HH66-28-01-2022_mixdown.mp3 Harpreet: [00:00:09] What's up, everybody, welcome, welcome to the artist. Data sighs. Happy hour. It is Friday, January 28th, 2020, to hopefully you guys got a chance to tune in to the episode that was released today with a friend from Down Under, Louis Marini. We talked about talk about Data man. We talked about Data, you know, the foundation discovering Data. It was a great conversation. Check it out released and Harpreet: [00:00:32] It was great. So there's also the comet Harpreet: [00:00:34] Office outage that happened on Wednesday. We talked, we did a panel discussion with the CEO of Super Conductive, the company behind great expectations as Abe Gong had Jimmy Whitacre from pachyderm and also a good friend of the show. Matt Blosser was also there Harpreet: [00:00:50] Talking about Harpreet: [00:00:52] Connecting the dots between Data Harpreet: [00:00:54] Governance, Data Harpreet: [00:00:56] Versioning pipelines, Data expectations and things like that all about getting data prepared for machine learning. So let me say that that's on the comet YouTube. We got working sessions going on with that as well, so definitely check that out. Shout out to everybody in the building. Van Can Erik Serge Russel, what's going on, everyone? I'm excited to to to be here. Harpreet: [00:01:19] So does my Harpreet: [00:01:21] Audio seem off? All right, I can try to fix that. Let's see what's going on here. What's it sound like? Can somebody describe Harpreet: [00:01:29] It to me? Sounds a bit like Speaker3: [00:01:30] The game's Harpreet: [00:01:31] Very up. Oh, OK, Harpreet: [00:01:33] Maybe this helps a little bit. Yeah, that sounds cool. Yeah, sorry about that. I hope they didn't blow up you guys speakers. For some reason, I had my input level on the audio set all the way up to the highest possible spot. Should we get now a shout out to everybody tuning in on LinkedIn? What's going on if you guys got questions, comments whatever. Please do, let us know. Leave a message in the in the channel, the comment section. If you got [00:02:00] questions in the chat, please do let me know as well as feel free to drop your questions right there in the chat. Harpreet: [00:02:07] So what are we on the Harpreet: [00:02:08] Kick off say? What are we going to talk about? Man, I had a question cued up. I forgot what it was, so it's take taking a second of killing time to remember. It's been a busy, busy week. Let me tell you that much, man I've Harpreet: [00:02:18] Done, I think Harpreet: [00:02:19] Three or four presentations this week. Two webinars, yeah, two webinars and the live session. So it's been been all over the place, man. If anybody got questions, please do. Let me know. Let's let's kick it off so well, we'll kick it off again. Ken, how are you doing, man? Speaker4: [00:02:33] Doing great and I've been working hard. I have some fun projects in the pipeline, so I am really excited to be able to to come clean about those on Monday. Harpreet: [00:02:44] Yeah, I'm excited for this big reveal. I'm excited for this big reveal, man. I excited to see what this is all about. Speaker4: [00:02:50] I have two big ones back to back so I can tease the first one. It's my first community project coming out Monday where I actually share some of my own Data and everyone is open to analyze it and and look at it with me. So a little bit different. Everyone's looking for projects. This is one that will have real utility on impacting hopefully my life. Some deep dark secrets might also come out, but Harpreet: [00:03:16] You know, so how can people join this? Is this just like something that's like the tune into a YouTube channel? You have all the details or how can people get, you know, be a part of this? Speaker4: [00:03:26] Yeah, the video will be out Monday. It will also be available on Kaggle. So I'm posting the Data. There will be the first Data set. I have a repost there. So yeah, just a little bit different. A lot of fun, and I'm excited to see what people come up with. Harpreet: [00:03:43] So they just putting the data out there and letting people kind of do what they do with it or are there guidelines or parameters or you just have to wait until Monday to find out? Speaker4: [00:03:51] Yeah. I mean, there are a couple things that I think would Harpreet: [00:03:53] Be really cool, but Speaker4: [00:03:54] To be honest, a lot of it is. I'm interested in seeing the creativity of the community and seeing what they [00:04:00] come up with, and obviously it will be giving away some prizes in those types of things for the the most interesting work, the most shared work and things along those lines. Harpreet: [00:04:09] Awesome. Well, that sounds Harpreet: [00:04:10] Like a lot of fun, and that's very brave of you to put to put all your data out there. I'm excited to see what data points you got to play around with. Actually, Eric and I, we connected a couple of days ago. We thought it'd be interesting to Harpreet: [00:04:23] Get Harpreet: [00:04:23] All of our friends, i.e. everyone here in this room and everyone watching to lend us their Spotify listening data. And maybe we could do something with that. We didn't flesh out the details of the project, but I think that could be cool like, Harpreet: [00:04:37] You know, music Harpreet: [00:04:38] Recommendation based on the artist and data science folks. Yeah, I'm excited for that, too. Speaker4: [00:04:42] That would be way more embarrassing than the Data that I'm sharing. Harpreet: [00:04:46] Yeah, we see some of the crazy shit I've Harpreet: [00:04:48] Been listening to BLEEP Harpreet: [00:04:50] Garbage Truck Song and Monster Truck Song. It's yeah, bleep man. That guy, that guy won the game for sure. Harpreet: [00:04:57] Right on, guys. Harpreet: [00:04:58] So if you guys got questions in the chat on LinkedIn, wherever you are, please do let us know. Definitely, we'll be taking questions. Speaker5: [00:05:05] I don't have question, but I have a follow up on last week. Yeah, please. Yeah, so last week I asked about, you know, how to test school or whatever without adding increasing to the collective misery of the planet. And so I I was thinking a lot about, you know, Ben and Dave. We're talking about about how could you just make like a hard question and then just have somebody just go forth and Google and figure it out tried to come up with a few different things that kind of build on built on one another, just like kind of a realistic task would. And then I asked a couple of other analysts just for feedback on it. I haven't actually had anybody do the test yet, but just for feedback looking at it, and I did get a couple of interesting responses of just good responses of like, wow, like, you know, yes, that's accurate. And that second question would be really challenging. And so I don't want to make it like too too hard because I only have a short [00:06:00] period of time, but I was glad to see like I felt like I was kind of striking a decent balance between, you know, Oh, I know how to do this and I'm going to have to google it because, you know, because that's just how work is. Harpreet: [00:06:11] So I want to come on it. But it's been been going Speaker5: [00:06:14] Well so far. I was glad to get some feedback. Harpreet: [00:06:17] Nice, man. Well, I'm glad that was helpful to you. I was gonna say, Antonio, you're quite quite good with SQL as well. Like if if you were to create like an assessment of someone's SQL skills, like Harpreet: [00:06:29] What are the Harpreet: [00:06:30] Two essential topics that you would structure your question Harpreet: [00:06:34] Around? So Eric, I Speaker5: [00:06:36] Was trying to follow, but is that for like who? Who's your audience or who's your target audience? So target audience would be people applying for senior analyst or above positions. I think I mean, that's very tough because I'm in the boat of, if you know, how to join and like Group II and like do some basic stuff, then you can pick up other things. Depends how heavy it is. Harpreet: [00:07:01] So I might Speaker5: [00:07:02] Not be the best because because I'm in the boat where I mean, honestly, for Google, they ask me a sequel question, right? And before that, I heard about technical assessment and I went on an interview query and I practiced, like advanced the medium stuff and I go there, and he asked me something about Harpreet: [00:07:21] A like, How do you Speaker5: [00:07:22] Transpose the Data thing with like cross supply or something? It was literally like a three line SQL code, and I totally blanked out during my interview and I was like, I was like, I have no idea how to do this. And so I was like, All right, this is going down the drain. And as soon as I ended up in my interview, I googled it for a second and I ended up emailing the same person I found him on like LinkedIn found his email emailed, and I'm like, Hey, I know I totally bombed the SQL interview, Harpreet: [00:07:49] But I actually know how to Speaker5: [00:07:51] Do this. I just was preparing for something totally so much more advanced than you asked me something Harpreet: [00:07:55] So simple, and I Speaker5: [00:07:56] Ended up sending him the solution, and I ended up passing the interview Harpreet: [00:07:59] Around. [00:08:00] And this is how I can Speaker5: [00:08:01] Get job still somehow. But I guess that's why I might not be the best person to ask because I am very against kind of like making advance like Harpreet: [00:08:11] Assessment because I've been Speaker5: [00:08:12] Doing SQL for a good amount of time. And when it came to like me getting assessed, I completely bombed it. Harpreet: [00:08:20] That's it. That's an interesting point. Like, I like how Harpreet: [00:08:22] You you Harpreet: [00:08:23] Sent a follow up email like, I mean, I'm guessing it was relatively shortly after and you're just like, Hey, look, I know a blanked out. But here's what the answer should have been. I'll be curious then like, what are your thoughts on on that? Like if a candidate was in an interview with you and they blinked, let's say, Harpreet: [00:08:40] On the Harpreet: [00:08:41] Coding assessment. But then like, let's say within within the half hour after that, Speaker5: [00:08:46] They were like probably 20 to 30 min the range where I would be like, just email them right away. Harpreet: [00:08:54] Yeah. Like, what are your thoughts on that? Do you think that that should help a candidate hurt a candidate? Do you think companies should be less lenient towards that? Speaker6: [00:09:02] To remind me so much of myself? And like how like how I would act that? Yeah, that'd be huge. Just because I think all of us are like that. You know, when you get to a certain level, Harpreet: [00:09:13] You Speaker6: [00:09:14] Get something wrong and it's just you Harpreet: [00:09:15] Can't let it go. Speaker6: [00:09:16] And that's something that I look for is, you know, it's it's almost like, I don't care if you give me the job. Harpreet: [00:09:22] I just don't want Speaker6: [00:09:24] To leave this thing this way. I just can't do that right? You know, whether it's reputation, whether it's, you know, I hate being wrong, whether it's like, I can't leave that question, you know, just hanging there. There's something there, you know what I mean? Harpreet: [00:09:39] It just resonates. Harpreet: [00:09:40] Ken, Ken has a little story here. Third person, he knows that did that. To us a little bit about that, by the way. Shout out to everybody watching on LinkedIn, on YouTube and in the room. If you guys have questions, let me know. Go ahead and add you to the queue. Speaker4: [00:09:52] Yeah. You know, there was someone I had on my podcast, Ray Harpreet: [00:09:56] Ojl, Speaker4: [00:09:56] O.j. I can never Harpreet: [00:09:58] Pronounce his name. Awesome guy. Speaker4: [00:09:59] But [00:10:00] he landed his first Data analyst role because he had an interview with the, you know, a technical interview and then interview with the CEO. Technical interview. You did not do well, but he did that exact same thing. Harpreet: [00:10:10] He's like, oh, you know, like, Speaker4: [00:10:12] I knew this. I like researched it or went in and followed up with a guy and worked perfectly. I mean, he landed the job. They really like that. That approach, I mean, just as Vin said, isn't that exactly what you want Harpreet: [00:10:25] In a candidate is that they Speaker4: [00:10:27] Want something to be done right? They want to go the extra mile to make sure that that the stakeholder or the end user is not given faulty information and they're willing to admit when they're wrong and go back and revise Harpreet: [00:10:41] Their own work. To me, it's really Speaker4: [00:10:42] Hard to to fault that person in any way and like, there's no possible way that that looks bad for them. Harpreet: [00:10:49] Russell's got a great comment here Speaker5: [00:10:50] That actually looks like it. Maybe you don't want to work for that person yet. Harpreet: [00:10:54] Russell, go ahead. Go ahead with the comment. And then we'll go to mark after that. Go ahead. Speaker3: [00:10:59] Sure. So it's just a very generic comment saying that Harpreet: [00:11:02] I think Speaker3: [00:11:03] A person's excuse me person's knowledge of the language of basic Harpreet: [00:11:08] Syntax and Speaker3: [00:11:11] Functions and which of the functions are optimal for which particular challenge in any form of analysis and then which of the basic forms can interplay with each other and which you can nest to get functions can be a very good indicator of someone's understanding of the language, even if they struggle under interview conditions to answer a specific problem. So if you kind of take the here's a problem, solve this from two hours or you won't progress out of the equation and just talk generally about some of the basic elements of the language and how the structures work. I think that can be a good indicator, but it's perhaps not directly applicable to your question, Eric, but a fancy supplement. If [00:12:00] there's if there can be a, you know, a, you know, a heads off Harpreet: [00:12:03] General check Speaker3: [00:12:04] Kind of section outside of specific challenges. Do you know what I mean? Speaker5: [00:12:09] Yeah, it's been like, that's a like a kind of a hard balance or whatever like you're saying, like, is it the general general stuff? Versus I also like trying to give somebody a look at Harpreet: [00:12:19] Like, this is Speaker5: [00:12:20] Like, this is what I do. This is the kind of the specifics of what I do is something you're interested in or not, you know, because they're interviewing me as much as I'm interviewing them, right? Harpreet: [00:12:32] Let's go to Marc. Yeah, I just Speaker5: [00:12:34] Want to share a story just to balance out what has been shared here. Unfortunately, I didn't get that like, Oh wow, thanks for coming back and giving this amazing detail. So I was interviewing for Facebook for a role and did the tactical screen. I did OK for the school thing, but typical Facebook ads like a very product focused question and I just completely blanked out and for context, like my background and research, Harpreet: [00:13:00] Design and experimental Speaker5: [00:13:01] Design. So it was real experimental design for product. So it was like my bread and butter and I just blanked out. And so somewhere I was like, Wow, how Harpreet: [00:13:09] That messed that up? Speaker5: [00:13:10] So over this research, how can I best answer that? This whole thing, the next morning, I was like, Let me just get a fresh mind. I was taking out the email and I sent to. I was like, super excited kind of similar advice. I didn't notice that five minutes before I sent the email, I received a rejection email from them and I sent this email saying like, Hey, like, here's what I would have done differently. And then I opened my inbox and saw that I was like, Please disregard this last email. So sometimes it Harpreet: [00:13:42] Doesn't work out, but Speaker5: [00:13:44] It was a good exercise where if I actually had to do that question again, I would totally crush it. Harpreet: [00:13:50] Yeah, that's Harpreet: [00:13:51] That. That's interesting for me because actually, here Harpreet: [00:13:53] In the chat Harpreet: [00:13:54] On LinkedIn surprise Google at that past, Dave Brown says, I hate interviews because [00:14:00] of how easy the questions are Harpreet: [00:14:01] Set. My most Harpreet: [00:14:02] Hated one is to write a Harpreet: [00:14:04] Join in SQL. While it should Harpreet: [00:14:06] Be explained why you would use a join. That's interesting. Also, Dave Brown, another comment is a two way street. Harpreet: [00:14:14] There were a few Harpreet: [00:14:14] Companies that I've interviewed Harpreet: [00:14:15] For Harpreet: [00:14:17] That I would love for a recognition of a mistake. While they Harpreet: [00:14:20] Are the ones providing Harpreet: [00:14:21] The job, it would be great indicator of someone who I don't work for. Yeah, dude, I've done that before. Like, I've definitely Harpreet: [00:14:26] Messed up a lot of interview Harpreet: [00:14:28] Questions and then Harpreet: [00:14:29] Just ten minutes Harpreet: [00:14:30] Afterwards were like, Oh yeah, by the way, this is actually what I meant to say and how I would have done Harpreet: [00:14:34] It. Harpreet: [00:14:35] Sometimes it works out and you move on. Sometimes you don't, but it's worth the shot. I totally agree with Van. It's just something that bothers me. It's like, I hate like, like, I know how to do this, and I'm just to prove it to you that I do know how. It's just your artificial Harpreet: [00:14:48] Constraints made me so I was, you know, unable to Harpreet: [00:14:51] Do it efficiently or go for it. And then just, oh Speaker5: [00:14:55] Yes, as a quick follow up. So I can't remember somebody here posted this on LinkedIn, maybe even more than once. Like, I'm not 100 percent like into the idea of. Tests, anyway, right, because like, really like literally the only time coding tests I ever took while interviewing, I failed and it was terrible and I felt just bad about myself, even though it was like one stupid thing, right? And so I'm not necessarily thinking, OK, this is going to be this like magical test that's going to be a useful thing for forever and ever. So like, if you weren't like, if you weren't going to have a coding test and we were just going to ask, I'm just going to ask you like three questions in a conversation that we're going to have over the next 25 minutes. And I want to understand if you understand Harpreet: [00:15:44] Stuff without Speaker5: [00:15:46] A select statement involved. What would you like? What would you ask? Harpreet: [00:15:53] Mark, go for it. Speaker5: [00:15:54] I think a good question. We're actually just talking about within my within, my job on our team is [00:16:00] when's the proper time to use a Harpreet: [00:16:01] Cte as compared to a Speaker5: [00:16:03] Sub query? They're both very similar. They basically do the same Harpreet: [00:16:08] Thing, but it starts Speaker5: [00:16:09] Going into like, how do I write my code and how do I think about readability and how others can interpret that? And so like Harpreet: [00:16:15] The kind of Speaker5: [00:16:17] Logic that we came about with is like, you know, we could be wrong. There might be some different best practices. But you know, if it's an individual piece of logic, I want you to use a CTE to kind of bring that out. But if it's a sub query that's like tightly, I need it for a CTE, like some subset within the CTE. Like I might include that term for Harpreet: [00:16:38] Like, Hey, this is like a cup of logic. The main Speaker5: [00:16:42] Thing is just being able to explain my reasoning of like, why do I structure my SQL code this way with the goal of like being able to explain to others or let Harpreet: [00:16:51] Them have like an idea of like Speaker5: [00:16:53] What's happening? And for me that that tells me like what they actually care about their code and to they care about communicating to others at versus sub queries seem to be like, I mean, it's going to bring out everybody's SQL philosopher, I think Harpreet: [00:17:10] Search any, any input. Speaker7: [00:17:13] Oh, I was just typing something. I think I kind of find that my experience would push things off track a little bit. But yeah, recently I had a recruiting experience. I'm not actively looking for a job, but CEO of a company reached out to me and the job they were offering was right up my alley. You know, for those that know, I wrote a book on interpretable machine learning. So like, the job involved that pretty much that I'm facing customers, which is something I've done throughout my 20+ year career. Harpreet: [00:17:52] So like, Speaker7: [00:17:53] He knew that off the bat and they put me down the same rabbit hole, they put all the other recruiters. The [00:18:00] recruits are applicants, if you will. Harpreet: [00:18:03] And it Speaker7: [00:18:04] It was annoying because, you Harpreet: [00:18:07] Know, like from off Speaker7: [00:18:08] The bat, they, you know, the CEO told me one process. Then the hiring manager told me a different process. Then they put me in contact with the recruiter and the recruiter set up a meeting with me. And it turned out that it was not a meeting with the recruiter, but with a, you Harpreet: [00:18:28] Know, Data scientist. Speaker7: [00:18:30] And it was a technical interview, which I iust, but it was completely unexpected. And then after that interview was supposed to come apart that I knew was part of the whole process, which was to take a paper on interpretable machine learning, you know, academic paper and and create a presentation on it. I knew I was going to do that and that was it. But apparently there was that and even more technical interviews after that. And I said, you know, like, I didn't want to be a jerk about it, but you know, like, I already have a job that I like. I mean what? There's no guarantee I'm going to be wasting my time for hours here. You know, I'm OK with creating a presentation because that's something that could be useful for me later on. But you know, a coding exam is not useful to me. And you know, there's like maybe like less than a 10 percent chance I get the job if I'm being, you know, maybe pessimistic or optimistic, I don't know, but I thought, you know, it's it's there's a chance I'm going to be wasting my time for a few hours. And clearly they don't value it, you know? And the CEO contacted me, so I don't know that should put me in a different position than other people. Or am I being too fussy or or a jerk? Harpreet: [00:19:46] Let's go to Ben on this one. And after Ben, I can't. I think you had a comment in here about case studies might be appropriate or not, but then go for it. Speaker6: [00:19:54] I think at a certain point, I know I've said this before, but at a certain point like you don't fit in the interview process Harpreet: [00:19:59] If you [00:20:00] want to Speaker6: [00:20:00] Hire super senior people, you know, when I call them the senior plus plus, because you get past like their senior year 10, 10 to 15 years, but you break that 15 to 20 year horizon like putting somebody like that unless they're transitioning into a new career or a new role, putting somebody like that through an enter the normal interview process, it's like, how would you even why would you do that? What's wrong with you, you know? And so when I because there's people that I have hired just because they said they were interested in a job and I'm like, OK, let's you know, let me get you an offer letter. Why am I interviewing somebody with 15 pages worth of Google search results when it comes to Harpreet: [00:20:40] Research and body of Speaker6: [00:20:41] Work? It's like, what? I'm not going to, who do I know that's qualified to interview you? And so I get you can hear me getting flustered because I don't want to say that the the truth. That is coming to my mind, but some companies really have to look at hiring very senior people if they do hiring C-suite roles because it's exactly the same process you don't hire your CEO doesn't show up for a job interview like a VP does. I mean, that's a different vetting process. It's an entirely different workflow for recruiting, for interviewing, for assessing. It's one hundred percent different. So why would you then have somebody who is just as experienced from a technology standpoint going through something different? It doesn't make sense. Companies have to figure out how to create this second process, and really successful companies have most of the largest tech firms have, you know, really three or four different tracks that you go down Harpreet: [00:21:39] And depending Speaker6: [00:21:40] Upon what type of role they're putting you in. There is no interview for some of them. I mean, some of these jobs are sales pitches. So and companies have to understand they need to build out tracks of interview process. And if they don't, they can't attract top talent. And I can tell you, you know, for certain that there are companies [00:22:00] who are like, Yeah, we only hire the top and it's like, I see you process. No, you're not. No one who is a top talent would go through this. And so you're kind of, you're lying to yourself, you're hiring the top people Harpreet: [00:22:10] That will Speaker6: [00:22:11] Put up with the pain of going through what you're putting them through. And so, you know, if I could, if I could say one thing to companies who are trying to hire people who are very senior, who are going to build organizations, build departments, build practices like ethically explainable AI, all of these things that companies desperately need and don't have is if you've got an expert, Harpreet: [00:22:32] Stop just off Speaker6: [00:22:33] From the job. If they've published credible work, if they have peer reviewed body of work, stop Harpreet: [00:22:40] Just off from the job and think, Harpreet: [00:22:42] Thank you so much. There's a lot of a lot of people vibing with it on LinkedIn. Harpreet: [00:22:46] A lot of +100 Harpreet: [00:22:47] From for being from Mexico, who is on LinkedIn today Harpreet: [00:22:51] And Harpreet: [00:22:52] As well as Trey. I mean, I heard about this story about the dude that invented homebrew like the thing that we all use on the terminal. Harpreet: [00:23:00] And he went to go interview at Google, and I guess Harpreet: [00:23:04] Because he couldn't invert a binary tree, they rejected his application and he had like this angry tweet about it. Harpreet: [00:23:11] I'll have to find it, and they use the Harpreet: [00:23:13] Exact words, but it's like, yeah, they invented something that pretty much the entire team uses and, you know, modernized computing in a way. Ken, what are your thoughts? By the way, if anybody has questions or comments, please do. Let me know. I hear in the chat as well as on LinkedIn shout out to everybody, not on camera. That's not blazers in the building genus and building Eric or Chris Murphy. Good to have all you guys here. Harpreet: [00:23:38] Yeah, I mean, Speaker4: [00:23:38] Just a real quick follow up with what Ben was saying. I think that like I am by no means a senior leader, but I also have designed my career in a way that I shortcut a lot of things. But if you're you're hiring senior leaders for the body of work that they've that they possess what they've done in the past, there is no way you're going to Harpreet: [00:23:56] Evaluate accurately Speaker4: [00:23:57] What they're capable of in any form [00:24:00] of assessment, even early Harpreet: [00:24:01] In an interview. Right. You're looking at it, the Speaker4: [00:24:03] Impact they've had in other companies. You look at the impact that they have in their own personal work, books, research, whatever it is. And that interview is almost exclusively to see if you can work with that person. It shouldn't be to assess anything Harpreet: [00:24:16] Because all of those Speaker4: [00:24:17] Things, if they're reasonably senior or out there in the public for the world to see or you heard about it or it's on the resume, you could at least ask them about that. But to me that the idea of some sort of assessment related to that is is ridiculous for Harpreet: [00:24:31] For Eric, Speaker4: [00:24:33] I'm a huge fan of case studies Harpreet: [00:24:35] Just talking Speaker4: [00:24:36] Through problems. I mean, it could even be something similar to what you use for a technical interview. But rather than having them coat it or rather than having them actually write sequel, just talk about the process that they would use and what they're thinking about when they're evaluating a problem like that. Or if you're like, How would you design this, like this? Or like, how would you design this process? Harpreet: [00:24:56] What would it look like Speaker4: [00:24:57] From end to Harpreet: [00:24:58] End based on Speaker4: [00:24:59] The specific Harpreet: [00:24:59] Problem? Speaker4: [00:25:00] I think the verbal communication of technical Harpreet: [00:25:03] Concepts, at least to me, Speaker4: [00:25:05] Is more important than if they can write the code because you can just Google the code and the syntax and those types of things. Harpreet: [00:25:11] So in the past when I've when I've Speaker4: [00:25:14] Done interviews Harpreet: [00:25:15] Like that as the Speaker4: [00:25:16] Interviewer, I've always put a huge premium on the applicability if they want to use a whiteboard to describe it, to show how things joined, to do those types of things. Harpreet: [00:25:25] Why not? Speaker4: [00:25:27] I consider that like semi technical. Harpreet: [00:25:29] But you know, that's Speaker4: [00:25:30] Also an opportunity for you to have a dialog and for you to dove kind of Harpreet: [00:25:34] Further and see if they Speaker4: [00:25:36] Truly understand the concept. You can see where they're strong very clearly. And also you can evaluate some of the weaknesses and pry into those just a little bit more that does get away Harpreet: [00:25:45] From like having a Speaker4: [00:25:46] One in zero like this person pass through didn't pass, obviously. But if you really want to know the candidate how. My conversation with them and using that time Harpreet: [00:25:55] To see Speaker4: [00:25:55] The thought process, at least to me, is like huge premium over a coding test in [00:26:00] most Harpreet: [00:26:00] Circumstances. Absolutely love that. I think one thing that I Harpreet: [00:26:04] Would love to see assessed in especially modern knowledge work type of jobs Harpreet: [00:26:08] Is gauging someone's Harpreet: [00:26:10] Ability to find an answer Harpreet: [00:26:13] Because that's Harpreet: [00:26:14] Really, really cutting, you know, kind of a differentiator in it. Knowledge work, just be able to find answers quickly, right? Harpreet: [00:26:21] So do you know the Harpreet: [00:26:22] Space well enough to come up with the appropriate search terms and you know how to search them and how to find answers that are credible and then, you know, put them to work? I don't know that there should be a Googling component to most to most interviews. Harpreet: [00:26:37] Maybe Google should Harpreet: [00:26:39] Should have a part in their interview where they interview you on the ability to use Google Harpreet: [00:26:45] Search. Go for it. No, no. I was just Speaker7: [00:26:47] Laughing. Harpreet: [00:26:48] I just said, Speaker7: [00:26:49] Yeah, I definitely think it's a it's a good skill. You can't Google if you're not proactive and you know, you can search for things you know in creative ways and find them, you know, like even if it's like, you know, you get a coding error, there's different strategies you could find to. It's not just going to stack overflow things kind of ways of using the keywords to your advantage to find the most, you know, the latest, you know, information about that. That is very important because it can increase your productivity and also shows into how the person thinks. Harpreet: [00:27:31] So I think it's definitely I think someone Speaker7: [00:27:34] Tweeted about that, that a person put among their skills Googling and and the person said in the tweet, Yeah, we're definitely interviewing this person because it's a skill that it's completely underrated. Harpreet: [00:27:49] Yeah, yeah. I mean, he could be sitting there for hours on end trying to find the maybe not the solution, but something that I'll get you on the right track. But without having the right combination of domain knowledge to [00:28:00] have the right search terms ready at hand to look up. Harpreet: [00:28:03] Plus the Harpreet: [00:28:03] Ability to effectively search for them like, you know, you could save hours and hours on end shout out Harpreet: [00:28:08] To a to Harpreet: [00:28:09] Great skill in the building. Also, see Antonio's baby that that that's a cute Harpreet: [00:28:14] Kid, man. Speaker5: [00:28:15] Anyone's hiring. He's looking for a job. How old is he now? Harpreet: [00:28:19] He's he's about three and a half months, and Speaker5: [00:28:22] He still hasn't paid me rent. So he better find a good salary because I'm going to start charging trust man. Harpreet: [00:28:30] It's a lot easier when they're blobs. All of a sudden they started moving around and it's like, Where are you going? Are you getting into everything? Please stop. Calm down. But if anybody has got a question or anything, please let me know. Harpreet: [00:28:45] Looking through the Harpreet: [00:28:47] Chat on LinkedIn, don't see anything there. Harpreet: [00:28:50] Don't see anything here either. So what's up, everybody? Somebody is asking something about DAX in Essos. Harpreet: [00:28:57] Dave Brown. How advanced are more DAX expressions? I don't think I know what DAX is DAX. I've been extremely difficult DAX expression, but I choose to try to program any of these values before entering a visualization Harpreet: [00:29:10] Program into a Harpreet: [00:29:12] Table to keep the visualization report fast or simple. Harpreet: [00:29:15] Dax expressions good enough. Harpreet: [00:29:18] Interviews are saying they use DAX for everything writing ten fifteen expressions per day. Are they talking about simple expressions? Curious if anybody has an answer to this question. Harpreet: [00:29:30] What's the question? Harpreet: [00:29:32] How advanced are most DAX expressions DAX? Hopefully I'm seeing that the right way. I don't know how else you'd say that. Speaker8: [00:29:39] Yeah, I think you can make it as complicated as you want, right? Depending on what you're searching for me, DAX is kind of like has a syntax Harpreet: [00:29:47] Of like conditions Speaker8: [00:29:49] That you put like, has a language of case win. Harpreet: [00:29:52] So case when Speaker8: [00:29:54] You have this variable, then Harpreet: [00:29:56] Aggregate these numbers Speaker8: [00:29:58] And put them in this weird [00:30:00] time window and things like that. So and this is definitely something Harpreet: [00:30:05] That you know is Speaker8: [00:30:06] Behind Power BI. So when you're Harpreet: [00:30:09] Kind of I think when you Speaker8: [00:30:10] Put that in the side, so it's more of a one understanding your Data, understanding what you're trying to pull and then practicing. For me, I played with a lot of I need a lot of, you know, errors by trial and error until I got it right watching videos. But you can get really complicated. But also it's a fun exercise to kind of test your creativity, too in terms of how you want to pull the data and things that are super interesting. And I love RB, that's by the way. So and you can even use docs in Excel through power query. So it's pretty cool. Harpreet: [00:30:49] Thank you very Harpreet: [00:30:49] Much, Greg. Yeah, I've never, never heard of Dash. I've. Merely not done like any business intelligence in my, you know, quantitative career seems to be Harpreet: [00:30:58] Something that I'd like to work Harpreet: [00:31:00] On and get better at. There was a a resource for DACs that Eric Sims had mentioned Calvin de Harpreet: [00:31:09] Wilde, Harpreet: [00:31:10] Calvin the Wild Wild Harpreet: [00:31:13] West on LinkedIn. Harpreet: [00:31:15] I've seen him come over my newsfeed, but I've never actually chatted with him. Harpreet: [00:31:19] Would love to. We'd love to Harpreet: [00:31:20] Chat with him. Gina, you've got a question. Go for it. Speaker5: [00:31:24] Oh, thanks. Harp. So sorry, I missed the first couple of these happy hours this year, but it's cool to see everyone again and I love the conversations. They're always so wide ranging. I was listening to a podcast recently, Super Data Science Podcast, talking to I believe it was Sadie St. Lawrence about trends in Twenty Twenty Two and from, you know, kind of career corner, job search corner. You know, one Harpreet: [00:31:54] Thing that I, Speaker5: [00:31:56] As a recent, relatively recent bootcamp [00:32:00] grad and Harpreet: [00:32:02] Job seeker Speaker5: [00:32:03] Kind of struggle with. I know we've talked about this a little on other happy hours, but the sheer the rate of change in this field and Data science, which is really intense. And also how, you know, job search or starting out might incorporate some of the trends people see for twenty twenty two. And so I'm not going to cite the trends they talked about. Rather, I'm interested to hear Harpreet: [00:32:29] Your guys's thoughts Speaker5: [00:32:31] On trends for twenty twenty two and those that are specifically relevant for, you know, for job searches, maybe those who are career changing or kind of early on in their data science careers Harpreet: [00:32:46] And out to to say these lures and John Chrome, both both friends of mine, hopefully guys are tuning in. But yeah, anybody got any insight here on trends. So it sort of trends on what candidates need to know in terms of tooling. I'm not sure if I caught that bit. Harpreet: [00:33:04] Yeah, I mean, Data is kind of big. Well, they even Speaker5: [00:33:07] Talked about some what they called micro trends and macro trends. So I think they talked about tools. Automl type stuff is one kind of area. And then they also in a way Harpreet: [00:33:20] That might be more what they considered Speaker5: [00:33:21] Micro or more kind of tactical, I guess. And then maybe macro trends in the sense of, you know, they talked about Gans and deepfakes and some of the the, Harpreet: [00:33:35] You know, the big kind of hype Speaker5: [00:33:37] Up and coming. It's already here to some extent, of course, Harpreet: [00:33:40] But Speaker5: [00:33:42] Things that will continue to change and those areas that can have really broad ramifications for our society. And so, yeah, again, I'm thinking about it from the perspective of I'd love to hear you guys thoughts on those trends anyway. And then if there's kind of a direct correlation to Harpreet: [00:34:00] Job [00:34:00] seekers who Speaker8: [00:34:02] Might be early on in their data science career, I'd love to hear that too. Harpreet: [00:34:05] Greg, go for it. Speaker8: [00:34:07] I'm not going to give you an answer in terms of like what tools since, you know, I'm probably the least technical guy here. But what I would like to say to my younger self or whoever wants to adventure is to follow the Data. Where is data being generated? First of all, talking about trends in twenty twenty two is already a thing of the past. What is happening, what is going to happen in twenty twenty three and beyond? And where is the data going? So what are the things that we're starting to talk about nowadays? When you think about metaverse, we need to think about cloud everywhere. What does that mean to me? Iot is going to get big and you will need to do something with that data you're collecting from all of these billions of devices that will collect from manufacturing sites everywhere you go. What can you do to generate value? So start thinking about how can you run tiny models on devices? How can you do things about, you know, real time inferences? So nowadays you hear ML around batch inferences a lot. So what about real time, right? Can you create a new market with that? You know, try to think about beyond Harpreet: [00:35:18] 2022 and Speaker8: [00:35:20] See where you can start, you know, keep getting your skills around to kind of excel in this space. You also have quantum computing automated vehicles. I mean, quantum computing. Can you imagine a world where, yes, you may have a couple of establishments or data centers that create the state of the art error correction, quantum computing that will distribute over the cloud? And then now you as a Data professional, you can leverage CPU. You for quantum processing unit versus CPU GPU. You can jump from from either one of them when you need to anticipate [00:36:00] those things and kind of position yourself in this case. And then you can work backwards Harpreet: [00:36:05] To figure out what kind of Speaker8: [00:36:06] Tools you Harpreet: [00:36:07] Need Speaker8: [00:36:08] To learn it. So is it going to be Python is going to be something else, Julia, whatever that secondary, in my opinion. Harpreet: [00:36:16] Eric, in response to Greg's Harpreet: [00:36:18] Comment of being Harpreet: [00:36:19] Quote unquote the least technical person here, Greg is also the guy who knows more about quantum computing than any 10 people I've ever met. And Serge says he's actually a humble quantum guru, Schrodinger's guru. Mark, go for it. Speaker5: [00:36:36] I guess another kind of going back to the original question of like, Hey, Data space is moving really fast. Like, what do I do? Especially when it's like new machine learning papers coming out every other day? One thing I really try to focus on is really, I think Ken was talking about this, like crafting the career you want. And so start thinking about like, what type of companies do you want to work for? Are they like Fortune five hundred or they like the Big Tech or the Harpreet: [00:37:01] Startups, right? Speaker5: [00:37:02] And then within that, like what type of industry that you want to be in? And then from those type of companies that profile like what's the typical Data maturity of these companies? Harpreet: [00:37:11] And from there, Speaker5: [00:37:12] You can really work backwards of like, well, then given this Data maturity, what technical stack be really good at? So, you know, I originally started in health, tech Harpreet: [00:37:21] And health Speaker5: [00:37:22] Care is really slow to integrate machine learning because there's a lot of regulation. So, you know, I didn't need any any more experimental design and like regression analysis. Right now, I'm an HR tech, but again, I'm still in startups. So these advanced machine learning, deep learning stuff's not really important. I've been shifting way more towards Data engineering because simply, we just you get the data into a state where we can do those cool things. And so that's where I've been focusing on. I've been really enjoying and a big reason why I've shifted more towards Data engineering is, well, I love startups and I plan on being startups. And every time I go to a startup, the Data quality or Data processes and pipelines are just a dumpster [00:38:00] fire. And so I enjoy fixing that. But say, for instance, I wanted to go work for like Google or Facebook or things like that, then I would totally work on more so like a B testing, like machine learning and like things about that stuff in production. And so again, just really trying to craft what industry, what space and then being very critical about that Data maturity. And that'll give you a leg up in interviews, Harpreet: [00:38:21] Because now when you do your Speaker5: [00:38:23] Research and you're going to be going to those interviews and saying like, Hey, you know, this is why I notice, you know what? What is your current processes? What do you think about this? And you can now be like strategic and thinking like, well, what can value? Can I add to move you forward in Harpreet: [00:38:37] Your Data maturity or give Speaker5: [00:38:38] Them these Data maturity constraints? What value can I add on the business while also taking into consideration like, Hey, we can't do ML, right? So, yeah, Data maturity is the main thing I will focus on. Harpreet: [00:38:51] Go for it. Speaker5: [00:38:51] I kind of piggyback a little bit off what Mark was saying. So like when we talk about where the industry is going to me, that's like that feels really overwhelming because we're the industry is going is so big and it's really led by, you know, probably just like a small handful of players. And but Harpreet: [00:39:08] Like where most of the Speaker5: [00:39:10] Companies that you might be applying for are going is probably like light years behind where the industry is going because some of them are, you know, sure, some of them are incorporating cutting edge models for things, and others are just kind of like starting to embrace using Data to make decisions. Period, you know, and so like the way that I like, keep that in mind is just trying to remind myself, like the things that I'm interested in, the things I'm good at and the things I'm interested in stretching to learn are needed somewhere. And it's just like, it's kind of a numbers game of like, you know, keeping keeping that Harpreet: [00:39:43] In mind and and not, Speaker5: [00:39:46] I guess, pressuring, pressuring myself to, like, be perfect because no company is either. You know, and so anyway, that was just a little a little piece that helps me remember that most companies aren't on the bleeding edge. And [00:40:00] so there's there's room out there for pretty much any talent set. Harpreet: [00:40:04] I'm curious, though I would love to get Vin and Mark's input on this, Harpreet: [00:40:08] Like Harpreet: [00:40:09] Talking about like Data maturity and companies, right? Like, obviously, there's a lot of companies who are, quote unquote legacy companies. You got manufacturing Harpreet: [00:40:16] Companies, maybe brick and mortar Harpreet: [00:40:18] Retail Harpreet: [00:40:19] Companies. But as these Harpreet: [00:40:21] Startups start popping up, for example, like, you know, a lot of SaaS solutions, Harpreet: [00:40:25] There are a lot of apps. Harpreet: [00:40:27] For example, these are companies that are digital native cloud native companies. Do you think even from the outset, even with having, you know, software people, highly technical software, people Harpreet: [00:40:40] Be the Harpreet: [00:40:40] Ones that are starting these companies that they're going to suffer from these Data maturity issues as well or like, I'm not sure like. And then if so, then how does somebody learn about Data maturity like? Somebody got cut in a boot camp, something that you can't be taught. You only have to learn the hard way. Harpreet: [00:40:57] I don't know Harpreet: [00:40:58] If those questions make sense, but Mexico is actually in the building. So Mexico, if you want to answer that, go for it. And I'd love to hear from. From Mark Harpreet: [00:41:05] And Harpreet: [00:41:06] Then on that. Speaker8: [00:41:07] Yeah, I think in general, like starting a startup is always like a really risky endeavor. Harpreet: [00:41:12] And I know, Speaker8: [00:41:13] Like for me at multiple Harpreet: [00:41:15] Startups Speaker8: [00:41:16] Or like the teams that I was on, like one of the hardest challenges was and one of the most expensive propositions was actually like getting Data. So there's like buying the Data and then there is getting the people of your platform to generate, said Data. Like all the computation and storage itself and like the processing Harpreet: [00:41:36] Layers, you know, like Speaker8: [00:41:37] It's if you got enough money, you can buy it. But like, I feel like the Data is the hardest part. Harpreet: [00:41:43] And then the Speaker8: [00:41:43] Other aspect Harpreet: [00:41:44] Is, you know, Speaker8: [00:41:46] A lot of startups, they kind of get away with with doing shady stuff because at the end they the year they're trying to get to product market fit. They're not necessarily concerned about like they're concerned about security, obviously, but it's like they're not concerned about [00:42:00] necessarily ethics or or bias or, for example, whether or not they're violating GDPR. Because SERPs can kind of get away with stuff like, well, they're so small. At some point they do have to Harpreet: [00:42:10] Worry about it, but there's Speaker8: [00:42:12] Like a really kind of long road before they do. But I do think like we're kind of seeing we're seeing like both specialization and also generalization. You're seeing specialization that's led by like the really, really big companies, which they have the infrastructure maturity and they're dealing with the unique challenges. But I also feel like you're seeing generalization in that as we kind of build really nice like abstractions on top of layers. So an example would be like the blockchain crypto Web3 world, right? Like part of the chaos there is that some of the lower layers of abstractions are still being built out. But once you see those abstractions come into play and people really start building on top of it. For example, like Web two frameworks, then you'll see stuff really kind of take off. But it's tricky when it comes to Data because you have Harpreet: [00:43:00] To get it, you have to store it, Speaker8: [00:43:01] You have to make sure you're using it. And more importantly, you have to generate that Data flywheel to keep it going. And to some degree, Data maturity might actually just be aspirational as opposed to like a real thing that a lot of people have. Harpreet: [00:43:14] Let's go to a mark. And then Ben, Speaker5: [00:43:17] I would actually Harpreet: [00:43:17] Argue that you can have Speaker5: [00:43:19] Small startup Harpreet: [00:43:20] Companies be extremely Speaker5: [00:43:22] Data mature because maybe they have like a ML based solution. They've maybe come out of some university and they they're like, have to be Data mature to kind of to enable that. Or you have startups where like, you know, they have a SAS offering. But Data isn't the focus, but Data is what drives it forward. And so, you know, they may have cloud, but like they don't have the proper like Data modern Data stack, right? But then you may have like these Fortune 500 companies, you know, Google Crazy Data, which they're on the bleeding edge Data maturity, right? But then you may have like Walmart, where and I don't know Walmart's infrastructure or anything like that, [00:44:00] but like because it's such a massive organization, you actually may have pockets of Data maturity where it's like one team is just like extremely Data mature, but they're siloed and another team is just like struggling. So I remember I had I was talking to a recruiter from Walmart for this health care aspect, and their whole description of their Data stack was just like atrocious. But I know, like they have like, they're like pushing these crazy models because they have an online aspect as well. So I think it's really just dependent on on just the I would say how, how, how can I put it, how much is Data like important for driving business value? And I think that will determine how Data mature they have to be and how much they have to prioritize it. Because if it's a blocker to becoming profitable or reaching their customers, it's going to be prioritized. But they can reach that avenue through focusing on other things and just get by with the bare minimum. They're totally going to do that, too, so it's going to be like a balancing play. You know, do I focus on Data infrastructure or like when he goes saying, you know, get into product maturity or product market fit, I'm very careful not to say he has greatly more experience than me and seeing a lot more things. Harpreet: [00:45:10] Yeah, definitely. Definitely want to hear what Vince says, but I mean, just a question that popped into my head. Maybe we can just let it kind of simmer in the background is is, you know, I guess as somebody who is, you know, let's say you are trying to get into data science, then just be laser focused on the type of company you want to work at. Because if you Harpreet: [00:45:26] Are trying Harpreet: [00:45:27] To work for, quote unquote legacy companies or, you Harpreet: [00:45:30] Know, companies that have these Data Harpreet: [00:45:32] Maturity issues, you will probably need a different skill stack that is well beyond just machine learning and computer science stuff to be successful. But if you are laser focused like, yes, I want to work at a company where there are Harpreet: [00:45:44] Major offering is machine learning Harpreet: [00:45:45] Or their major offering is data Harpreet: [00:45:47] Based and Harpreet: [00:45:49] Then double down on Harpreet: [00:45:49] That. I don't know, just putting that out there. Harpreet: [00:45:52] Let's go to a real Speaker5: [00:45:53] Quick context I wanted to add just kind of building on that is like I specifically wanted to go to a startup or the Data maturity was there so I can actually [00:46:00] do my job but not fully developed because for me, like I want to start my own startup in the future and I've tried in the past. And for me, I'm like, I'm going to face those problems, so I'd rather learn on someone else's kind of budget to figure out, like, All right, we started here and this is what took to get to this next level, you know, so I can go through those growing pains, not with my own company. Harpreet: [00:46:21] Then let's let's go to you. I know the conversation might have meandered a little bit, but let's go back to that original question that after we go to Mexico, Speaker6: [00:46:29] I think just context, any company that does tons of acquisitions that buys tons of companies buys tons of startups. You have pockets of like you have the main line, which is super mature. Harpreet: [00:46:43] You know, that's the reputation Speaker6: [00:46:45] That you get for a company like Wal-Mart. Harpreet: [00:46:46] But then you Speaker6: [00:46:47] Have, you know, like Wal-Mart has bought so many companies, they've probably forgotten about some of the companies they've bought, like there are companies that just sit out there. They're like, Wait, we own you guys, Harpreet: [00:46:56] You know, and Speaker6: [00:46:57] When they're doing an audit, they discover it. And that's when, like your technology organization gets into that group and starts Harpreet: [00:47:04] Bringing them up to Speaker6: [00:47:04] Speed and getting them more mature. So, yeah, any big company that does a lot of acquisitions, it's it's chaos down in those acquisitions. But when you start talking about where the field is going, Harpreet: [00:47:16] You know there's Speaker6: [00:47:18] There's going to be a movement towards and I understand the thoughts around generalization. I understand the thoughts around, you know, the practical nature of some companies just aren't that mature. Harpreet: [00:47:30] But we're kind of at the Speaker6: [00:47:31] Point where if you're not mature by the end of this year or the middle of the next year, you're done. I mean, there's just companies that aren't going to be here in two or three years and it's going to be, you know, I talk about this a lot. It's like a great machine learning die off of companies. That's coming Harpreet: [00:47:45] Because at some point you're so Speaker6: [00:47:47] Far behind, you're never catching up. Your companies are just your competitors are just going to lap you. So be really careful when you look at your first company because, you know, and I'm not being. This isn't hyperbole. Like in two years, a bunch of companies just [00:48:00] aren't going to be here. And some of those are going to be massive names, you know, and you're going to look at it and go, Wait, I started my career with this great company and it just died. You know, it's and it's a horrible realization when you're in one of these Harpreet: [00:48:12] Companies because of your Speaker6: [00:48:13] First year, you don't see it for your first year. Everything makes sense. Everything's great. And so if that company has 18 months left to live, you can almost your career walks off a cliff. So be really, really careful about the company that you go into. Low maturity companies are not a bad thing as long as they have a plan and a strategy, like if they have a roadmap to get better than definitely, you know, you're looking at a good opportunity if they don't run screaming. And that also leads to the necessity for specialization because in order for companies really to mature and deploy reliable, reliable models into production that actually work, that actually make money that will do what the business needs it to do or do what the customers need it to do. Harpreet: [00:48:56] There's just another level of capability Speaker6: [00:48:59] That you need that requires people to start really noshing into areas of specialization. I mean, you even see machine learning. Engineering is kind of splitting in half where you have an automation tools and automation Harpreet: [00:49:11] Side and you have Speaker6: [00:49:13] A customer facing platform side. And so watch really what people are building. And you know, the follow the Data is a great, great piece of advice. Harpreet: [00:49:21] Follow really where Speaker6: [00:49:22] Startups are going and follow the research as another great piece of Harpreet: [00:49:25] Advice. Speaker6: [00:49:25] But look at what's getting to market and follow the products, follow where people are deploying and being successful. And if you want to have a really successful career over the next five years, Harpreet: [00:49:38] Look at what they fail at. Speaker6: [00:49:39] Because if you get good at what companies don't seem to be able to execute on, that's a valuable career path. So it's another way to kind of figure out what you want to get into and where you want to go. Just, I mean, I can't overemphasize be really, really careful about which company you go into because there are some that are just headed off a cliff. Harpreet: [00:49:59] Let's go to Mexico, [00:50:00] then, Greg. After that, sorry, I Speaker8: [00:50:02] Totally lost my train of thought because I was making jokes about my relatives. I'm sorry to hear that. Harpreet: [00:50:08] Oh, do you want to come back to you or trends? Speaker8: [00:50:12] Trends? Yes. One thing I would say, though, because a couple of us we were speaking on a panel yesterday for people interested in machine learning, engineering and that question of like, how do you get into the field comes up and what you should know? And it's like, Well, you probably shouldn't take your career on like the super like entirely all like the super new, like sexy tech that just seems to just because of the risk. So I would say that like if you're a new data scientist or bootcamp grad, like really, make sure you're you're creating like a portfolio of experiences. The other reason why that's kind of important is that like, I think until you sort of get into your role like Harpreet: [00:50:49] It's a filtering problem, right? Like if you. Speaker8: [00:50:51] Have experienced you're not going to know what to like, filter out in terms of skills and topics and stuff like that. So it's one of these things where like sticking with patterns that are sort of tried and true. So for example, instead of trying a fancy new like like open source like Python Library Harpreet: [00:51:08] That maybe Speaker8: [00:51:08] Only has like five stars and maybe one or two like, you know, computers, contributors, maybe instead going with Typekit learn Harpreet: [00:51:18] Or something like that, right? Speaker8: [00:51:19] And instead of jumping straight into Haskell or Julia or Scala, it's like maybe spend your time learning SQL and Python and or like and or Python and R and Harpreet: [00:51:31] Or that like Speaker8: [00:51:32] Super cool language. Right. So that's all I would say for campgrounds is it's very easy to get super distracted by fancy stuff. But the reality is that there's still a lot of people that like program and see, they probably don't want to, but they make good money. They might like rust might be the most popular language of a Stack Overflow. But guess how many people are actually like making their living off of like, see, Java, C++, Python? All [00:52:00] of that, all the all the good stuff. So I would just say that make sure you're like appropriately balancing that portfolio of skill sets. Harpreet: [00:52:09] Yeah, the lesson is, if you're kind of Harpreet: [00:52:11] New, then just Harpreet: [00:52:13] Focus on the stuff that you know you'll need to Harpreet: [00:52:15] Do and not get Harpreet: [00:52:16] Distracted by trends or shiny objects. Just, you know, there's stuff that you're invariably going to need to know to be successful as data scientists focus on that, get good and then explore after after that. Harpreet: [00:52:28] Great. Go for it. Harpreet: [00:52:29] I hope that was the lesson I could have been put together if it wasn't. Harpreet: [00:52:32] Now that was great. Great. Go for it. Speaker8: [00:52:34] Mine was kind of like more of a question for Vin and anybody who was to give their opinion. You mentioned, like some companies will go would go away in the next years or so if they don't Harpreet: [00:52:47] Really Speaker8: [00:52:49] Adjust their usage or their ability to create value in Data. I also see an opportunity for like legacy industries like the oil Harpreet: [00:52:57] And gas, for example. Speaker8: [00:52:59] I feel like I think they are getting more open to accepting that getting value from Data is super important to kind of understand. For example, you know, the landscape Harpreet: [00:53:12] Of Speaker8: [00:53:12] Of Harpreet: [00:53:13] Their product, Speaker8: [00:53:14] I guess, like when they're trying Harpreet: [00:53:15] To plan Speaker8: [00:53:16] Or for cost optimization and things like that with this industry being so old and the manufacturers being Harpreet: [00:53:24] So mature in their Speaker8: [00:53:25] Process, but not Harpreet: [00:53:26] Necessarily up to par Speaker8: [00:53:28] With the way they leverage Data. Harpreet: [00:53:30] I mean, could a Speaker8: [00:53:31] Petroleum engineer say, Look, I'm going to learn email to kind of bring value to this industry and I can be a winner Harpreet: [00:53:37] Then, right? So, you know, is there Speaker8: [00:53:40] Still opportunity for these Harpreet: [00:53:41] Old industries Speaker8: [00:53:42] To win back the shipping industry, for example? Right? We know it's not going away, but there is a Harpreet: [00:53:47] Ton of there are tons of Speaker8: [00:53:49] Opportunities for AI and ML to kind of Harpreet: [00:53:51] Streamline it and make it better. What are you guys? Speaker8: [00:53:53] What do you guys think? Harpreet: [00:53:55] Let's go to end for this one and then let's go to Serge after that and [00:54:00] lots of great comments coming in to you, so I'll make Harpreet: [00:54:02] Sure to get to that as well. Over in, Speaker6: [00:54:05] Yeah, I'm not going to wish death on any company. Macy's, I wouldn't in a million years, you know, say something like, Yeah, no, I wouldn't do that. That kind of person. But when you look at companies that are in industries like what you said, I mean, you look at Maersk. They are. If you look at the way they consume and utilize Data for optimization, world class Airbus old company, you would think legacy know some of the absolutely most cutting edge data and analytics and edge computing or embedded systems. Sorry, I hit the buzz word my bad. But you look at Harpreet: [00:54:46] Just Speaker6: [00:54:47] Some of these industries where you're like, Oh, that's legacy. No, and they are completely modernized, using a ridiculous amount of data, intelligently not like chasing shiny objects like a whole lot of other industries are doing. They're doing, you know, what's the simple solution that we can deploy that works? And then you have hardware constraints with Airbus that is possibly the most realistic company on the planet when it comes to how to deploy and monetize data science and machine learning. So I don't think it's the legacy industries that are in trouble, like shipping is going to be there forever. But I think they're going to start moving more and more towards our autonomous because I mean, it's a ship in the ocean. They don't really have to worry about hitting anything, you know? So when people talk about autonomous use cases, it's like, why wouldn't you? You know, we've got Autopilot's already. What? Why wouldn't you continually just work towards, you know, and that's what those kinds of companies are working on. Those are the high value use cases for them. So there are a ton of these legacy companies that have massive opportunities to use Data, and they've figured it out. You know, like I said, there's so much more realistic. You know, it's really just this pragmatism where, well, you know, the Pirates are automating too. Speaker6: [00:55:59] I [00:56:00] mean, they've seen the writing on the wall. They understand that, you know, they don't modernize, they're not going to be able to get the AI driven ships to where, you know, like the Terminator as they're on board. But, you know, and that's the that's the other side of this is, you know, in response to competition, companies are going to quickly modernize when they realize, you know, this is my last chance. And so there's going to be a ton of opportunities and legacy legacy business industries. But I don't think that, you know, when we think about them as existing or older, I don't think they're going to stay that way for very long. But you're going to definitely see consolidation in those industries because smaller players just don't have enough money and money's not free anymore. Used to be, you know, it was free last year and it's going to start getting expensive. Startups are going to be getting a good influx of cash, but everybody else is just, you know, the pipes off. So yeah, I think legacy industries have a ton of opportunity, but it's only going to be for a very small number of consolidated players. And I think interesting, interesting, interesting failures are coming over the next year, year to 18 months. Harpreet: [00:57:06] Thank you. A lot of Harpreet: [00:57:07] Questions coming, not questions, but comments coming down LinkedIn and things, which might have been some points earlier, but close to the same interesting point, then sounds like what's been happening with autonomous vehicles. A number of those companies won't be around in 18 months. Rodney Beard saying We have the same problem. We are facing competition with global NGOs that are doing data science and machine learning. This will eventually impact funds. Harpreet: [00:57:28] Oracle is like Harpreet: [00:57:28] That, then some things super advanced and other services or software is acquired frozen in time and neglected Harpreet: [00:57:35] Question coming in. Well, actually, it's a comment. Harpreet: [00:57:37] Then a question from Dave Brown. By the way, if anybody has a questions here in the chat, Harpreet: [00:57:41] Let me know. Dave Brown said something Harpreet: [00:57:43] About local government had a choice between getting a job or taking an internship didn't have the confidence, so I took the internship great learning experience, but they're both behind. Harpreet: [00:57:52] Their answer to anything was don't worry, they have RBI, Harpreet: [00:57:56] So we could do this in Python. But if we used R or Python [00:58:00] and got done where we have power, I finally asked for a real answer. I just I to read the rest of that. But he's asking, Have you ever been in an interview and rejected because you were too good? I've been Harpreet: [00:58:11] Told that Harpreet: [00:58:12] I need to dumb myself down, improve my chances, but I don't know how. Harpreet: [00:58:16] I don't even know how to respond Harpreet: [00:58:17] To that mark. Go for it. Speaker5: [00:58:19] Maybe, maybe not necessarily too good, but I knew I was looking Harpreet: [00:58:22] At a train wreck. Speaker5: [00:58:23] Essentially, it's coming out of boot camp and I had an interview for a Data engineer Harpreet: [00:58:29] Role, and they Speaker5: [00:58:31] Basically described to me, Yeah, Harpreet: [00:58:32] I ask like basic question Speaker5: [00:58:34] Like, what's the tech stack like? Wait, what are you doing? And I'm like, Oh, it's all excel. We want you to manage this five hundred person company. Harpreet: [00:58:41] Like all their data via Speaker5: [00:58:42] Excel, that's Data engineer. And I instantly knew that was a train wreck for Harpreet: [00:58:48] Anyone, and I felt bad Speaker5: [00:58:50] Because my friend recommended me the job. And so I sent them a kind letter, hopefully kind saying like, Hey, you may want to reconsider this Harpreet: [00:58:57] Because Speaker5: [00:58:58] Any Data engineer that's going to fix this Harpreet: [00:59:00] Problem is not Speaker5: [00:59:01] Going to use Excel. Harpreet: [00:59:03] Matt, I go for it. Speaker8: [00:59:05] I had something similar happen to me, so I was interviewing for a data analyst and it turned out what they wanted was the Tableau developer. So I went through four rounds of interviews. Did the SQL test Harpreet: [00:59:16] Got like really good, Speaker8: [00:59:17] Went straight through? It did really Harpreet: [00:59:19] Well. Final interview was Speaker8: [00:59:21] The manager straight up asked me, Are you sure you're not going to be bored here? And that right, there is one. I knew that it wasn't that they weren't really looking for what I was, what I had. So it was eye opening experience. Harpreet: [00:59:35] I'm curious if anybody else has had an experience. Yeah, I mean, I don't know how it Harpreet: [00:59:39] Respond to anybody. I ever said, dumb yourself down. Harpreet: [00:59:41] That would be me. I don't know. Harpreet: [00:59:43] Sure, Ben probably Harpreet: [00:59:44] Has some stories to share. Have you ever been? Have you ever showed up to a place? And this is like, this is not Harpreet: [00:59:51] Not even worth you can pay me as much as you Harpreet: [00:59:54] Want to pay me to still be worth my Speaker6: [00:59:55] Time? Yeah, I think so. My early, early in Data science days, [01:00:00] I had to politely Harpreet: [01:00:01] Nope out of Speaker6: [01:00:02] Interviews. Harpreet: [01:00:03] And I've gotten, you know, I've Speaker6: [01:00:04] Gotten good at politely just saying, though, I, you know, I would not provide value. That's my new my new no is. I don't believe I would provide the kind of value that you're looking for. But yeah, I mean, the story I always tell is I was in an interview in 2012, and I had somebody asked me for my GPA and I was just like, what? Harpreet: [01:00:23] You know, and I Speaker6: [01:00:24] Didn't graduate in 2010. Let's just put it that way. I mean, we were not even that's not even the right decade for my graduation. So. And yeah, you almost have to self screen at some point during interviews where if things get too surreal, just I mean, I know you, I know every opportunity feels like a good opportunity when you're trying to break into the field and you're trying to get just trying to land a job in the first place. And so. Feels like any job will do, but don't if you get into an interview where Harpreet: [01:00:54] You just start getting Speaker6: [01:00:55] Asked stupid questions or Harpreet: [01:00:57] They start talking about, Speaker6: [01:00:59] You know, particular technology that just doesn't make sense for what they're trying to apply it to. You know, sometimes people get stuck with legacy, and that's fine. But if you don't hear them sort of acknowledging this isn't the perfect stack and we're going to move someplace else. Yeah, yeah. Don't dumbing yourself down. You know, unless they're bringing you in is like that, that senior plus plus person who's going to save the Harpreet: [01:01:22] Day, you know, if they're asking you to Speaker6: [01:01:23] Put on a cape, that's one thing. But if they think they're the ones who are right, run screaming and don't be afraid to run screaming for an interview. Harpreet: [01:01:31] You know, if you have to. Speaker6: [01:01:32] If you have to politely ask them to find you an exit, just do it. Harpreet: [01:01:36] That'd be that'd be entertaining. Harpreet: [01:01:37] Yeah, this is. We're going to go ahead and end. Harpreet: [01:01:39] This guy's not not worth any of our times anymore. Let me know if anybody else has questions. Scanning the Harpreet: [01:01:45] The chat Harpreet: [01:01:46] Doesn't look like there's any questions coming in there or Harpreet: [01:01:48] On LinkedIn. So let me know. Harpreet: [01:01:51] Matt ask questions Harpreet: [01:01:53] Does not Harpreet: [01:01:54] Look like. All right, Speaker8: [01:01:55] Greg. Yeah, one thing I had, Harpreet: [01:01:58] I think I can't Speaker8: [01:01:59] Remember if it was [01:02:00] Yuvan or somebody else. You guys mentioned a lot about, you know, using I is the I think it was one Harpreet: [01:02:07] Of your Avago Speaker8: [01:02:09] To your articles then where you think Harpreet: [01:02:13] Talking about AIs, Speaker8: [01:02:15] The engine behind the Harpreet: [01:02:19] Metaverse, Speaker8: [01:02:20] Could you could you explore a little bit of that without spilling the beans in your Harpreet: [01:02:24] Article? Speaker8: [01:02:25] Because I know, you know, a lot of people will find Harpreet: [01:02:27] Value behind your Speaker8: [01:02:29] Article. So just wait to understand like what is your thought process there and how do you think is going to Harpreet: [01:02:34] Solidify that Speaker8: [01:02:36] Reality, I Harpreet: [01:02:37] Guess. Yeah. I mean, Speaker6: [01:02:38] When you think about Metaverse, I think the majority of people think too grandiose to begin with. And I think that's our biggest problem Harpreet: [01:02:47] Is we think of this automatic, Speaker6: [01:02:49] You know, all of a sudden, everybody's going to wear some headsets and we're all going to be doing virtual meetings in in the Metaverse. And, you know, the Multiverse of Madness will come up and like, I made a joke on machine learning plus ml in way home, Harpreet: [01:03:02] But we're not going to go there. Speaker6: [01:03:05] That's not where we go first. Metaverse already exists, and once you realize that, then you can start digging into the technologies Harpreet: [01:03:12] That support Speaker6: [01:03:13] What the actual metaverse is right now. Contents in the Metaverse. How do you navigate content? I mean, when you have a Harpreet: [01:03:20] Massive metaverse, you Speaker6: [01:03:22] Need something like Google, which can help you find the kinds of content experiences that you want to have. Harpreet: [01:03:28] And so that's Speaker6: [01:03:29] That's machine learning. Like, you can't navigate that catalog, you can't navigate a virtual space. You can't do that unless you have advanced machine learning platforms that enable that experience. Otherwise, it's just a content catalog that no one can consume. If you have, you know, like Disney's, the example that I continually use because if anyone keeps Facebook up at night, it's Disney. Because Facebook is all online. Disney has parks, Disney has its own content and it's a juggernaut of content creation. [01:04:00] They have a robotics division that they don't talk about very much, but it's really, really good. And you might be seeing like a show in the park over the Harpreet: [01:04:08] Next year that starts using robotics. Speaker6: [01:04:12] You might be seeing some attractions very soon that use robotics, and you've got to start thinking about that. They have an app and multiverse right now is apps. You know, there's a Disney genie. They are incentivizing people to use their apps when you go to the parks, and now they have a metaverse component because and people overlook simple overlaps like this, where you're at a Disney park, you have an app, you don't need a VR headset, you have this experience with you. And now how does Disney tie in games? Do it through the app? How do you tie in serving content to people who are bored in line? Do it through the app. How do you monetize when you create a marketplace? All of that's driven by machine learning. You can't do any of the things that we've talked about Harpreet: [01:04:58] Without Speaker6: [01:04:59] Machine learning, because without that, as the platform behind Metaverse, there's nothing there. You simply can't create the kind of experiences unless you put machine learning as a platform at the very center of it. The build your experience is on top of machine learning. And that's really, you know, when I talk about credibility in the multiverse or in the metaverse, credibility in the metaverse is really, have you ever deployed a machine learning platform, a sophisticated machine learning platform? Harpreet: [01:05:26] Have you? If you haven't, you all want to be. Speaker6: [01:05:28] And it's that stark. So I think when you look at metaverse going forward, I think if you look at those really simple use cases that are out there now that are working in those companies that are intelligently integrating Harpreet: [01:05:41] What they already Speaker6: [01:05:42] Have into the multiverse and using machine learning as a platform to support that, those are the ones who are going Harpreet: [01:05:49] To win because. Speaker6: [01:05:51] They're not trying to create like this pie in the sky version of it, they're going to incrementally get people Harpreet: [01:05:57] There and they're going to be the Speaker6: [01:05:58] Ones who Harpreet: [01:05:59] Have Speaker6: [01:05:59] Multiverses [01:06:00] that people want to be a part of because the Metaverse isn't like this one thing, it is a multiverse. And the companies that are being pragmatic about it right now, like Amazon, your company has a scary, scary where you guys are right now because there is so much opportunity and you already have the platform of all machine learning platforms. So, you know, the opportunity for Amazon is just ridiculous. And and you have that overlap between physical and virtual and you have a content catalog, you know, and so you start looking at companies that way and you look at who has that opportunity, who's deployed the platforms Harpreet: [01:06:40] That could support Speaker6: [01:06:41] A multiverse. And you start realizing, OK, these are the winners. Those are there's just no chance because if it's just a multiverse by itself, you know, if it's just content, it's not enough to be a multiverse or a metaverse of any sort. Harpreet: [01:06:56] And it's actually pretty interesting, right? It's one thing to put on headsets and just look, it's one thing to put on headsets and just look inside of, let's say, a Whole Foods. It's another thing to put on headsets, walk around the Whole Foods, pick up items, check out with them and then have them delivered the next day type of thing. Harpreet: [01:07:13] Yeah, that's a lot of Harpreet: [01:07:15] Generative models, for sure. Harpreet: [01:07:16] I wonder what gnarly Harpreet: [01:07:17] Gans those are going to be. Greg, any follow up or. Speaker8: [01:07:20] Oh, that's I think that's pretty good. That's a pretty good way of looking at it and making everyone think and especially how you see things going forward. You know, the follow up question that I have is kind of like you guys probably heard the announcement from Facebook talking about this supercluster. So you're talking about sixteen thousand five flops, whatever that means Harpreet: [01:07:46] Or the buzzwords, Speaker8: [01:07:47] A giant computing thing. Right? You know, a lot of people are saying this is going to be the engine behind all of that computation. All of this persistent, reliable [01:08:00] computing power that will power this AI model that you speak of for the Facebook version of the Metaverse or what are you guys seeing that will come out of this super cluster? I see them smiling. So you probably have an opinion about that. That's good. It'll be a cluster of clusters. Speaker6: [01:08:23] I mean, there's a whole bunch of press releases that are cool, and I like them because you know, what Facebook is doing is they're trying to move everyone forward. They're trying to get people to invest in what Facebook understands Harpreet: [01:08:39] Must Speaker6: [01:08:39] Move forward in order for all of the stuff that they want to do to be supported. And that's really what Facebook's announcement was. It was basically, look, you as companies not named Amazon, Microsoft and Google need to start building out more advanced compute. Or you are going to be reliant on companies named Amazon, Microsoft and Google. And that's the strategy for those big three companies is you will build your metaverse on our platform. So we make money, you know, even if you don't use our metaverse, you run yours on ours. So, you know, and that's the strategy. And Facebook's really saying, look, you have to create more advanced compute or you as a company will always be beholden to someone who you may also be competing with Wal-Mart. So there is, you know, there is this sort of awareness that they're trying to bring to the marketplace with these hyperbole type announcements, like talking about the metaverse and talking about advanced compute, they're going to roll out a model that has a trillion parameters here pretty soon or something like that, you know, because they want to let other people, other companies know it's like, Hey, there's all of this other, more advanced functionality that needs to happen. You know, they're basically sounding an alarm for other companies because they need help competing against Greg. Basically, they need help. They need to build a coalition [01:10:00] and they, you know, they see the writing on the wall. And so you're going to hear announcements like Harpreet: [01:10:05] This, and it's Speaker6: [01:10:06] Really Harpreet: [01:10:06] A lot of sense. You know, Speaker8: [01:10:08] It doesn't make sense. You know, I was going to say to ask you a little bit more about that. It doesn't make sense, right? So they're building this supercluster on their own and having, you know, given that they have the the technical skills to maintain it right and even grow it at some point to Harpreet: [01:10:24] Remove their dependencies Speaker8: [01:10:26] On the main cloud Harpreet: [01:10:28] Providers. Right. Because the Speaker8: [01:10:29] Main cloud providers can, all they can do is Harpreet: [01:10:31] Sit there and wait for Speaker8: [01:10:33] People to build metaverse on top of what they already have, right? Harpreet: [01:10:36] So that Speaker8: [01:10:38] Gives analysts out there or investors confidence that Facebook is not going to be at the mercy of another Apple Harpreet: [01:10:46] Game, right? Speaker8: [01:10:47] So where Apple can just squeeze that little juice? The minute they want, they're kind of removing themselves from this, so it totally makes sense Speaker6: [01:10:56] When Facebook is preparing for what you know, what's coming next is quantum internet quantum cloud. They're getting ready for that. I mean, that cluster, they don't really care about this cluster. Harpreet: [01:11:06] They're starting to get smart so that Speaker6: [01:11:08] When the next replacement cycle comes up, they can insert themselves Harpreet: [01:11:12] Into it. And this is Speaker6: [01:11:13] How you do it. That's how AMD started taking on, you know, Intel. That's how Nvidia started taking on Intel is they took small steps in sort of in advance of major pushes in advance of these major next generations that were coming up. Harpreet: [01:11:30] And that's how they've Speaker6: [01:11:31] Always, you know, that's sort of the the playbook is how you compete with an incumbent is you get in, you Harpreet: [01:11:36] Learn at the end Speaker6: [01:11:38] Of the technology cycle to prepare to jump into Harpreet: [01:11:42] The next Speaker6: [01:11:43] Technology cycle. That's what Facebook's doing is they're getting ready for that. They're getting ready. I mean, they're going to go in the same direction when it comes to physical experiences. They're going to get into content, they're going to Netflix, they're going to get into, you know, other [01:12:00] directions where they can start merging. What they have right now is a social network, which is possibly the most powerful Harpreet: [01:12:07] Dataset ever assembled. Speaker6: [01:12:10] And that's their competitive advantage is that things are scary. Know if they never gather another piece of data again. I bet you their models will be accurate for the next five years. Like they have that much data right now and they are that advanced as far as machine learning capabilities are concerned. Harpreet: [01:12:24] So they have strong Speaker6: [01:12:26] Competitive advantages that they're trying to put behind. Some of these new announcements are really aimed at competing with companies they've never competed with before in the past. And this is, you know, I talked about this in one of my posts talking about Disney and Facebook. You know, if they don't figure out a way to partner and if Facebook and Amazon don't figure out a way to partner, and if all of these mega companies that you could never see competing with each other Harpreet: [01:12:51] Don't figure Speaker6: [01:12:51] Out ways that they can partner on this. It's going to be a war of Harpreet: [01:12:55] Annihilation, and you Speaker6: [01:12:57] Don't want to see what that looks like Harpreet: [01:12:59] Because the Speaker6: [01:13:00] Businesses that depend on, you know, that depend on Amazon, whether it depend on Facebook, whether they depend on Disney, Harpreet: [01:13:07] You take Speaker6: [01:13:08] One of those out of Harpreet: [01:13:09] A top tier Speaker6: [01:13:10] And drop them down a peg. All those businesses evaporate. And I mean, we're talking about some serious collateral impacts. And so it's scary to think about. I mean, it's fun to think about the big guys beating each other up. But the implications of that, you know, from a consolidation standpoint and collateral damage standpoint is really scary. Absolutely. Harpreet: [01:13:30] There's a show on Amazon Prime Video. It's called the feed, which Harpreet: [01:13:34] Is essentially the Harpreet: [01:13:36] Metaverse that came out a couple of years Harpreet: [01:13:37] Ago. Harpreet: [01:13:38] And and it just to Ben's point about these companies pairing up in this version of the feature, there are companies like like Facebook, Amazon, whatever, like these giant companies had merged together that are standalone companies now that are so disconnected and seemingly unrelated. But in this vision of the future, they're all merged together. There's the feed and there's another one. I [01:14:00] can't remember the name of it, but it's all about uploading consciousness to the cloud after you die and creating a metaverse there where living people can also visit. I remember the name of that show. I'll get to you. It's on the tip of my tongue. Though this is the biggest risk for Web three, mark talks about what you mean by that. Definitely. Speaker5: [01:14:18] So there is this really great Harpreet: [01:14:19] Article Speaker5: [01:14:21] By forgetting the name, but it's moxie. But you know what I'm really into right now? It's taken up a lot of my free time. It's just Harpreet: [01:14:28] So interesting. Speaker5: [01:14:29] But, you know, I'm always looking for articles are against it because I want kind of a balanced view. Like, what are the pros and cons? Most articles Harpreet: [01:14:37] Suck. Speaker5: [01:14:39] They really don't dove into it. But this person basically like, you know what? I want to get a first impression. Harpreet: [01:14:44] I'm going to build some Speaker5: [01:14:45] Dapps, which are decentralized apps. I'm going to create some entities and just poke around. And they actually did their research. And what they really came up with is, Harpreet: [01:14:53] You know, a huge Speaker5: [01:14:55] Floor of Web3 currently is that a lot of the stuff relies on cloud Harpreet: [01:14:59] Computing to Speaker5: [01:15:01] Run, and by that very definition, that defeats the purpose of decentralization. You're stuck on a centralized source. Now there are movements towards the create decentralized compute power, but you know that's not going to compete with Amazon or Google in the same type of scale. And so people are making a lot of tradeoffs between, you know, being fully decentralized versus using these established services for cloud computing, which is a no brainer use that especially if you're like a small team and you're scrappy and trying to build something. Of course, I'm going to use some compute power from Amazon from a free server, but you know, as you grow, and I think Ben was kind of alluding to this, like if you become kind of a competitor to to those services, you know what's going to stop them from? Ending your contract or saying you violate terms of Harpreet: [01:15:53] Service, you know, Speaker5: [01:15:55] I'm not saying they will do that, but you know, it defeats the whole argument of like, we're decentralize [01:16:00] and that's what makes us secure. When you're fully run in the background on a centralized compute resources Speaker8: [01:16:07] And after this mark is at the end of the day, the fundamental platform that enables everything Harpreet: [01:16:13] Right now outside of Speaker8: [01:16:15] Just the transportation industry that we're business built on top of the transportation industry is the internet. The internet is that layer that allows everything to exist right now in the digital world that allows people to distribute just about anything in the digital world. And and it's kind of like we're going to run in circles and in terms of how we know about 50 percent of folks don't have access to internet. And then you have a company like Tesla launching satellites to create a new layer of way of thinking about the internet, which means you're going to depend on that too. So the other 40, 50 percent who don't have access, we rely on that satellite of there to get access, which means if you're going to build a true decentralized network, then you're going to have to rely on something like Tesla for access to the internet to distribute your content or distribute your information. Harpreet: [01:17:05] So it's kind of like a it's a Speaker8: [01:17:07] Convoluted way of looking at things where there's an entity or an organization always on top, where you have to build on top of it to kind of to kind of win. So the idea of that where intrapreneur is still maturing. But hopefully you'll get to a point where it's really generating value for those who put time to put products out there in a distributed way. Harpreet: [01:17:30] Every super intelligent, smart person I know right now is obsessed with Web three. I just wish I had time in the day to search that, and you all know how efficient I am with my time. Maybe I will make that something to. I dove deep into, Harpreet: [01:17:44] You know, Harpreet: [01:17:44] In quarter three and beyond. Yeah, excellent insight. Thank you very much for kicking off that discussion, Greg. Let me know if there's any other questions coming in. It doesn't look like it. Questions coming in from Russell. I guess [01:18:00] we can table that for next week. All. Thank you so much for four for joining me today. I hope you guys had a great week. Hope you guys had a great session. Tune in next week for the all stars, but also next week on on Wednesday for the comet office hours during a panel discussion talking about all about best practices for managing experiments. Harpreet: [01:18:21] We got a lot of good friends Harpreet: [01:18:23] Coming on the show. Harpreet: [01:18:24] Jonathan Hartley is going to be there. Harpreet: [01:18:27] She's been here in the office hours or very happy hours numerous times. Also on the podcast as well. Susan Shou Chang The guys don't know Susan. Follow her on LinkedIn. She's she's Harpreet: [01:18:37] Awesome data Harpreet: [01:18:39] Scientist and a Harpreet: [01:18:40] Game developer. And then Harpreet: [01:18:41] W Ronnie Harpreet: [01:18:42] Huang, who's a Harpreet: [01:18:43] Research scientist over at Google. Yeah, I'm excited to chat with them this Wednesday, 10:00 a.m. Central Time. It's going to be streamed live on on YouTube and LinkedIn, so if you miss it, you can definitely catch that. Also, shout out to Matt Blosser for coming on to the Comet Office hours early this week. I see Matt still in the room. You guys didn't get a chance to tune into that discussion. Definitely do do so. It's we got Abe Gong, who is the co-founder and CEO of Superconductor. That's the company that found great expectations. Jimmy Whitaker from Pachyderm and Matt Blousy himself. A lot of great insights being shared. Episode released Harpreet: [01:19:16] Today with Lawrence Harpreet: [01:19:18] Marini. So check that out. Last week's episode was Dr. Joe Perez. Before that, Scott Taylor, the week before that was all about blockchain for Data scientists, Harpreet: [01:19:25] With Jeremy Harpreet: [01:19:27] Rush on until Jonathan Rotondo, Harpreet: [01:19:30] Who Harpreet: [01:19:30] Also has a number of courses on LinkedIn Learning. So definitely check out that episode was released on January 7th. Cool. That's a lot of updates, guys. That's a lot of killing time, making sure there wasn't any of the questions. So thank you all so much. Carlos Ben Carlos on the show. Harpreet: [01:19:44] Yet there's a, you know, I've got Harpreet: [01:19:46] Such a long delay with my episode releases. So there is another episode being released with Carlos, but it's one that we had Harpreet: [01:19:52] Recorded almost a year Harpreet: [01:19:53] Ago. So I'm curious to see how his how his predictions line up with what he had predicted back [01:20:00] then. But yeah, like, I've got enough episodes in the backlog right now to release into like July of this year, just taking a break from recording. Yeah, yeah, I'll bring Carlos on again. Definitely, I'm going to read up on Web3 and everything. Just just start educating myself. That's a big goal. In the next couple of months, I might go deep down the rabbit hole with Web Harpreet: [01:20:22] Three Harpreet: [01:20:23] Interviews on the podcast. So if you guys know people who are notable and knowledgeable in that field, that would be willing to come on to a podcast with, you know, a no name podcast like myself, I would appreciate that because that's that's I'm committing right now. That's going to be the next rabbit hole I'll go down into. It's going to be Web3 stuff off the clock and deep learning on the clock. Y'all take care. Harpreet: [01:20:47] Have a good rest of the evening. Have a Harpreet: [01:20:48] Good weekend. Remember you got one life on. Planning, why not try to do some big cheers, everyone?