HH49-10-09-2021_mixdown.mp3-from OneDrive

Harpreet: [00:00:07] What's up, everybody, welcome, welcome to the artist Data Science, happy hour. It is Friday, September 10th and I can't believe it. I really shout out to Antonio for taking over at the wheel. Last week, he did an excellent job. Just listen to the episode yesterday. I thought he did an amazing job. So Antonio,

Harpreet: [00:00:24] Thank you so much for

Harpreet: [00:00:26] Taking over, man. Really appreciate that. Also shout out to Angela Valdez.

Harpreet: [00:00:30] I don't know if you listen or not, but word on

Harpreet: [00:00:31] The street is that you just had a baby. Congratulations. Super excited for you. Super happy for you. Hopefully, you guys got a chance to tune into the episode that was released

Harpreet: [00:00:40] Today with Max

Harpreet: [00:00:41] Frenzel. He's a physicist. He has PhD studying like quantum information, something or other, some amazing stuff turned AI researcher. So we had an opportunity to chat about a wide number of things. He's also he's also a like A.I. research and just like generative music, type of space is really interesting. But he's also an author. He wrote a book called The book is called Time Off and it's all about cultivating your rest that they can the importance of having rest built into your daily life.

Harpreet: [00:01:11] So I think it is a

Harpreet: [00:01:13] Damn good episode that we recorded this back in February and February. I was extremely heavily burnt out, so you can hear a very stressed out Harp in that episode if you want to see what that sounds like. Speaking of Antonio, there he is, Antonio. You just missed the huge shout out I did to you for a taking over last week. Thank you, man. Really appreciate that. Appreciate you all the help and having you, you know, just crush him. And my favorite happy hour episodes seem to be the ones where I'm not at the at the wheel,

Harpreet: [00:01:42] So I might have to make the cut. I might have to

Harpreet: [00:01:45] Make this happen again. But again, man, thank you so much. Appreciate that.

Harpreet: [00:01:50] What else?

Harpreet: [00:01:50] What other needs do I have?

Harpreet: [00:01:51] Oh hey, I just completed my first

Harpreet: [00:01:53] Week at

Harpreet: [00:01:53] Comet this week.

Harpreet: [00:01:54] That was pretty awesome. Shout out to Austin Austin in the building.

Harpreet: [00:01:57] Speaking of like, OK, so I started to

Austin: [00:01:59] Feel to be [00:02:00]

Harpreet: [00:02:00] At a new job.

Harpreet: [00:02:02] I was great, man. Like, I mean, this is it's exactly what I want to be doing. I think it aligns

Harpreet: [00:02:06] Well with what I'm all about and where I'm trying to take my career and

Harpreet: [00:02:12] Just it. Just a lot of things just kind of come together in this role. So happy to

Harpreet: [00:02:17] Step into something that makes me

Harpreet: [00:02:19] Extremely satisfied. A bunch of my

Harpreet: [00:02:22] Mentees actually got jobs in the last week

Harpreet: [00:02:24] Or two. There's like six of them that that got jobs. So shout out to all you guys who are leveling up. I just started a job. I'm wondering, do you guys have any advice for somebody on how they can ensure

Harpreet: [00:02:36] That they crush it in the first

Harpreet: [00:02:39] 30 days? I'd love to hear some advice. I guess we could start with that. Let's start with Tom. Let's go to Antonio after that

Austin: [00:02:46] One word answer. Python always up your python skills.

Harpreet: [00:02:50] Yes, Python skills. Was that a laugh track that you just played, or was that an actual just people laughing in the background?

Austin: [00:02:55] This is people laughing in the background. That's why I'm muting a lot.

Harpreet: [00:03:00] Yes, Python. Happy Python skill, especially if you're a data scientist. 100 percent agree. Antonio, what about you?

Austin: [00:03:06] I'm starting a new job as well, as well as they have our AIs,

Harpreet: [00:03:09] So

Austin: [00:03:10] I'm probably going to need some place to thank you. But I don't know. I think what

Austin: [00:03:16] My my my

Austin: [00:03:17] Last director when I started Verizon,

Austin: [00:03:19] Ask him, what

Austin: [00:03:20] How can I be successful the first 30 days? He's like, I don't expect you to be doing anything. The first six months, man, let alone the first 30 days. You know, by the time we get set up and everything, it takes time.

Austin: [00:03:32] But yeah, I'm going with a mentality.

Austin: [00:03:35] Just be confident, be myself. And, you know,

Austin: [00:03:40] Like that was

Austin: [00:03:41] Since when was my last day at Verizon? All the people are like, you know, like saying buying stuff. Nobody's saying like, Oh, you know, I mean, everybody, like, we've worked on a phone project and what people remember, like the jokes, they remember, the fun times, how you made them feel as especially like the Data Science Data analyst. I [00:04:00] always kind of like people are give me that feedback like you always made us feel like our project is in good hands. You know, even though I messed that up many times, I'm like, Hey, yeah, of course we can do this, you know? So it's been that kind of stuff. So I'm excited to be starting a new journey as well alongside with you.

Harpreet: [00:04:18] I like that. I maintain like a good, positive attitude. Have a good time and share your concern. I guess let's go to Austin for some advice that you know it's weird because, you know, report report to Austin. So let's see what advice nonspecific advice.

Harpreet: [00:04:34] Yeah, yeah.

Austin: [00:04:34] Yeah, I know you need to get in the car.

Harpreet: [00:04:36] No, I'm just kidding.

Austin: [00:04:38] For me, I think the most valuable

Harpreet: [00:04:41] Thing

Austin: [00:04:43] It's like,

Harpreet: [00:04:44] I think just taking taking an interest,

Austin: [00:04:46] Being curious, asking asking questions

Harpreet: [00:04:48] Like sort of being present in that way.

Austin: [00:04:51] It is a really important thing because I think one of the

Harpreet: [00:04:55] Things that either hiring

Austin: [00:04:57] Managers or your team is they're in the sort of work

Harpreet: [00:05:00] Every day and they start to

Austin: [00:05:02] Take things for granted that they assume that, like everyone knows.

Harpreet: [00:05:06] And when you're brand new,

Austin: [00:05:08] A lot of those things, definitely you can't take this for granted. So I think it's really a huge thing is just being curious, asking good questions, you know, finding the right people to talk to and then just just soaking what you can from them. And that should be enough. Like, if anyone's really asking you to do like a whole lot more than that,

Harpreet: [00:05:25] Then I think there have unrealistic

Austin: [00:05:27] Expectations. But I really think like the questions of the curiosity is what really helped me get oriented a comment and sort of realize what we were doing and where we were headed.

Harpreet: [00:05:35] And that's that sort

Austin: [00:05:37] Of helps, I think, in the medium

Harpreet: [00:05:38] To long term as opposed to

Austin: [00:05:40] Trying to overdo it in the

Harpreet: [00:05:42] Short term.

Austin: [00:05:43] I think that's sort of the key for for the first couple of weeks or so, at least, even if not more great tips.

Harpreet: [00:05:49] Thank you very much. Austin also shout out to a couple of people I ain't seen in a very, very long time. We go, What's going on, man? Good to see you again. Carlos in the house. Carlos must have heard all the NFT talk we've been having over the last month. [00:06:00] Where were you during during that time? We really, really would have liked to have you there. So let's let's hear from you, Carlos. Any tips? Let's go to Carlos and after Carlos surge concert. That's a good comment here in the chat. But any tips

Harpreet: [00:06:13] On, you know, for people who are

Harpreet: [00:06:16] Starting a new job on how to, you know, just crush it in the first 30 days?

Carlos: [00:06:20] Yeah, that is exactly why I'm here. But, you know, actually, I'm leaving my job. I took the first plane up, so I'm going to be doing entrepreneurship stuff.

Harpreet: [00:06:33] Oh yeah.

Carlos: [00:06:34] Graduations, man. Same advice. I need that advice. What should I do to really impress clients if I meet them? But I think to answer the question, there's still

Harpreet: [00:06:45] A little things.

Carlos: [00:06:46] So like when I first started in consulting, I noticed like I would put a lot of meetings and no one took notes that and they were highly technical meetings, and I was like, Oh, no one's going to take notes. I guess I will take like insanely good notes and email them out, and I'll just run all of the PowerPoints and read all the new taking and do all of the boring admin stuff. Because when you do that, you'll start to be the person who knows most of what you're supposed to be doing and everyone's going to go to you and just be like, Hey, like, how can I best contribute? So I kind of forgot I wasn't paying attention. Whatever new things happened

Harpreet: [00:07:17] And just being the person, just a

Carlos: [00:07:18] Little extra things. People will call you up really quickly to do the important things. So I would get pulled into proposals like way soon as other people in my life, in tasks like way more stuff, a little more high value stuff, so do a little things memorable.

Harpreet: [00:07:34] You very much Carlos. After Carlos, go to let's go to Serge and then Monica just sharing advice, general advice on how to just crush it in the first 30 days. Then after that, I'm just going to go to Russell for a question on

Harpreet: [00:07:45] On

Harpreet: [00:07:45] Nfts or crypto because you've been having a bunch of those questions. I think

Harpreet: [00:07:48] Carlos is.

Harpreet: [00:07:50] Carlos is here. We don't get him very often, so go for it.

Carlos: [00:07:53] Well, I think it's good to start like, you know, like with fire. Yeah, [00:08:00] because I think the notion is that you should start like, take it easy, kind of soak it in, like get to know everybody and and you

Harpreet: [00:08:10] Know,

Carlos: [00:08:12] It takes a while to get you truly onboarded. But I think it's important to figure out what are the biggest challenges in the organization as soon as possible and then think of all the crazy ways to fix them. And a lot of people along the way will tell you things like, No, nobody's figured this one out or it's not a big deal. We've we're going around it, you know, figuring out other ways around it. And and that outside of the box thinking over time, you have less and less of

Harpreet: [00:08:43] It because you just become kind of, you

Carlos: [00:08:45] Know, it becomes like part of the organizational culture and structure that you kind of deal with it the way the organization deals with it. But early on, you have all these these ideas, and I'm saying it's it's best to kind of, you know, get on with it as soon as possible and and challenge the system

Harpreet: [00:09:04] Before you become

Carlos: [00:09:06] Part of it a little bit. I don't mean that in a bad way. I mean, like every organization I've

Harpreet: [00:09:11] Been in in the first

Carlos: [00:09:12] Few months, I remember

Harpreet: [00:09:14] That of all

Carlos: [00:09:15] This. But why? But why do you do it this way? And it's important to speak out because, you know, they they you, they start to see the reasons why they hired

Harpreet: [00:09:26] You to begin with. And also,

Carlos: [00:09:27] You might hit something some home runs early on that are not only memorable to others, but they also kind of set like a foundation for what you're going to do moving forward.

Harpreet: [00:09:40] So the amazing advice. Thank you so much, Serge. Let's go to Monica and then after Monica will go to a we'll go to Russell for some crazy and PhD question. And I have a good one. By the way, if anybody has questions, whether you watch it on Twitch or on YouTube or on LinkedIn Live,

Harpreet: [00:09:57] I see you guys just

Harpreet: [00:09:58] Go ahead and drop your comment right [00:10:00] there in the chat or your question right there

Harpreet: [00:10:02] In the chat. Make sure I'll get to you.

Austin: [00:10:03] Hey, I just got to acknowledge before. Monica speaks

Harpreet: [00:10:07] That her new do is

Austin: [00:10:09] Freaking awesome. Thank you so much. Where did you

Harpreet: [00:10:15] Get your hair? It looked like it was tied back.

Harpreet: [00:10:16] You did. Oh, nice.

Austin: [00:10:18] I cut it and dyed it back to red, so I did. I did the whole color spectrum, and now I'm back to red.

Harpreet: [00:10:26] So thank you. These are

Austin: [00:10:30] Great. These are great. Great tips. Also, it seems like everybody has a new jobs lately. I also have a new job. Not even two months in. So the asking the questions, the why questions are the best. And I'm already starting to kind of like be absorbed in it and like, everything's now more normal. So the why questions are kind of fading away. So get those out as fast as you can, especially if you're in a high pace environment, you're going to be absorbed super, super quickly.

Harpreet: [00:11:03] Another thing that I've found

Austin: [00:11:05] Helpful is not only reach out to your team members and have like coffee chats, get to know them as well, but reach outside of your direct apartment or your direct organization that you're working with to get a perspective of, like your company as a whole or like an outside perspective of your department and how you can help others within the company. And that's somewhere where you can

Harpreet: [00:11:29] Really

Austin: [00:11:31] Help the company as a whole.

Harpreet: [00:11:33] Thank you very much, Monica. Appreciate that. Great tips. Yeah, I mean, like, this is the first time I worked at a company that's, you know, kind of small. There's like, I think, 40 something people had comment when I was at bold. I was like 300 people, 30 30 people. So I was like the smallest up until then. So now it's cool to be in a smaller organization, I think gives more of an opportunity to

Harpreet: [00:11:52] Create these kind of relationships that

Harpreet: [00:11:54] You're talking about. Eric Sims had to bounce, but he left some great tips in the chat. I'm just going to read them. [00:12:00] Eric said he started a new job about 100 days ago, and his boss told him that in the first 30 days, he

Harpreet: [00:12:05] Should focus on three things one,

Harpreet: [00:12:08] Establish strong

Harpreet: [00:12:09] Relationships with my key stakeholders and figure out

Harpreet: [00:12:12] Who they actually are to figure out how to actually find the Data. And three learn what levers can be used to move the KPIs. Great. Great tips. Up top. Greg, good to see you again. So, Greg and Russell, you guys always have some amazing PhD questions and you're lucky because Carlos is here, so

Harpreet: [00:12:31] We'll let Carlos

Harpreet: [00:12:33] Take those questions on. But Carlos, this year, I need it as well. Do I need questions or comments or are you just waiting for the question from Russell? Yes. All right. Go for Russell.

Russell: [00:12:43] Thank you. Evening, everybody. Just before I get on to the NFT stuff, one quick comment about starting a new job. It may be a a no brainer, but I think it's important to not be an ass in a new job, you know, especially if you're triggered easily by things, if things in a different way. Just kind of let it flow, come back and pick it up later on. Soak up everything like a sponge and try. And you know, if you if you receive negativity transferred into positive and put it back out there. I think that's a great way to start a new job.

Harpreet: [00:13:10] Big tip. Yeah, absolutely.

Russell: [00:13:12] So the the NFT and the crypto stuff, we've had a lot of conversations recently about NFTs. But if I can put it to call us firstly about crypto, I posted something earlier today about MasterCard buying or acquiring a crypto company called CipherTrace and put it out there. But this is one of the first big financial organizations I'm aware of that have invested in in the crypto

Harpreet: [00:13:38] Market, really,

Russell: [00:13:40] And then went on to extrapolate from that. You know, what is the likelihood that crypto can become a more dominant currency, a competing currency with

Harpreet: [00:13:50] Orthodox

Russell: [00:13:51] Currencies? And then further asked, you know, would people be happy to be paid in crypto? So I mix a whole lot of stuff in a single post. It was perhaps a little [00:14:00] confusing, but really interested on currencies take. He's already put some great comments back on the post in the first instance.

Harpreet: [00:14:07] But if he's if he's able to

Russell: [00:14:09] To elaborate on that now, that'd be great. Thank you.

Carlos: [00:14:12] Yeah. So Twitter owned Square, where I think they also don't cash out. I think the fintech companies have gotten really into making crypto available within their moat, which is important to say I think banks are a natural extension of that. They want to moat as much of that activity as possible. And to that extent, visa, which I think, you know, I think Visa, MasterCard all related to these other fintech companies like Zelle, Venmo, which is owned by PayPal. I mean, they're all connected, right?

Harpreet: [00:14:46] So I think it's

Carlos: [00:14:47] Weird to say like MasterCard or Visa's first versus JP Morgan being first, so it's really all so interconnected. But to your point, Visa got was in the news for buying a CryptoPunks, which is one of the the original first. But it's one of the big original NFT is they spent about $100000 on it. Other companies doing this, Budweiser just bought beer, which is like Ethereum Name Service, address

Harpreet: [00:15:14] A bunch of money for that. I think we're seeing like broader

Carlos: [00:15:17] Is a more broader comment of like big companies are getting in the space. So we need to bring up too is like, you know, why are these companies making these acquisitions? It's because they have regulatory reasons that they're not allowed to invest directly in crypto. So what

Harpreet: [00:15:32] Do they do if

Carlos: [00:15:33] They want bitcoin exposure and they can't buy bitcoin ETFs? They can buy a MicroStrategy, they can buy Tesla so they can diversify their treasury kind of in a derivative way into this space. So I would say, like, I think a lot of institutions are deep in this space and like these roundabout ways, and it makes people not really aware that this is not

Harpreet: [00:15:54] A hypothetical anymore, like it's

Carlos: [00:15:55] Just happening before our eyes. We're going to see more NFT [00:16:00] NFT marketing. We're going to see more companies go around ICO and IPO laws by doing NFT offerings. Yeah, I mean, I think the adoption here, I don't think I don't think it's appropriate to, like, have a

Harpreet: [00:16:12] Conversation around like, is it going to

Carlos: [00:16:14] Come? It's like, it's like, it's your it's here, man. It's like, it's not Netscape anymore. We're deep into Google territory, timeline wise. But then the other point about, like cryptocurrencies competing, there's a very deep economic answer that I won't give you, but I can pull you to a paper that goes into it. The shorter answer is, I think it's an inevitability that the globalization of certain assets will change how government functions like countries that have proven incapable of maintaining stable inflation. They are citizens are just going to opt out. They're going to buy USDC on a theory in a decentralized way, permissioned way. If they can't buy that, they'll buy other stablecoins like stablecoins. Is the revolution first. So in terms of like competing, no ether and bitcoin

Harpreet: [00:17:05] Will never be U.S. dollars.

Carlos: [00:17:07] That's actually that would be kind of bad for them because they're not growing very much. People who want like a stable way of living with a stable currency will just use assets of other countries and opt out of their government in some way. That's going to be hugely destabilizing. For governments that don't know how to run budgets, but in a lot of ways, it's going to be like voting with your dollars like literally I mean, the number one country right now for cryptocurrency adoption is probably Vietnam. People don't really think of that. But Vietnam off the Vietnam War keeps concerns about their currency stability. So give me the longer answer when I want it. But I mean, people are going to opt out of their currencies for the best government cryptocurrencies in the most literal sense, like central bank currencies. See the CBDCs.

Harpreet: [00:17:56] Good economics lesson right there.

Harpreet: [00:17:57] Thank you, Carlos. A follow up question to [00:18:00] that

Harpreet: [00:18:00] Russell or anybody else have questions on NFTs or crypto at all?

Russell: [00:18:05] Definitely. Yeah. So the follow up question is building upon the questions we've had in the in the last two or three sessions here talking about how NFTs are becoming more widespread. Firstly, for the for the more superficial element, which seems to be basically buying collectibles. You know, some of these high value art items or these are iterative pictures like a bored ape and other things like this. But yeah, yeah. But we went on then to suggest and I think it was Mark Mark Friedman's suggestion in the first instance. I don't think he's with us today, but he suggested some two or three months back now that, you know, could NFTs be used to validate the authenticity of a set of Data or even a training model that could then be used in other ML systems later? And how we could do this? And I think Anthony then came up with another great option last week, saying, basically, it could be like a loyalty scheme, you know, like a diner's club card or a or a grocery card that you wouldn't have a physical card, but you'd be given NFTs that you could buy with that would then predispose you or make available to you, preferential rates, et cetera. So we were just kind of spit balling all of the alternative options for PhD, so it would be really interested to hear your take on that also.

Carlos: [00:19:28] I guess push back on some of, yeah, you can NFT anything that you want to be scarce and digital and like perfectly documented and also nonfungible, like you could do that stuff, but really like if you wanted to time stamp data sets. I mean, I don't think you it seems to be that it's like more complex, like directed as silly graphs and like Data ontology things that you can do. There's, of course, like hashing, so you can just have like a Data. So that stamp with the hash of all of its data, which be insanely [00:20:00] large or extremely long process to do. But that would be like perfectly timestamp some data. But I think I think the question is more so like like what's possible and it is NFTs will be like your identity on the internet. That, to me, their most immediate new use. So for example, like why would someone pay hundreds of thousands of dollars for aboard a yacht club? They might do that because it gives them access to the Discord server. And if people who spend $100000 to buy a boat, Abe know this is an exclusive social club and all they have to do is have that in their wallet, which then allows their wallet address to effectively serve as a login to a server or log into a website. Then they're paying for a private digital club, which gives them access to celebrities, basketball players, people who are well-connected. Yeah, I mean, NFTs are definitely going to be a means of like not just collectibles, but like they'll be a pass to clubs that you can resell. And that resell ability is massive because now you can do it a lot more clubs,

Harpreet: [00:21:03] If it

Carlos: [00:21:04] Might let you sell it for profit, like more clubs, if I thought they were like, make a profit playing it, play to earn as the whole model. And NFT gaming world. But yeah, I mean, you

Harpreet: [00:21:15] Can do that right now. Like there's an

Carlos: [00:21:17] Mmorpg called Trivers. I bought way too much money of land in this digital game. It's fixed land. I'm one of the only people with a house that only ten thousand houses in the whole game, and I can flex my NFT in this little world. And I know that whenever I'm bored of this game, I'll just sell and I'll definitely make profit. So that's like a play to earn model there, too. With social club, we're like financialized every human interaction, including ones on the internet. Harp you got to cut me

Harpreet: [00:21:47] Off, bro. Yeah, no, no. That's good. I love it. I mean, it's just like you think about far off into the distant future. Once humans are able to say, let's say humans are able to upload their consciousness, whatever into [00:22:00] some virtual realm like these, NFTs are going to be like the cryptocurrency that you use in this type of environment. Just thinking crazy thoughts way off in the distant future.

Carlos: [00:22:09] Social Security numbers probably safer as an NFT than it is as just a number that someone who knows can just call their bank and be you like Greg.

Harpreet: [00:22:19] So I think Greg has a follow up question here. Go for it.

Russell: [00:22:24] Yeah. So thanks for the for this rundown, Carlos, I know some of the things that I continue to listen to in over and over until I finally get it.

Harpreet: [00:22:35] One thing that's quite

Russell: [00:22:37] Fairly easy for me to understand is how NAFTA can benefit folks like basketball players or celebrities, for example, right where I can see a world and correct me. If I'm wrong, I can see a world where they own NFTs that they can pass on to their children in terms of inside of their wills or while they alive. They can benefit from these digital assets to sell to the public and have an added streamline of stream of of revenue, right using these NFTs. So when you think about the gig economy, the

Harpreet: [00:23:17] Regular folks out there,

Russell: [00:23:20] We know, you know, notoriety or they're not known in the public, but they just want to find a way to earn a living. What can they do

Harpreet: [00:23:30] To profit

Russell: [00:23:32] From NFT? How do they find out what to engage in to earn a living using these NFTs? Do they need to be technical? What is the minimum?

Carlos: [00:23:44] Yeah, yeah. I think you think you brought up a really good point. So much already. Being able to finalize anything is the fundamental thing to get. So like I can financial as my social network because people who know me, [00:24:00] I have some. I have quantifiable amount of clout, I guess, and I can make entities and people might buy them from me. And if I'm very famous, a lot of people buy them from me. So what do you do for people who don't have that? I would argue everybody has some amount of social network and can be financialized in some way, whether it be through no loss lotteries and like collective investment. There's lots of ways that people with very, very small social networks can turn those networks into value. But to get directly to your games on this question,

Harpreet: [00:24:36] We're financially losing everything. That's the main

Carlos: [00:24:38] Thing to bring up. So that means like small artists and, you know, turn their true fans into money. But let's skip art and skip saying that people have a specific talent. How can people with like no particular talent or social network like you? Zenefits There's a game called Axie Infinity, and it is like the number one downloaded game in the Philippines because you can be moderately bad at it. All you do is have to make a talent and like, play the game as if you're infected. To play your little Pokémon rip off monsters will get better, and then you can buy and sell your team and people in the Philippines who are earning hundreds of dollars a week doing this game and actually infinities revenue like skyrocketed with making hundreds of millions a year and the amount of money moved just like the volume of exchange people. I mean, like, that's life changing money in countries like I don't know what the average median income is in the Philippines, but in Mexico, for example, the daily minimum wage is seven

Harpreet: [00:25:39] Dollars per day.

Carlos: [00:25:41] So it's like the poorest we make making $7 a day, and you can make hundreds a week of playing video game on the phone.

Harpreet: [00:25:49] You're going to do that, and that's the thing that makes it possible.

Carlos: [00:25:51] It allows you to financial AIs anything including like time, like a unit of time.

Harpreet: [00:25:55] Playing this game has a price,

Carlos: [00:25:57] And it means lots of games that you used to [00:26:00] pay to play the games. Now you can earn out of those games because people are happy to pay for that exchange of time. Like you grind my Runescape character for an hour and that saves me an hour. I'd pay like five dollars for that because I want to hurry up and level up, so I keep doing this are pretty damn good ones.

Harpreet: [00:26:21] Gregory also have another question.

Russell: [00:26:22] No, I think it's it's it's quite interesting. So because the way the way I look at it is, it's kind of

Harpreet: [00:26:31] Like somebody who wants

Russell: [00:26:32] To go into to commerce. You know, you have a set of products. You're good at manufacturing, which are hands and you want to find a market to sell

Harpreet: [00:26:43] It and kind of like

Russell: [00:26:44] Automate the contracts that you have between yourself, your products and the people consuming it. So I see NFT as definitely a great vehicle for that. So which and for that reason, I'm I'm continuously interested into this in time is definitely

Harpreet: [00:27:08] Something

Russell: [00:27:08] That I need to to monitor, to invest myself a little bit more into it. So I appreciate your answer there, and I also owe you a follow up car lot. So I'll reach out to you offline to discuss more about about that and. I know you're working

Carlos: [00:27:24] On Marketplace, where I won't sell anything for free, no worries, but just to add something to the point. But I was actually talking to one of our Typekit Public Health Consulting, and I was talking to someone on our account team, which is someone who had interfaces with our customers directly about blockchain stuff. And they're asking like, what's the future state of our clients are like not really thinking about? And I was like, Well, I think in the future, we will be able to, like have health insurance companies that have deals with hospitals for specific units of care. But for example, side of things like stroke as this [00:28:00] kind of aftercare for this many days and there's an average for the stuff that they're modeling, they would cause an NFT. Those units of care at a certain amount

Harpreet: [00:28:08] Per year based on what the

Carlos: [00:28:09] Hospital knows it can do like it can handle this many stroke victims. As many of these per day, they would like many units of care. And then in the marketplace, they would sell those units of care and then people. Insurance companies would effectively collect units of care based on their own analysis of their own people who have that insurance. Kaiser Permanente, they have my Data. They'll study me and they'll try to essentially personalize and predict my care as a unit of care level, in which case if they're wronged or whatever they can sell. Those units of care, secondary markets and hospitals can have a percent of revenue from secondary markets. And anyway, that's the feature. Say that things probably less than 10 years away, it might be less than five years away.

Harpreet: [00:28:52] Carlos drop a link to a

Harpreet: [00:28:53] Marketplace, a

Harpreet: [00:28:55] Cb CDP protocol. Oh, you printed it out. I appreciate that. Of course it's hard. Read everything.

Harpreet: [00:29:02] I'm the man I told you.

Harpreet: [00:29:04] I'm ready to invest, man.

Harpreet: [00:29:05] I got faith in you. Me up, man. Iraq's ready to rule.

Harpreet: [00:29:10] I got a question for Greg, though, man, by the way, if anybody has questions, whether you're on LinkedIn, which YouTube like I said, I'm watching, please do drop your questions. I would love to. I'd love to have them. Greg, man, what the heck is quantum ml? Break this down for us at a high level manam. I'm really interested. I know you talked about Quantum L. I believe it was in the podcast with my good friend John Krohn Super Data Science Podcast. But what is this stuff? You know, give us the lowdown on this.

Russell: [00:29:40] Well, this is going to be my non-technical attempt at explaining what quantum email is. It's really speaks for itself, right? So it's bringing

Harpreet: [00:29:51] The

Russell: [00:29:53] Mechanics of quantum physics to to to machine learning. So everything [00:30:00] that you can think of, that a classical computer does it with a bit. You bring that, you convert it into a

Harpreet: [00:30:10] Qubit to

Russell: [00:30:12] Increase or beef up your processing capability.

Harpreet: [00:30:17] So when you think

Russell: [00:30:19] About the different setups in the classical world that you can do

Harpreet: [00:30:24] In quantum, I mean

Russell: [00:30:27] Machine learning, whether it's support vector, linear regression and things like that,

Harpreet: [00:30:34] You can

Austin: [00:30:36] Take a

Russell: [00:30:37] Few bit

Harpreet: [00:30:39] And transform

Russell: [00:30:41] Your variable inputs from a linear

Harpreet: [00:30:46] What do you say?

Russell: [00:30:49] I say so, so when you think about a classical system and you think about a linear regression and you take your inputs and you transform that into an output, so your inputs in the

Harpreet: [00:31:02] Classical computer

Russell: [00:31:04] Becomes

Harpreet: [00:31:05] Qubits in the quantum

Russell: [00:31:08] World realm, where you can design gates, that tells you how to make that computation to arrive at your output in a faster way. So the problem with quantum machine learning is that there's not really a clear advantage to what it can do simply because just because you make

Harpreet: [00:31:31] A model

Russell: [00:31:33] Faster doesn't mean it's great. Right? So they're still studying

Harpreet: [00:31:39] What they can

Russell: [00:31:40] Do to

Harpreet: [00:31:41] Claim an

Russell: [00:31:42] Advantage for quantum machine learning. And right now, what they're doing is leveraging the

Harpreet: [00:31:49] Speed of quantum

Russell: [00:31:50] Machine learning with the accuracy of classical machine learning

Harpreet: [00:31:56] To arrive at greater achievements.

Russell: [00:31:59] So [00:32:00] I'll give you a quick example. Now you have some new term devices that are taking, for example, a quantum system that can provide you with probability of occurrences but will not be accurate until the classical computer reviews. These probabilities adjust the weights and adjust it back to the quantum system for that quantum system to recalculate probabilities until and iterate like this until it arrives at a high confidence level probability. So when you have a classical machine learning or classical system beefed up with a quantum

Harpreet: [00:32:43] System, you can have

Russell: [00:32:44] This iteration done faster to arrive at your higher probability outputs faster

Harpreet: [00:32:51] Than,

Russell: [00:32:52] Say, a classical can do it. So think about millions of variables that you are billions of data that you want to process to drive this. And I'd like to finish by thinking about the different,

Harpreet: [00:33:06] You know, use

Russell: [00:33:07] Cases that you can think of out there. Think about drug discovery. Think about an airplane manufacturer who would like to simulate the best way

Harpreet: [00:33:20] To design

Russell: [00:33:21] Their wings or something like that that ingest tons of Data would like to evaluate tons of designs to arrive at the best design ever that minimizes faults or defects, et cetera, et

Harpreet: [00:33:37] Cetera.

Russell: [00:33:38] So those are the things that are happening right now on the near-term side. But the long term side is that the way I see it is that you will only hear about or you will hear cloud service providers, for example, offering you something like CPU, GPU [00:34:00] or CPU for quantum processing unit and then TPU, right? So our tensor. So it will only be one of the offerings that customers or consumers can switch back and forth with regards to their business cases that they want to analyze.

Harpreet: [00:34:21] All right. A couple of other questions here like, I don't know anything about this,

Harpreet: [00:34:25] So they're dumb

Harpreet: [00:34:25] Questions. Please forgive me. Can I just go to Best Buy and ask for a quantum computer? Will people look at me like, I'm crazy? If I do that,

Russell: [00:34:34] I think people will look at

Harpreet: [00:34:35] You like, you're crazy.

Russell: [00:34:38] So far, the way I understand it, it's very unstable,

Harpreet: [00:34:42] Simply because the way

Russell: [00:34:45] They're designed right now, nobody knows what is the best design to build a quantum computer, put a quantum computer

Harpreet: [00:34:52] Together where they

Russell: [00:34:54] Can eliminate noise completely noise. What I mean by noise is when you think about a qubit. Think about of an electron inside of a closed system. There are noise like vibrations, other things happening in its environment. If it if it interferes with these electrons, the calculation will be at, say, affected and you have what you call error. And the only thing they can do right now knowing that they cannot fully eliminate

Harpreet: [00:35:29] Noise, they come up with

Russell: [00:35:31] Some algorithms to correct these errors. So it's still an active research where you have

Harpreet: [00:35:39] Multiple ways of

Russell: [00:35:41] Designing a quantum computer, whether you want to use photons where you, you want to use a trap ion. You want to use a superconductor?

Harpreet: [00:35:52] It depends

Russell: [00:35:52] On it, so it's still there,

Harpreet: [00:35:54] Still.

Russell: [00:35:55] There are still some competing companies who or [00:36:00] researchers who are trying

Harpreet: [00:36:01] To find what is the best

Russell: [00:36:02] Way of designing that. Once you arrive at the best way to design it, then you're going to have to focus on efficiency. How to how do you compact it into a smaller place, right? So I like to imagine it as one day you will have the same thing that happened to transistors in the classical computer. Hopefully in the far future is the same thing will happen right where you can design it in a small chip that will be stable. And then the better we get into it, the smaller they will get until you can no longer design it. That's more right. There's the size that you reach that is so small that is going to cause

Harpreet: [00:36:47] Instability into the

Russell: [00:36:48] Design. You will arrive at that, too in the in the far future for quantum. So long story short, right now, what's happening is research into the hardware piece that is still ongoing, but also you have software that can simulate what a good quantum computer can be. Right. So right now, researchers are

Harpreet: [00:37:14] Hopping on these simulations.

Russell: [00:37:17] So you have brackets that can offer you some

Harpreet: [00:37:23] System

Russell: [00:37:24] Softwares to kind of simulate what a quantum computer can do. So you can start beefing up our research and things like that.

Harpreet: [00:37:32] But there isn't a, you

Russell: [00:37:34] Know, known hardware that is stable out there. I know Google, for example, the other day, they claimed that they have a quantum data center, right? So you know, to me, it's just a research center that's looking for the best way to design the hardware that will give you stability. Will we ever have

Harpreet: [00:37:59] A phone [00:38:00] that is

Russell: [00:38:01] Beefed up with a quantum chip? Maybe.

Harpreet: [00:38:04] But to me,

Russell: [00:38:05] It's in the far future.

Harpreet: [00:38:07] Yeah. All that noise

Harpreet: [00:38:09] With those electrons in different multiple universes and stuff like that, right?

Harpreet: [00:38:13] Go, go on, Andrew. Go for it. Yeah. Just to add

Andrew: [00:38:15] On to that, you can't get one at Best

Harpreet: [00:38:17] Buy, but if you've got

Andrew: [00:38:19] A billion dollars or so and you want to go up to Canada and talk to D-Wave, they they might be able to hook you up.

Harpreet: [00:38:25] And it's interesting

Andrew: [00:38:26] Because for a project, probably about a year ago, I'll put a link in the chat. You can actually go to IBM and sign up and play around with one of their quantum quantum computer simulators and how you would have to adjust a program to work on a quantum computer. I'll put that in the chat now. But as Greg was mentioning, you know, you have to keep these tremendously cool. It's like almost at zero at actual zero Kelvin. So the interesting thing about it is that we're trying to exploit that quantum phenomena at that temperature, while at the same time on the silicon processors as they get smaller, it's the quantum properties at that level which are actually getting in the way of getting even smaller. So there's a little bit of an interesting irony there and that

Harpreet: [00:39:12] We have, you know, you can

Andrew: [00:39:14] Stabilize a few qubits, but like not enough to meaningfully do anything right. And so I think there's I don't know, there was some there was some study out recently. And to Greg's point, it was like mentioning like you had to get over a certain number of stable qubits that didn't have a certain level of error versus to actually be useful.

Harpreet: [00:39:36] Right.

Andrew: [00:39:37] So it's just really fascinating. But the IBM quantum computer playing around with that and understanding how the qubits work is really, really fun.

Harpreet: [00:39:45] I mean, the link you sent, though, is that summary of 12 AI after scenarios from the future of life. So might be a

Harpreet: [00:39:52] A different link, Greg. Go for it,

Harpreet: [00:39:54] Go for it. Definitely respond to that. But somebody, please tell me what the hell's a qubit as a different from regular bit. Break that down for me [00:40:00]

Russell: [00:40:00] Qubit as a regular versus a regular bit again. All right. If I were to explain to the child, Greg,

Harpreet: [00:40:11] What is the purpose

Russell: [00:40:12] Of quantum

Harpreet: [00:40:13] Computing is just

Russell: [00:40:15] To

Harpreet: [00:40:15] Accelerate adult Harp

Austin: [00:40:17] Adult Harp?

Russell: [00:40:20] What is that purpose? Well, what are we trying to achieve here? Simply said to me is we live in the universe of information and possibilities,

Harpreet: [00:40:34] And we lack the

Russell: [00:40:36] Means to compute

Harpreet: [00:40:38] These possibilities and understand

Russell: [00:40:41] Them in a in a fast way. If you think about time, right, we may have

Harpreet: [00:40:45] Some tools to perform

Russell: [00:40:47] These calculations. But it takes forever, right, who's going to live ten thousand years to wait for the response of a classical computer in a very complex problem, right? So they're trying to shorten that time

Harpreet: [00:41:00] To kind of evaluate these

Russell: [00:41:02] Outputs to

Harpreet: [00:41:03] Shorten that time. Now Cubitt

Russell: [00:41:05] Versus bit. Right? A bit can be somewhat like turning the light on and off.

Harpreet: [00:41:14] Right on

Russell: [00:41:15] Is one of is zero a

Harpreet: [00:41:19] A cubit

Russell: [00:41:21] Can be both one and zero and you can influence where you want the value of that qubit to be by manipulating its angles. So a qubit is more of a sphere where a bit is more of a lie like this, if it's pointing up is it's light, it's on. If it's pointing down, the light is the that bit is zero, but the qubit can

Harpreet: [00:41:52] Be both up or down

Russell: [00:41:54] And you can influence it by tweaking where [00:42:00] you want that point, that thing to point. So if you if you see my finger here and I'm moving inside of a sphere, so I'm doing things to that ion or that photon or that qubit to influence where I wanted to land. So if you're not measuring it, if you're not really looking at it, it could be either one or it could be either zero. If you're not looking at it, you do something to it and then you measure it. You may end up here, you may end up here, you may go here. So all of these angles can be measured by exactly like a matrix, right? So you can read it as part zero. Part one. Regards to its angles, and then I'll let you explain that a little bit better than I can, if you, if you will. But the way I understand it is, you can manipulate it

Harpreet: [00:42:50] To make it adopt

Russell: [00:42:52] Certain angles to arrive at a certain results.

Harpreet: [00:42:56] Say, for example, you want

Russell: [00:42:58] To achieve twenty four twenty five percent probability x plus, uh, seventy five percent probability why you tweak it at a certain angle to arrive at some somewhere between zero and one.

Harpreet: [00:43:15] Right.

Russell: [00:43:16] So therefore, if you think about a system where you have one qubit, you have already

Harpreet: [00:43:28] Zero

Russell: [00:43:29] Zero zero one possibilities. If you have two qubits, then you increase your possibilities by

Harpreet: [00:43:35] Two race to the end

Russell: [00:43:37] And being how

Harpreet: [00:43:38] Many qubits you put

Russell: [00:43:39] Together. So your possibilities increase in terms of computation capability. Ok. But hopefully,

Harpreet: [00:43:48] Well, that's like when you're talking about, you know, you see something that it's either one or there. That's like the uncertainty principle with some of that observer effect, right?

Harpreet: [00:43:57] Like that that

Harpreet: [00:43:58] That kind of thing going on.

Andrew: [00:43:59] So [00:44:00] there's a part of this where they're talking about like the collapse of the wave function where, yeah, yeah, there are some fun algorithms that you can experiment with that. And I know some game designers have used that to kind of develop the procedurally generated worlds and everything.

Harpreet: [00:44:16] So that's a great description.

Andrew: [00:44:18] Fantastic.

Harpreet: [00:44:19] So the philosopher who wants

Harpreet: [00:44:21] To know, is there

Harpreet: [00:44:22] Such a thing as quantum logic? Because isn't? I information in terms of bits like you're building logic gates, so I guess other quantum gates now and then quantum logic, does that make sense? And there's a couple of questions coming in from LinkedIn that I'll get to from from Khost. But yeah, quantum logic. Is that a thing like, I'd love to learn more about that or quantum logic gates.

Russell: [00:44:46] So quantum logic gates, I don't remember visiting that subject too much, but I went out here, gates. To me, a gate is simply an instruction that you give to the qubit to perform. It's kind of like a mathematical function

Harpreet: [00:45:06] Or a rotation

Russell: [00:45:08] Rotational instruction that you provide to that set up qubit to perform right. So you can say if you started at that top angle here, once you arrive at the gate, I want you to do a flip, you know, towards the x axis and rotate your angle by forty five degrees or something like that. That's what a gate does. And those are the mechanisms that the quantum system uses to manipulate the qubit to arrive at a certain result. So really, you know, when I think about it by expanding it this this way, you can think about that. This is in fact a logic that you use because you can design it by different gates until you arrive at the final measurement that you where you can

Harpreet: [00:45:56] Check if you

Russell: [00:45:58] Arrive at that result. So [00:46:00] if you think

Harpreet: [00:46:00] About a search

Russell: [00:46:02] Algorithm, so that's why you have these. Grover search, for example, has claimed that they can achieve their search logic faster than the classical way because of the way they design their logical

Harpreet: [00:46:20] Gates to arrive

Russell: [00:46:22] At that at that point. So it's quite interesting how they prove how faster they can arrive at that expected value when you search

Harpreet: [00:46:37] Into large

Russell: [00:46:39] Amount of databases

Harpreet: [00:46:40] Or, you know,

Russell: [00:46:42] Numbers and Google Voice searches. One thing that you can take a look at to find out more. But long story short, I do believe that those gains are simply instruction set of instructions that the

Harpreet: [00:47:00] Qubits go through. And if you can

Russell: [00:47:01] Think about it, these gates are very

Harpreet: [00:47:05] Linear where you have the initial state

Russell: [00:47:08] Of these qubits and typically there are parallel to each other and they move in parallel together, right? And each time you can arrive at a gate right next to each other, so each time they pass a gate, they do something and they can be linked to each other. These qubits through a phenomenon called entanglement. And I think somebody put it in there. If you want to find out what entanglement means and as they go through these gates, they understand what they need to do based on the logic placed inside of these gates until they arrive at a final result. So and that's why you have different things like quantum key distribution where you have what you call. [00:48:00] The same thing that the encryption technology use. You have quantum teleportation where they can have sending a qubit to

Harpreet: [00:48:13] Another entity that is far

Russell: [00:48:16] Away from the other person sending that that qubit. And you have other business use cases that you can leverage qubits for.

Harpreet: [00:48:28] Thank you very much, Greg. Appreciate that. Yeah, that's that super fascinating. I looked it up on Google. Quantum logic is a thing as an entire Wikipedia page,

Andrew: [00:48:36] And I'm just going to,

Harpreet: [00:48:37] In the chat, put a

Andrew: [00:48:39] Link to another resource on like the building

Harpreet: [00:48:42] Of quantum

Andrew: [00:48:44] Logic gates. There's a little actually a really fun app. You can play around to see how they work. I'm going to put a link to that. Also in the chat.

Harpreet: [00:48:51] That'd be great. Thank you. A question coming in from LinkedIn that's kind of connected to what we're talking about here from coast up.

Harpreet: [00:48:57] One of the big leaps

Harpreet: [00:48:58] Along with GPU computation, was its implications on energy use when we're evolving into quantum

Harpreet: [00:49:05] Computing for you and

Harpreet: [00:49:07] Ends, which I assume means quantum neural networks,

Russell: [00:49:10] Neural networks

Harpreet: [00:49:11] And the power they unlock. Do we have some projection on the power implications on quantum processing? Here's the

Russell: [00:49:19] Question. So the way I understand it, the power computation for KUNM like the power of computation, like how it will be achieved or is it? Is it better,

Harpreet: [00:49:33] I guess, that the power implications on quantum processing, because I guess GPU was they take a lot of energy, right? Gpus take take a lot of energy to power. Do we have that same kind of issue with quantum computers?

Harpreet: [00:49:46] I guess the

Russell: [00:49:46] Question is we don't know, simply because most of these things are just simulated on classically built computers. Write these classical

Harpreet: [00:49:58] Computers with designed [00:50:00]

Russell: [00:50:01] Quantum systems on them to perform

Harpreet: [00:50:04] Calculations

Russell: [00:50:06] For a. And so I don't think we're there yet to understand what would be the energy consumption of these, you know, systems. But my guess is that if I think about the hardware that it takes, like if you can Google right now, the the computer that Google used to achieve quantum supremacy, I mean, these things are huge taking like a whole room. And I can tell you that it looks like it consumes a lot of power. So like any technology when it's designed, it's not designed at its optimal state. So maybe it will consume a lot of power at the beginning. And then this will be one of the key things that will be improved over time when researchers find more stability inside of the design. That's my take. I mean, I'm not too sure.

Harpreet: [00:51:00] Yeah, it's interesting to think about, right? Like you mentioned right now, the quantum computer Google used to achieve quantum supremacy took up a whole room to claim. Yes, right? They claim people the whole room. Imagine back in the days how big a room it

Harpreet: [00:51:13] Would take

Harpreet: [00:51:14] To power this thing and how that fits just

Harpreet: [00:51:17] Right here. It's thin.

Harpreet: [00:51:18] It's like a quarter of an inch thick. I think that's an overestimation, right? And in how many years, 30, 40 years, right? So who knows? Thirty four years

Harpreet: [00:51:27] In the future, man, what wildness

Harpreet: [00:51:28] Will be up to? I think it's a super fascinating to think about. Yeah, yeah, man. Don't see any of the questions coming in through the chat,

Harpreet: [00:51:37] Which is it's all good. I mean, we had this conversation.

Harpreet: [00:51:40] We're talking about some fascinating topics, man. That's what the artist Data science is all about.

Russell: [00:51:46] I like the questions because it kind of tested whether I was paying attention in class or something like that.

Harpreet: [00:51:51] But so talk to us about the program you're working on. Is this like a graduate program in quantum computing? What's what's this all about?

Russell: [00:51:59] Yeah. So I participated [00:52:00] in a program sponsored by IBM and I've used I mentioned eight years bracket as one of the simulators of quantum systems. There's also a test kit, which is the system that I've used, and you can leverage it as an environment to bring in Python notebooks there and code using Python.

Harpreet: [00:52:18] And they have

Russell: [00:52:19] A quite well documented package for quantum system, which then. And where you can learn how to design systems using Python. So I've studied that or practiced it for a little bit, and it was a two semester program.

Harpreet: [00:52:41] And then the

Russell: [00:52:41] Third one was like a crammed one over the summer, which was an introduction to quantum machine learning. I promise I will do a better job at explaining how quantum machine learning works, but it's quite fairly new at its infancy. But there is. So a lot of promise there. And we've performed some exercises

Harpreet: [00:53:06] And you pass the

Russell: [00:53:07] Program once you reach a certain level

Harpreet: [00:53:11] And inside

Russell: [00:53:12] Of your labs or exams that you take and things like that. With that, the key is to continue to read and practice and learn more and things like that. One of the things I'm interested in if anybody wants to take a look, there is

Harpreet: [00:53:29] The package

Russell: [00:53:30] Python package called Penny Lane. Penny Lane is quite fascinating company that is offering a package for building quantum machine learning systems. So this one year is probably one that I will try or make an effort

Harpreet: [00:53:50] To visit

Russell: [00:53:51] To better understand how quantum machine learning can offer an advantage over classical machine learning, or how to [00:54:00] build hybrid systems where both work together to optimize, you

Harpreet: [00:54:06] Know, a system. This is great. I can look at the website, got got, you know, you can learn, got tutorials and all that stuff.

Harpreet: [00:54:12] Definitely. I'll look

Harpreet: [00:54:13] Into this.

Harpreet: [00:54:14] Probably see if

Harpreet: [00:54:15] There's anybody that can find on

Harpreet: [00:54:16] Linkedin that can reach out to

Harpreet: [00:54:18] Bring them onto the podcast, because I would love to talk about this a little bit more. So if you wouldn't mind break, like,

Harpreet: [00:54:22] Just let me know some some of the

Harpreet: [00:54:23] Textbooks you're using so I can look at the authors and look them up and see if they'd be interested in coming out of the show. I think that'd be super, super interesting.

Harpreet: [00:54:32] Just, you know,

Harpreet: [00:54:33] I'd like to expand my intellectual horizon. I like being interviewed and interviewer ask them questions because it just it. It's the best feeling to feel dumb in a situation and walk out like, Oh my God, if I learned so much more. That's awesome. Absolutely. Yeah. So it doesn't look like there's any more questions. But real quick I we want to give a shout out

Harpreet: [00:54:54] To and I've had a lot of

Harpreet: [00:54:55] Help from the community in reviewing this course that I'm launching. Shout out to some

Harpreet: [00:55:01] Of the community

Harpreet: [00:55:02] Members who have been reviewing it and the leaders in the industry, such as Tom Greg Vyn, Mark, who have been reviewing the course that I'm creating, you know, putting putting a lot of work into this and I'm hoping it's going to provide benefit to people. You know, the reception so far has been positive, so hopefully, you know, hopefully I can deliver something of value for you guys. Actually, there's another question

Harpreet: [00:55:27] Coming in as about

Harpreet: [00:55:28] Shut it down, but Kosta is asking the question Does the nature of the computations inform a

Harpreet: [00:55:35] Completely different way

Harpreet: [00:55:36] To approach Data in general? Or is it the

Harpreet: [00:55:40] Same Data

Harpreet: [00:55:41] But different processing style?

Harpreet: [00:55:44] We're still figuring out what Ml

Harpreet: [00:55:46] Ups looks like. I'm wondering what that looks like with quantum involved. The deep question, man. Very interesting. What do you think you guys think?

Russell: [00:55:57] Could you could you read that again?

Harpreet: [00:55:59] Yeah. [00:56:00] I'll also post

Harpreet: [00:56:01] Right here into the chat. Does the

Harpreet: [00:56:02] Nature of computations inform a

Harpreet: [00:56:06] Completely different

Harpreet: [00:56:07] Way to approach Data in general, or is it the same

Harpreet: [00:56:14] Data but different

Harpreet: [00:56:16] Processing style? So like quantum

Harpreet: [00:56:18] Quantum Data, right?

Harpreet: [00:56:20] We're still trying to figure out what ml obstacles like.

Harpreet: [00:56:23] I'm wondering what that

Harpreet: [00:56:24] Looks like with quantum involved a heavy question.

Andrew: [00:56:27] Yeah, that's that's really interesting because it seems like some of the

Harpreet: [00:56:31] Discussions I've

Andrew: [00:56:35] Been hearing in this world are about really like translating some of our traditional data types and traditional programs into something that can then run on a quantum computer. So it's interesting to think about it. The Data is not quantum native

Harpreet: [00:56:51] In a way.

Andrew: [00:56:52] And so what might that look like? And when we might get there

Harpreet: [00:56:57] And if we ever will, because it might not make

Andrew: [00:57:00] Sense. I hear a lot about like potential futures of hybrid, like large quantum computers elsewhere, potentially available via the cloud, but like you're still mobile

Harpreet: [00:57:09] And whatnot because of the tremendous cost

Andrew: [00:57:11] In space. Maybe not now, maybe in a hundred and 200 years, but in 50 years. It still makes sense to keep like traditional silicon processors on these sorts of things and just tie in to the quantum network when you need that computing power. But that's a really interesting question. Like when could we get to quantum native Data?

Russell: [00:57:30] Mm hmm. Yeah, I think I think this also when I was following the quantum machine learning program, one of the limitations or difficulties was to take classical Data or what I call classical. Data is kind of like a structured Data with with columns and rows and transform each of these columns into a qubit. That transformation is [00:58:00] quite. Difficult and you need practice to, and according to my learnings, too, is that the more you add columns, it skills with the number of qubits as well. So designing a system with a large number of qubits may be a little bit paralyzing or not. Not the best to manage. So there's one thing that's that's one thing to to to observe a little bit

Harpreet: [00:58:31] More in terms of

Russell: [00:58:32] How to make it a little bit easier. So if you think about the design of a system that is physical, like a physical unit and it it contains a certain amount of qubits in

Harpreet: [00:58:52] There, you may have

Russell: [00:58:54] Limitations when it comes to processing large amount of data that may require an excessive amount of qubits to perform the computation that you need. So you're going to be limited there. Where you can have a little bit more flexibility is if you can run that classical Data into a system that can convert this classical data into quantum features that can scale automatically based on software or things like that. But on the physical side, I think we're going to be very limited.

Harpreet: [00:59:33] Question Paul question from coast up here for context to the previous question. For example, testing and machine learning model monitoring needs. Version Data state and a versioned model state. Is there another layer involved?

Harpreet: [00:59:48] Now we're talking about

Harpreet: [00:59:49] 15

Harpreet: [00:59:50] Datasets and 15

Harpreet: [00:59:51] Models, so would mean that's like quantum and NFTs binder. But I'm just being facetious. But but I'm yeah, it's [01:00:00] a good question. What do you? I think it's OK to a pet to pass on

Harpreet: [01:00:04] It because these are deep questions,

Russell: [01:00:06] But these are questions that need a little bit.

Harpreet: [01:00:08] Yeah, yeah, definitely. These are questions that when I get somebody on the show

Harpreet: [01:00:12] That's an expert

Harpreet: [01:00:13] In quantum

Harpreet: [01:00:14] Machine learning, I was

Harpreet: [01:00:15] Going to ask these questions because they're very good. Thank you. Because the yeah, don't know, don't know Andrew Andrew's like, yeah, nope.

Harpreet: [01:00:23] Either. Great.

Harpreet: [01:00:25] Great question.

Andrew: [01:00:26] Nfts are gone for the day.

Harpreet: [01:00:31] Yeah, man. So this is this is just a testament to just, you know, the future is crazy. Things are going to happen that we cannot even anticipate, you know, twenty years ago, if people would have been

Harpreet: [01:00:41] Talking about all of us

Harpreet: [01:00:43] Sitting together in rooms during a pandemic, talking about stuff nobody would have believed,

Harpreet: [01:00:49] As you know, we're talking about

Harpreet: [01:00:51] Quantum computing, NFTs and things like that as well.

Harpreet: [01:00:54] Stay curious and open minded. A little

Russell: [01:00:57] Curious man. Yeah. Another layer involved is, is the person talking about another layer in the quantum world? Right? I'm assuming. Is there another layer involved? Can the person elaborate a little bit on that? Is there another layer involved?

Harpreet: [01:01:15] Sounds to me like the leaning of the question is

Harpreet: [01:01:18] In essence,

Harpreet: [01:01:19] Quantum ml ops related type of thing, right? Like, do we need to? Because, you know, if we have traditional computers, right, we don't. There's the bits aren't probabilistic, they just compute, right? So if a quantum computer was to fit a model, the bits are in a probabilistic state.

Harpreet: [01:01:37] Do we need to

Harpreet: [01:01:37] Version control how those bits behaved when we train a model and fit a model? Does that does that? I'm thinking that is what this person is going to, because that's

Harpreet: [01:01:47] Kind of what bouncing around in

Harpreet: [01:01:48] My mind,

Harpreet: [01:01:50] But I'm not entirely sure. Does that make sense

Harpreet: [01:01:52] Like Kent can? Does it? We have to version control

Harpreet: [01:01:56] You at state when we are

Harpreet: [01:01:58] Trying to reproduce a

Harpreet: [01:01:59] Quantum machine [01:02:00] learning model.

Russell: [01:02:01] Yeah, I think once you reach a certain level of stability. The version control is probably going to

Harpreet: [01:02:13] Happen in

Russell: [01:02:15] A classical world where you can register the angles that each qubit arrived at to arrive

Harpreet: [01:02:24] At the probability

Russell: [01:02:26] Numbers. But that's why I think people are designing new term devices that are combining the power of a quantum computing

Harpreet: [01:02:38] With a

Russell: [01:02:40] Classical computing

Harpreet: [01:02:42] Like the example I took before.

Russell: [01:02:44] So you're twisting and turning the qubits into certain angles until they arrive at a measurement that is properly arriving at the expected output. Think about your searching for that needle in the haystack,

Harpreet: [01:03:05] So you tweak it and

Russell: [01:03:07] Find out probability at 80 percent, for example, in your first iteration. Then you download these Engel's of these qubits and pass it to a classical computer, which in turns performs some sort of weight adjustment through back propagation, for example, and then passes those weights back to the quantum device to perform. Now with these

Harpreet: [01:03:35] Weights, you have

Russell: [01:03:37] Ok, now you're instructing this quantum system to change the angles of the gates or the instructions to increase

Harpreet: [01:03:45] That 80 percent

Russell: [01:03:47] Probability in the output until you reach a certain version that you're satisfied with in that version can now be, I guess, stored in the classical world and [01:04:00] then move forward with that. Right. So that's the way I understood it. Not sure if

Harpreet: [01:04:04] I've inserted the proper

Austin: [01:04:05] Way.

Andrew: [01:04:06] Yeah, that sounds right, because you would use the quantum computing to explore potential states.

Russell: [01:04:11] Yes, right?

Harpreet: [01:04:12] Quickly, which is you can't do in classical

Andrew: [01:04:16] Computing, you know, certain issues, right? Like, of course, like the most obvious use of quantum computing is for cryptography, right? So like RSA, if you're factoring primes right, the quantum computing can do what would take a classical computer a million years.

Harpreet: [01:04:32] And, you know, I don't know, like

Andrew: [01:04:33] Eight seconds and pulling it up my ass, but like something short like that. So you explore those possible states. But then once you find it like once you find the key, you would store that in a classical computer. Same thing with drug discovery, right? You're exploring all the permutations of the features of the molecular structure. You find the right one.

Harpreet: [01:04:51] You save it around. Great questions. Great questions, then Coast

Harpreet: [01:04:55] Up is happy with with the responses. And thanks you guys so much for just being honest with the with the unknown. Mark, now what's going on to get to have you here? Also shout out to people who haven't heard from in a while, talk, etc. You're probably sitting in some beautiful bar in the south of France. Serge A. What's going on, my friends? Yeah. Mark Mark missed. Carlos, yeah, we had Carlos back after man. I don't know how long. It's true, man, Carlos

Harpreet: [01:05:23] Is one of the OG ones, always

Harpreet: [01:05:25] At office hours from the get go. And this is officer number forty nine, which means in three weeks I'll be one year of doing this thing. Man, that is a that is insane to me. One year straight of Data sized happy hours, Markman as a good man. How's your week been?

Austin: [01:05:42] It's been. It's been going great, going great. Sorrows late, a little bit. Like I said earlier, I don't know for this one or the Sunday one, I'm working on a podcast. And so inspired by

Harpreet: [01:05:53] You, for sure.

Austin: [01:05:55] But me and my friends had like a whole like strategy session of like, what is this podcast mean? [01:06:00] And they went over. But it was like super fun. It gave me a lot of respect for the work you do.

Harpreet: [01:06:05] I'm excited to hear, man, I'm excited to tune in when this thing comes out.

Harpreet: [01:06:10] You want

Harpreet: [01:06:10] To share any more detail

Harpreet: [01:06:11] With us, like what it's

Harpreet: [01:06:12] Called, when we can expect to see it.

Austin: [01:06:15] We're we're still trying to come up with the name, but essentially the idea is like it's not even Data related. It's like an outlet outside

Harpreet: [01:06:21] Of Data, but essentially it's like

Austin: [01:06:23] Modeling vulnerable conversations and you know, how

Harpreet: [01:06:27] To share various stories

Austin: [01:06:29] From different backgrounds and having these really great conversations recorded

Harpreet: [01:06:35] Conversations. Is that a

Harpreet: [01:06:36] Name of a podcast, or is that the actual name of yours was about to be?

Austin: [01:06:40] It's not even a name. It's like, literally, we want to have bonorable conversations.

Harpreet: [01:06:44] Got you. You know,

Austin: [01:06:45] Ideally how I was like called dear past self

Harpreet: [01:06:47] Where like, you know or lessons,

Austin: [01:06:49] You'll tell your old self. We don't like the name too much right now, but essentially it's just like trying to think through like, what's what's a brand? What's a story? How we launch, where our goals are. Super fun stuff.

Harpreet: [01:07:02] I'm excited and excited to see and also giving a shout out to everybody who was checking

Harpreet: [01:07:05] Out my my course doing a review for it. Mark is one of those folks who so generously

Harpreet: [01:07:12] Putting in his time

Harpreet: [01:07:13] Marc Gregg,

Harpreet: [01:07:15] Then Vince giving me such incredible feedback. You know what he has to know. You can keep it completely honest with me if my course sucks. Please let me know what I could do to improve it because I'm trying to make this thing be the

Harpreet: [01:07:28] Best it

Harpreet: [01:07:29] Could possibly be. Russell is asking, Are we going for Hawaiian shirts to celebrate? Happy hour in three weeks and Winnipeg? It is. It's already like low sixties now, so I'm about to be switching to cashmere

Harpreet: [01:07:41] Sweaters in

Harpreet: [01:07:42] A week or two before I go into a full on sweatshirts. Shout out to Joe Reese and Matt Housley

Harpreet: [01:07:48] In the house. Very happy to have you guys here. What's going on, man?

Austin: [01:07:51] All right for the tardy.

Harpreet: [01:07:53] Not so good, man.

Harpreet: [01:07:54] You miss

Harpreet: [01:07:55] Out on a very interesting conversation on quantum computing, I

Austin: [01:07:57] Wonder. Oh, nice. And I have to go back and watch [01:08:00] that.

Harpreet: [01:08:01] Oh, and there are some, some interesting, interesting talks. And then also with NFTs and things like that, I'll open it up for four other questions. I don't see any questions coming in any questions or comments. Anybody want to chime in?

Harpreet: [01:08:12] Please do so.

Carlos: [01:08:14] Do you have any thoughts on quantum computing?

Harpreet: [01:08:16] Yeah, this is here, man.

Harpreet: [01:08:18] He's not really interested about that.

Austin: [01:08:20] Yeah, I mean, I know that in the late nineties, there were a lot of questions about whether coherent enough coherence would ever be achieved to actually get something useful out of quantum computing. I mean, there's been so much progress since then that it seems plausible that we'll see something useful where the practical limits are. I don't know. I mean, we'll we'll coherence ultimately be a limit where RSA algorithms will continue to work. We just LinkedIn the the factors and we'll be OK. Or will that be something that that type of cryptography will be completely broken? I think those are open questions

Harpreet: [01:08:55] At this point. But yeah,

Austin: [01:08:56] Over 20 years, a lot of things have changed in terms of the general outlook.

Harpreet: [01:09:01] I think that would be the most immediate use case for for quantum computing

Harpreet: [01:09:04] Is in the realm of cryptography.

Austin: [01:09:06] It seems to be the big motivation. I know that a lot of other algorithms have been developed to apply to other areas, but I fully follow that conversation.

Russell: [01:09:15] So you know what? What I think is interesting that I that I that I'm

Harpreet: [01:09:21] Interested to see is

Russell: [01:09:23] Nowadays,

Harpreet: [01:09:24] When it comes

Russell: [01:09:25] To cryptography, the world of cryptography is kind of losing their mind

Harpreet: [01:09:32] Right now because

Russell: [01:09:33] Of quantum computing making advancement, especially with quantum key distribution. And they're now they're thinking, OK, he or her holds the power to quantum computers can now crack the code, the hash code of crypto. And now I think there are movements in the crypto world where they're trying to design, you know, the hash that [01:10:00] is non breakable by quantum computing.

Harpreet: [01:10:02] And I ask my

Russell: [01:10:03] Question, Well, why

Harpreet: [01:10:06] Not leverage

Russell: [01:10:07] Quantum computers to design the hash code of crypto? In this case, if the quantum computer designs that hash code, then quantum computer won't be able to

Harpreet: [01:10:18] Break that hash

Russell: [01:10:19] Code, right? So it's quite an interesting thing to do. And we're talking about long key chains of numbers like up in the two hundred five hundred long, so I'm interested to see what's going to happen there. But for now, there's a

Harpreet: [01:10:39] People are scared that, you know, if if

Russell: [01:10:41] You're China, you have computation power through quantum computing,

Harpreet: [01:10:46] Then

Russell: [01:10:47] Any cryptos can be hacked by this very powerful computation power. So I'm I'm curious to see how things will happen.

Andrew: [01:10:57] Yeah. Just to

Harpreet: [01:10:58] Add onto that, I think there

Andrew: [01:11:00] Is a lot of impetus from the national security space just because there is this computational arms race.

Harpreet: [01:11:09] And I think that

Andrew: [01:11:09] That that's driving a lot of the discussion on quantum cryptography outside outside the commercial space and then coming back to what Greg just mentioned

Harpreet: [01:11:19] On using

Andrew: [01:11:20] Quantum computing to create these keys. And that kind of ties a little bit back into our discussion

Harpreet: [01:11:26] On how Data can be

Andrew: [01:11:27] Fundamentally different

Harpreet: [01:11:28] Because at that point, you're

Andrew: [01:11:30] Not working with RSA where you're trying to factor your primes. You can actually, if you're transmitting a key and someone intercepts it, which in a traditional with the traditional key, you would not know. But because of the nature of the quantum communication of the of the key,

Harpreet: [01:11:46] It would be interacted with.

Andrew: [01:11:49] And if a quantum key is interacted with in transmission, it disentangled. And therefore you

Harpreet: [01:11:54] Would know when you receive it

Andrew: [01:11:56] If it has been tampered with at all. And so there's [01:12:00] a lot of. Very interesting use cases there. And that's fundamentally a very different way that the Data operates.

Harpreet: [01:12:07] I'd love to get Matt's perspective on this

Harpreet: [01:12:10] Question that

Harpreet: [01:12:11] We're discussing a little bit earlier because it's a really interesting question that came in from LinkedIn.

Harpreet: [01:12:15] There's the nature of

Harpreet: [01:12:17] Computation in form,

Harpreet: [01:12:19] A completely different approach to

Harpreet: [01:12:21] Data in general. Or is it the same

Harpreet: [01:12:25] Data but different

Harpreet: [01:12:27] Processing style? We're still figuring out what Ml Ops looks like. I'm wondering what that looks like

Harpreet: [01:12:33] With quantum involved. Some more

Harpreet: [01:12:35] Context is that testing and ml model monitoring needs a versioned

Harpreet: [01:12:41] Data state and a versioned

Harpreet: [01:12:43] Model state. Is there something else another layer involved once you start talking about quantum computing?

Matt: [01:12:50] Yeah, I mean, from my perspective and I'm not an expert in this domain, let's be clear, but the real challenge with quantum computing is that you're you're so far outside the realm of procedural

Harpreet: [01:13:00] Thinking, the way

Austin: [01:13:01] We have been thinking about computation for the last like, you know, since the beginning of computation, at least since the 19th century, really. And so you fundamentally have to think rethink what you can do with these algorithms. Quantum computing can be very, very powerful, but you're very, very restricted in terms of the things you can actually do. So you have to figure out how to, like, restate your problems in a fundamentally new way. And so I think that's the big challenge. And then the question is, can you then apply these new types of algorithms in these new types of processes to machine learning in a way that's going to be valuable in a way that solves problems faster than in a way that fundamentally gets work done

Harpreet: [01:13:37] A lot faster, like it does

Austin: [01:13:39] With factoring primes, for example? Yeah.

Harpreet: [01:13:43] For those guys interested more in just like

Harpreet: [01:13:45] The theory of computation, I highly recommend checking out stuff from Stephen Wolfram. He's been to Lex Friedman's podcast a couple of times.

Harpreet: [01:13:52] Just talk about the theory of computation.

Harpreet: [01:13:54] Really interesting stuff. I definitely encourage everyone to check it out. You mentioned coherence,

Harpreet: [01:13:58] And

Harpreet: [01:14:00] Russell [01:14:00] also says that apparently coherence is a big

Harpreet: [01:14:03] Thing in

Harpreet: [01:14:04] Quantum computing. So what? Oh, here's what the heck does that mean?

Harpreet: [01:14:08] I'm talking about

Harpreet: [01:14:09] That. Somebody, anybody, I go for it.

Austin: [01:14:12] I so so fundamentally, what you have to do is create these entangled, really complex, entangled states

Harpreet: [01:14:18] And

Austin: [01:14:20] As shows up in other places of of where quantum mechanics becomes very

Harpreet: [01:14:24] Important, like superfluid helium,

Austin: [01:14:26] For example, any kind of disturbance of that system, so like thermal disturbance can mess up that coherence. And as I understand it, this is something that's already taken into account in these algorithms to some extent. So there's a degree to which I'm not an expert on this either, but you could sort of resample and do the experiment many times and then get a reasonable answer out of it as long as the states

Harpreet: [01:14:47] Aren't too disturbed. But fundamentally,

Austin: [01:14:49] Like your

Harpreet: [01:14:49] Bit, length is limited

Austin: [01:14:51] By your ability to maintain these coherent states in the system. And that's the key challenge. And so I think, like physicists and researchers have pushed this much further than maybe people thought would be possible 20 years ago. But they still have a long, long way to go to maintaining these entangled states over so many bits.

Russell: [01:15:08] Yeah, if I can if I can explain that in, you know, too little, Greg, it's good to understand what entanglement means as well. Right. So for me, my interpretation of entanglement is

Harpreet: [01:15:20] You take two

Russell: [01:15:22] Electrons and you have this initial energy state, so you have low energy, high energy and entanglement. For me, it means

Harpreet: [01:15:37] If you want,

Russell: [01:15:39] You know that if electron one has energy level one, you want the other to be zero or you want them to be zero at the same time, you want to be them to be one in one at the same time, or you want it to be flipping one in zero or zero in one kind of thing. So they [01:16:00] kind of win one changes the other, one changes or if you want them to be the same. So entanglement to me is kind of like if you know, the value of one, you know, the value of the other automatically right now. Coherence means as you go through the changes or gates for these electrons, when you push energy to these electrons, they will move certain ways and they will maintain their stability. Decoherence happens when it interacts with the external forces.

Harpreet: [01:16:37] When Noor is

Russell: [01:16:38] Traverses these systems, then that pair that you formed with these two electrons is no

Harpreet: [01:16:45] Longer valid.

Russell: [01:16:47] This way, when you perform your computations or when you perform your gates onto these electrons, then that expected result is no longer achieved. So that's why

Harpreet: [01:17:04] I

Russell: [01:17:04] Would advise you to look into the perfect quantum system is one that meets the diving chancellor's rules. And one of the rules is coherence, where when you perform a calculation

Harpreet: [01:17:23] Through that

Russell: [01:17:25] System that the system is

Harpreet: [01:17:26] Stable and noise

Russell: [01:17:28] Isn't strong enough to break that entanglement that you formed with the with the with the electrons, I call them electrons because it's easier for me to kind of imagine what a

Harpreet: [01:17:38] Qubit is. That's why I'm calling them electrons.

Russell: [01:17:40] It could be ions. There could be other things as well. But in chains, those rules

Harpreet: [01:17:46] Are what

Russell: [01:17:47] Researchers are going through right now to understand the criteria that is that it takes to have a stable quantum system.

Harpreet: [01:17:58] Dan has some interesting stuff. [01:18:00]

Harpreet: [01:18:00] Any anything to add to that? Matt, Andrew, anybody does not.

Austin: [01:18:05] That was a really nice explanation. Thank you, Greg. That's awesome. So, yeah, this looks like a cool resource as well. I haven't read this article, but I'll check that out, too.

Harpreet: [01:18:13] So, yeah, physics is awesome. Physics is pretty phenomenal. I wish I was smart enough to study it. I just kind of observe from the sidelines

Harpreet: [01:18:22] And pretend like I know what's

Harpreet: [01:18:23] Going on. But that awesome

Harpreet: [01:18:25] Conversation.

Harpreet: [01:18:26] Thanks, guys, for tuning in. Appreciate having all you guys here. Don't forget, tune into the conversation I had with Max Frenzel. We touch on quantum and some stuff like that. Quantum information theory

Harpreet: [01:18:38] Is what he is PhD and

Harpreet: [01:18:40] Plus some other physics related topic, so definitely check that conversation out. We also talk about the importance of resting and get to hear a very stressed out Harp share his woes with with Max. We recorded this back in a back in February when I recorded this episode with him during that time when I was going through like I had like nine interviews in one month and like. Festivals held baby wasn't sleeping as though it was like operating on like

Harpreet: [01:19:07] Four hours of sleep. It was madness.

Harpreet: [01:19:11] They were operating on very minimal sleep. Maybe still crazy. But anyways, thank you guys for

Harpreet: [01:19:16] Hanging out again.

Harpreet: [01:19:17] Chatterton Tony for taking care of last week's happy hour. If anybody in the

Harpreet: [01:19:22] Community wants to take the

Harpreet: [01:19:23] Wheels at a happy hour wants to be hosted, let me know, man, I'd be

Harpreet: [01:19:26] Happy to take a

Harpreet: [01:19:28] Break every now and then. So if anybody wants to host, let me know.

Harpreet: [01:19:31] Definitely, you know,

Harpreet: [01:19:33] Have you guys that have that opportunity again? Take care. Shout out to everybody who

Harpreet: [01:19:39] Helped review my

Harpreet: [01:19:40] Course.

Harpreet: [01:19:40] I'm excited to launch it.

Harpreet: [01:19:42] And my friends remember you've got one life on this planet. Why not try do some big cheers, everyone?