering it's also it also talks a lot about like how do you um actually approach it by building a product so a really big part i think is people who are just using like interfaces like chat gbt and you gotta like think a little more than that if you're actually building something how do i like use it from a developer's perspective i think that's a very very important skill to think about i think they also do have yeah they do also do have a really good specialization as well in d.a.s stuff so really recommend checking that checking out those courses engineering you know there's no debate about his expertise there um let's see so early adopters were

1:14:48win well as well gonna be available to roach yes so the webinar is going to be available you can watch the playback as well afterwards so oh yeah i also wanted to mention if you want to have like some notes and also like this deck if somebody mentioned that they're interested in this deck so i do there is a email newsletter um that you can sign up for in which we will just uh email you like notes specific things for the lunch and learns and also there's a calendar that you can add in for um the future of the interest so the lunch and learns are a weekly thing they're completely free they will always be free and we invite people like Alex for example last week we had a workshop that was from a company of x-meta AI uh engineering people so it's like it really depends on um like what it is like who it is that they're talking about but it always like centers around the topic of AI and data and sometimes it workshops and times of the talk so they're always going to be weekly let me just link that if you want to sign up for it there you go um thank you for the kind of reason

1:15:53currently looking to get the quest three ones it comes out oh we're talking about fun fact i used to work an Oculus back in the day um yeah i just posted about that on LinkedIn in twitter this morning sorry who did i did oh you did well let's flex Friedman who's like the best podcast in the world i listened to all everything i want to put out but he did one with mark his third one with mark and they were did it completely in the metaverse but it was now almost not it wasn't like photorealistic but it was very it's really close it's come a long long way and i just talked about in the thing i was like i called that like years ago i said they're gonna be they're gonna get there and when they get there which is like soon like they're really gonna get there soon when they get there it's gonna be like it's gonna be so awesome and so yeah i think uh i think the art AR is a super super awesome uh field it's really really interesting it is i think especially

1:16:53AR is where it's gonna be coming from without going on a huge rant about that like working on the products that were there it was like a huge so unfortunately a lot of layoffs that were happening were in that area because it was essentially just like the meta like sinking in so much money like pouring into it we used to call like there's like an internal joke that was like um it was like daddy vuk who was funneling everything all the money from the ad revenue into like the AR and VR sphere yeah like i think the quest um i never really found like that much uh in terms of in itself the market fit wasn't really there but i think what we can really appreciate is with these companies like they understand that this product is not gonna be like widely adopted but everything is built on each other like for example like working with like AR that was built on the VR stuff it was also built on instagram so a lot of this like things products that are being built up it's progressing towards some sort of future um in which all of these things are going to be incorporated like the glasses the AR glasses that came out that was like a really really big stuff um

1:17:58that was from that on so i think yeah with AR VR if i had to bet out the betting on the AR side i think VR is gonna be hard to get out of the specific issues cases but i was in a couple of years could all be wearing air glasses wouldn't be surprised about that in my case i would prefer AR AR contacts so maybe that i just i just saw a video about that the other day which looks really cool air contacts yeah really okay can you say me that way there was like a demonstration of like a very early version of it that was actually working wait so me this is so excited i'm so down for that i'll let me think i'll go find it um how rigorous is the lonely activist selection process so we do select it's okay so i wanted to make a point this lonely activist so it's pinned over there pinned on the comments it is for learning AI and data skills to apply to real freelance projects so first what i want to make is that we are rigorous not in the sense that you have to learn

1:18:58you have to know how to code or have to have like all these technological background we will teach you we'll meet you where you are with your custom study plans what we care about more than anything is your ability to learn and your ability to actually like take it to come to be part of a community like is it a good fit for you to have community-based learning and project-based learning and even if you don't know how to code we will literally start with introduction of Python um and we will like go off there and the freelance projects as well so we ask people to work on learning projects apply these skills and then after you get to a certain point you can sign up for a freelance project where you're working with a actual company to create a product for them which is generally from our past cohorts uh is what people were really excited about so yeah that's how the screening works um not in terms of your technical skill sets we don't really care that much about that but your ability to learn and whether it's a good fit or not in terms of the community-based learning and accountability systems thank you for that question um

1:19:58i would like a connector directly in my brain neural link it'll be there you will yeah i don't doubt that i think of this yeah i don't i don't i don't doubt you almost don't doubt he just like imagine like many years ago he's just spewing crazy shit right he's just spewing crazy shit about cars electric cars everybody's just like you're insane somehow it happened um let's rules i agree so quick question if you have avi skills how do you incorporate domain think of it like the reverse of the usual pattern interesting do you want to take that Alex say that again because i i didn't okay if you have AI skills how do you incorporate domain think of it like the reverse of the usual pattern because like i think a lot of people who transition to data analytics are coming from other fields i think is what he's talking about and then if you're starting off with like AI skills how do you incorporate domain so i'm not sure about analytics i have an answer for engineering and data science though you go ahead because i i know what i'll say afterward okay one day was going to be the same

1:21:03well i think the first thing like if you have AI skills what's really cool about this skill set is that if you're like indecisive as a person and you just get bored easily it's a great skill set to have because you can bounce around everywhere that you want like different companies that you're working for because the domain situation um are you i would argue that it is easier to learn than the technical skill set like even if you have to be partnered up with someone who has like deeper domain knowledge um about that so learning the easiest way easiest way to learn about this is go get a job like seriously say you have AI skills which is a very highly desirable skill set go get a job in like a tech company the domain will come um ask questions a lot try to understand like all the jargon that's there and then make sure that you're talking to people with that domain specialty um and then say you're joining like a healthier company you're like what is a protein right i've seen this happen like before you're coming with like technical skills just like be willing to learn about what the domain is and just make sure that you're always talking and partnering up with people who do have domain knowledge so that would be my answer and i'm curious about yours

1:22:07yeah somewhat similar i i learned a lot of domain just from working in the field even before i got into so i worked um at like an ER and then i worked at um you know in therapy and so i worked on the like actual provider side where i was providing services so i learned about that and then when i started learning the skills i was like oh see i know i know the domain already now you're kind of doing the opposite you know the skills let's say you know the skills we don't have the domain knowledge um i think if you were wanting to utilize those skills you just have to go find a job somewhere hopefully in a field they interested in um and i'll give you an example this always this one always makes me laugh um there's one i i used to mentor people are trying to break in analytics and he just worked in a warehouse but he knew he had a lot of technical skills but he just used to work in a warehouse and he was able to land a job his second job that he got he was working at paypal and he was working in um like uh when people would get refunded money and he's like i'm learning more

1:23:12about this process than i ever wanted to he's like i actually don't he's like this isn't what i want to do with the rest of my life because i like i know way too much about it so he had the technical skills and he just trying to he started that at a small company just learning some stuff about payments then he got that job at paypal and he has done really really well but he just used those technical skills to land that first job kind of at a place that would hire him gain some domain knowledge and use that to transfer up to a bigger job and so if you already have a little bit of domain knowledge and you're interested in it try to go into that's what i did with healthcare as i kind of the opposite i have healthcare experience then i went into data but you can do the opposite you can be in data and then learn the healthcare cool um let's see well oh from mebrahim what's the stack use for adolescent builder way issues of your face building made in lessons learned that's a great question i'm gonna uh i've been working with a fantastic team really smart people

1:24:17it's we built everything from scratch so at oh well okay let me not everything like we're integrating stripe for our payments processing we're not doing that vice from scratch but we didn't integrate any course platform we didn't integrate any editor platform we built everything all of that from scratch um that was really interesting super super tough um especially getting it's called a remote code execution engine so having being able to write code in your web interface is not is not super simple you have to we also have to build in a checker make sure your answer is correct make sure you've got the right output so it's a called a remote code execution engine that is uh that was that the toughest part of the entire thing really really complex um everything is built our front-ends built with react our back then uh we have um we're gonna be using AWS and that's that's literally the last thing that we're doing is we're transferring our staging and dev environment over to our production environment our staging and dev we were using um oh some cheap

1:25:23thing to host it so it costs us like 20 bucks a month but now we're going over to AWS which when people start using it um it's going to be you know very expensive and so we're transferring over to our production environment so that's going to take a while so we use AWS for that we're also using uh versatile um and then we're using because we have to store um videos so we're using vimeo for that right now but I think we're gonna we're gonna create our own video hosting through AWS to bring down the cost even more um because vimeo is expensive um and especially on the enterprise level which is what the tier we're at because I have like 500 videos on there right now it's a lot it's double the videos I have on my youtube channel there's so much content on on the platform and I have more on the way like I'm still recording so that's kind of like our general text stack but then in the in the background we also are hitting we're using um Kubernetes Docker um we're also using um you know we have a python we would be able to write code and python my SQL post-grace SQL microsoft SQL servers we have back end databases that are in virtual machines that are running

1:26:25those as well so um yeah I've learned a ton about web development I learned a ton of about APIs like you were talking about earlier APIs are super important we have an API for everything everything's built up APIs um so I'm in like APIs are like my our run butter on on our platform so that's just some like general overview the APIs again right API I'm making over they totally are like seriously you're gonna go learn something learn to arrest APIs like just learn about what that is and how to use it yeah make your life better um okay the let me see so I just wanted to want to check on LinkedIn as well so if there's any additional questions go still on youtube I need to have like a better system I'm also trying to like figure out the system as well so bear with me everybody um figure out what is like a good way of okay so some questions here will a data analyst need to be a data scientist the category I that's a very interesting question that's a good question I almost

1:27:28added that to the middle one which is like the near future um I almost added that because much like this one that you I can people still see this they're looking beyond yes okay so much like this bottom one uh domain analyst positions are gonna be because of tailored AI now we don't have to adapt a little bit um because data analytics is not going to stay stagnant like it has for the past you know 15 years I think it is going to change a little bit I don't think we'll actually move it closer to data science I actually think we're going to move a little bit closer to data engineering um just almost kind of what you were saying earlier uh in in it's it is true just because of knowing how the data flows knowing understand where the data comes from how to get it from one place to another that's a lot of uh data engineering stuff but also understanding schemas understanding how AI is going to be integrated with that I think data analysts will need to learn a little bit of data engineering skills which I think it I think kind of like what you said for data scientists are going to expect data engineering stuff I think that is especially at smaller

1:28:32companies that have the kind of like a one jack-of-all trades data analysts they hired they're going to want someone to do a lot of that stuff especially with AI you know some of those some of those data engineering concepts are going to kind of bleed into other areas I I really think that data engineering is um is going to be bigger than it ever has although it's already big I think data engineering is going to be extremely extremely popular in the future so much so that they're going to expect um you know data analyst data scientists uh to to know a little bit of it as well absolutely absolutely so um one more question from Philip um on LinkedIn so I have a quite few questions I'd love to hear opinion on but the one that interests me the most from my experts point of view what do you think of the idea of citizen developer ship process experts using no code like excel or power BI and where do you think AI will support it so no no no code platforms yeah process experts using no code like excel or power BI and where do you think AI will support it

1:29:37so I've used a note I've used some no code platforms in the past um they it really depends on your use case but let me tell you some of my experience um I used to use one for my job and we would always run into some use case where it just didn't work and the thing about these no code platforms though there are some and the luckily the one that we have there are some that then you can open it up and you can see the code behind it and then you can edit the code now true no code platforms where they do not let you do that um in my opinion are not going to be the future I don't think that's the way to go um because it really limits it let's say you have a really specific problem that you just no code it cannot solve you can't bring in somebody to consult and like fix that issue because you can't see the underlying code um so no code solutions can get you let's say even 95% of the way there for some uses as things get a lot more advanced I have found that no code is not that great but if you stick to kind of like whatever that because

1:30:40there's so many no code solutions out there whatever your use case is if if it works for like 80 90% of it that's great it can save you a lot of money a lot of time but then you're going to encounter something and you're like we cannot do it with our current thing and then you'll have to find another solution or whatever so have one that can open up you can open up the back end see how the code is being written and then change it and save it um that allows that flexibility to bring someone in for harder solutions but no code no code is not bad thing I am not against no code I just I've used it and I see limitations to it whereas when I've been coding I haven't seen limitations I've just see possibilities in the future enjoy that's what I see I feel like I'm a little more aggro than you are okay but okay I get I'm not coming from an analyst I'm gonna say like from there's like low code development from like uh like engineering or data science so I'm like no no because like I get it the appeal of doing these things but so much

1:31:44of your work like it's just a lot of it's like experimentation especially with like data science like how do you the idea of like even just black boxing experimentation that terrifies me because there's so many like specific use cases that you're just like hoping this thing is gonna work out and if there's anything about data is that it always surprises you like no I feel like no matter how much experience you have you're like I got this and you look at it you're like I didn't think about that it's always like that and I'm like okay fine if you're gonna go make like a website and you're like square space right fine I understand that the savior like actually building things that are more like customized um or like enterprise level stuff I just don't think no code is I'll use it right now like I don't think no code is good like I just think bit reasoning of like no code is gonna be big therefore I'm not gonna learn how to code is not gonna fly for at least that I agree that I agree no and it's something that a lot of people um I have a lot of questions I don't know if this has been asked by somebody but I'm gonna

1:32:47say it anyways a lot of people I'm asking me why do I even need to learn SQL why do I even need to learn Python why do I need to learn how to use whatever coding language you're talking about now or you know there's a hundred out there but why do I need to learn these if we have AI let me give you like a super small example and then I'll just talk about a little bit so on analyst builder I have you know you can check questions and practice stuff to questions I have questions that are considered very hard so I have easy medium hard and very hard I wanted to see if Chatchy B.T. could solve these problems and they're not crazy hard like when I was working this is like a medium difficulty when I was in like the real work it gets way more difficult and way more complex in the real world but even Chatchy B.T. was having a really tough like could not answer these questions and to me those are like medium questions like from a working professional now on an interview you won't get asked those in interviews but then expand that a little bit further in the real world you are going to run into a lot of things that

1:33:50an AI system is not going to give you the right output or it's going to give you an output that works but it's not the optimal way to do it and that actually is kind of scary because you're going to start people I've already seen this is not just like going to happen this is happening now where people implement implement systems or code into their current code base that is not optimal and ends up really biting them later on or they're like oh I didn't know because I just didn't know I just trust that they I knew what it was saying so that expertise especially when I've been using Python because I've used Python for quite a few things when I've used Python like oh it's telling me to do this which works but I know that it's not the best way to do it just because of my expertise I've been doing this for many years and so knowing the skill is going to be important for a long time I like I was saying in this looking beyond I said maybe even 20 years maybe you don't need to know that much it's not as important you'll still need to know it you still I don't think I don't see it going away where you're like you don't need to know code anymore everyone can code I don't think that's going to be true I I just don't especially when you're trying to implement things in development

1:34:55in development environments it's a lot harder than it looks to just like put code it you can put code somewhere but then actually implementing it is a lot tougher than it looks so all these factors I think coding is like here to stay for a long time and knowing how to code this year to stay 100% agree and from the chat as well following up people are asking um see what is very fast changing demand for skills that require do you think certificates are way forward to prepare a job market where it's pursuing degrees still better so I did answer this in the chat already but I kind of wanted to repeat that so um I don't actually think it's that much of a fast changing demand to be honest I think there's like a transition at least from data science and um engineering stuff to learn how to to learn to use more engineering things but I think again it's like what I said earlier going back to the fundamentals right if you know how to learn the fundamentals and you understand how things work these new skillsets that are coming out these new technologies it's not actually

1:35:59particularly like absolutely new right it's kind of like if you know how coding languages work and you like went through the pain of coding in assembly or C learning a new coding language is so easy because you understand how it works just like if you know how to use REST APIs just another API to learn is so easy so these like fundamentals are always going to be coming to play statistics for example right if you know statistics maybe there's a new model that's coming out maybe it's like a new thing that's coming out it's just statistics it's those statistics it's just like a different flavor different application different iteration of it and then for these certificates I don't personally don't think certificates are particularly useful anyway even right now um in terms of a certificate in itself because you can get like that certificate it doesn't really hold that much in terms that of employer because it's like okay great like you say that you know these skills you said you did the certificate and it says that you passed it but it doesn't really like hold that much value it's more so the projects so I would say certificates just learn things from certificates it's also why I recommend like we're going to go Coursera is amazing it's free just audit it um and then actually apply them to different projects

1:37:03especially if you can get like freelance projects where like pro bono consulting based projects those are the things that employers are going to be looking at like wow this person actually knows how to do this as opposed to if I just write on my thing like oh like I know Lua and I'm like do I really know it who knows but if I have a project in which I coded it they're like ah like this person actually knows it and um in terms of a pursuing degree I really think traditional education to catch up to the AI stuff at least is going to take forever like forever I think just now like after like 10 years they're finally like huh we should incorporate data science as a as a degree um without being said though if you do want to go for a degree again for like data science um and engineering data engineering stuff like that go for like a computer science or computer engineering degree um those would teach you a lot of the fundamentals if you're going for like more research stuff go for like math and stats if you I think double majoring like statistics and um computer science is like I think that is like a really really good combination

1:38:04what do you think about analytic side Alex yeah I'm I'm I'm very biased in this area uh I don't think I use my degree in the slightest uh I have the green recreational therapy I knowing and and it all it's it's I'm of course coming from it's like very biased I've already been in it else but knowing what I know now about how to break into the and that's what I teach on my channel that's why I made it knowing what I know now I'm like I could have done this right out of high school I know I could I could have learned these skills taught my skills stuff all these skills I could have broken in almost a hundred percent sure now when you're at that place when you're trying to do it it seems impossible but when you're in it then you're like oh this whole I now see behind the curtains and you're like that's it's a lot uh it's not as hard as it looks be saying that I still think for a lot of people um it they may not be a self-motivated or they may not have the um ability to self-learn or know the path to go down and so for some people something like

1:39:10a data analyst bootcamp might be right for them not a lot of people though I don't recommend them but some people for some people a bachelor's or a master's or a PhD that's for them um in general though I don't really see going beyond a bachelor's for data analytics for most people um so if you're thinking about going and getting masters you can it's just a lot of money and you already should have some type of educational level some experience before you go for a master's um because I think a master's is very tailored for specific use and if you're just going straight from a bachelor's to a master's you don't have any experience um so it's not super useful at least in my opinion certifications now for data analytics there's almost none maybe the like tablo data analyst certification or AWS or Azure data analyst certifications are okay to get but again just like you said they're not they're not that like credible what's up whoa this guy has the certification we got to hire this guy it's this doesn't happen you know they the how you talk about it showing your knowledge in the actual

1:40:14interview that's what's gonna you know really really sell you in an interview and get you a job cool all right thank you so much Alex we've won over by 42 minutes I appreciate you staying out as well thank we got through a more similar questions that we're seeing as if you have any like additional questions about say like Alex like best place to reach your public LinkedIn right I feel like you answered LinkedIn yeah LinkedIn is probably the one I respond to the most yeah and then if they want to do learn more about analyst builder and everything there so sign up for the wait list that's where you're gonna be putting on the information right yes yeah and all I post stuff online too so I'll just follow me on set like every social media platform just look for me I'll be there I'll say analyst yes yeah only octopus as well if you're interested in any of things related to that you can email me from tina lonely octopus.com probably that's by far

1:41:15the easiest way to get to communicate so cool all right thank you again so much Alex this was has been very exciting thank you it has been shared by fancy people that we did not know about apparently I don't I'll have to look this guy up yeah I mean I uh he was like he's digging into it so I'm like I'm very curious I because people are saying that they were coming I'm like is it sponsored by Jeff Dean I'm like not that I'm aware of yeah I don't think so cool all right thank you everyone for joining and I'll see you guys in the next uh like next weeks lunch and lunch okay bye