Transcript for:
Interview with Daniel Bourke on his Journey through Tech and Machine Learning

programming to me is just thinking like a lot of my time I think if if people get the um the idea of like if you're a software engineer you just spend all day um writing code I would say now majority like I do write a lot of code however uh I use AI tools such as GitHub co-pilot to write a lot of the boiler plate [Music] welcome back to the freec code Camp podcast I'm Quincy Larson teacher and founder of freec camp.org each week we're bringing you Insight from developers Founders and ambitious people getting into Tex this week we're joined by Daniel Burke Daniel is a machine learning engineer he's the creator of many popular tutorials on YouTube and he's a frequent contributor to the Freo Camp YouTube channel Daniel welcome to the podcast hey Quincy it's it's great to be here uh I actually started to learn to code back in the day I think it was about 2017 on free code Camp um I remember going through all the projects and so yeah it's really cool to be here absolutely man it's great to have you here and I just want to emphasize that uh it's an honor to count among our alumni somebody like yourself who not only goes out and practices like kind of Leading Edge uh machine learning techniques but also turns around and teaches other people how to do that that's super chill and I also want to acknowledge that uh a couple years ago you you made a you had a pretty good year on YouTube and you turned around and you donated the proceeds from your YouTube channel to freeo Camp uh to our charity and that was super kind of you and that helped uh that went toward teacher compensation that went toward keeping the 100 servers we have around the world up and running and uh so I just want to publicly thank you for doing that it was my pleasure I've gained a lot from free code Camp so yeah it was uh very very heartwarming for me to be able to give back in some way yeah well let's dive in and learn more about you man like I have really enjoyed watching your videos over the past few years and uh to be completely candid I still don't know that much about your personal life I know how like Ironclad your work ethic is and how you have all these different techniques to be incredibly productive which you you know delve into on your channel but you're based over there in brisban in Australia I hope I'm pronouncing it correctly Brisbane how do you pronounce it uh bris Brisbane so like bris bin okay so it's almost like there's no e on the end like bris bin yeah yeah but Bane is yeah it kind of looks like Bane bris well it's like I I always would pronounce Melbourne and when I went to Melbourne people were like oh it's Melbourne like it's very Australia we like to compress the vows as much as possible yeah yeah that's awesome well I mean Americans we do the same thing right like we have like these words like these French words that we inherited and we just chop them down to like it'll be like seven letters and it'll be like one syllable right so um yeah well let's talk about your childhood out there cuz it sounds like you had a pretty idelic childhood I understand that you live pretty close to where you grew up I do yeah so my uh family home is about a kilometer down the road uh the house I live in now I live in with uh my second brother I have three younger brothers so there's four boys in our family and uh yeah my other two brothers live a kilometer down the street with both my parents and we really love the area um so back a few years ago I was thinking about moving to the US but then that that just didn't happen cuz I found a job in Australia and I'm like you know what the internet's good enough I'm just going to stay here cuz yeah I love it we live really close to uh the sandate foreshore or Waterfront which if you look on Google Maps just a beautiful area the three biggest sand islands are actually just just off the east coast of Australia so we live opposite um three biggest sand islands in the world that is we live opposite Morton Island which is the third biggest sand Island it's a beautiful place we go there for family holidays um every year so um I turned 30 last year what is a sand island is it like a sand bar like literally just sand there's no so imagine just a huge Sand Bar yeah so it's like just pure sand um yeah I mean there is some dirt but it's like most majority sand so if you like dig dig underneath the the grass and whatnot that's sort of made its way there over time through wind and water um you'll just start digging into sand um but yeah it's a huge Sandbar like we're talking I think one side of the island is like 25 km long so maybe 16 Mi or whatever the I'm not sure the the conversion there but it's um yeah it's a it's a great great place um and yeah we go there as that's been our family holiday since we were kids and it's about an hour boat ride offshore and uh you can take your full drive there and drive around all the sand tracks and go fishing and the water's nice and protected and uh and calm so it's almost like a big Harbor SL Lake uh so it's not where where I go for walks in the in the afternoon and evenings is is really really calm it's not like it's a beach but it's not like a a wach walking on like an ocean beach but there's not like a bunch of waves exactly yeah so it's more like a big uh like Lagoon SL Lake yeah that's that's how I would describe it that's really cool going you're not going surfing there how often yeah maybe maybe like kitty surfing like getting your kids ready for like seeing if they can stand on the board uh yeah like like how often do you actually go out to the beach and go for walks almost every day like I was out there yesterday went for a walk with my two young brothers and dogs and we were walking for I think an hour and a half or or so but it's yeah it's a beautiful I try to get out to see the water almost every day and we call it the scenic route when you drive home is like driving driving home we take the long way instead of like going straight to the house we drive drive along the Boulevard and just look out in the big expansive water yeah well Australia has a lot of big cities uh of course like you know Melbourne and uh uh Sydney and you know Brisbane uh like is that like a a major Tech Hub or is it like what is the popular what what is the industry like there well I would say definitely Sydney is the like the tech Hub or like just a business Hub in general and then Melbourne probably follows that as a close second and then brisbane's probably third after that um so if you like if I wanted to be really into it and definitely get a job in the industry I'd probably move to Sydney or Mourne um to work at a place but because of I mean the internet these days like brison's not that far behind and I I've been to Sydney to me that's that's like the New York of Australia it's too it's too hectic for me uh I I quite like to just chill out and uh not sit in traffic for hours at a time and uh Melbourne is beautiful like just a beautiful um conglomerate of different cultures there's almost many cities within a city in Melbourne that's how I perceive it and then Brisbane is kind of almost like I'm biased obviously but brisbane's like The Best of Both Worlds of Sydney and Melbourne so it's got that just enough busy but not too crazy but it's also got it's really starting to now um pop off pop off with with different little pockets of of culture and restaurants and that sort of stuff yeah that sounds really cool I mean that's like kind of the city like I like to live in like I like to live in a city but New York City Los Angeles like a little bit too big and being you know and I've seen like videos you just going out and ride bikes or like those like really bikes what tricle things I don't know what it is a big wheel we call a big wheel yeah yeah dries oh they yeah so they're epic they're like um yeah so little tricycles you you're sitting probably 2 to 3 Ines off the ground um in a bucket seat you got one big wheel on the front and uh then you have two rubber tires on the back and you make them into drift trikes by putting PVC pipe over the tires so there basically on the bitchman there's there's no no grip uh so you can you can move but then as soon as you try to turn your whole back end slides Out and because my family home is on a on a pretty steep hill and pretty long like it's a few hundred meters long um we we'd have we built these drift trikes and then we'd we'd ride down the side street to get speed turn left onto the hill and then just you could just skid out down the whole bottom of the hill you'd have someone down there looking out for when there's cars they give it all clear and then you just shoot down walk back up the hill start again so it's like Initial D for kids almost yeah Initial D you're familiar with the anime the drift racing kid no I have oh wait is that the one with the um that's the where the 86 is that the Toyota 86 yeah yeah okay I know it from the car cuz my friends are really into cars and uh they love their Toyota 86 and so I've kind of I've heard of it from there I went back and rewatched the first season it's pretty pretty exciting like uh yeah is it's cool it still holds up uh but yeah like like the like whenever I play like Mario Kart or something that's the thing I like to power slide right and so the notion that you could actually physically do that how dangerous like do you ever tip over yeah yeah like you'll you'll hit the like you got videos of us like you hit the gutter and you kind of cuz your front wheel's so big it um you hit the gutter and then it goes you like tip over the front cuz your uh um your front reel doesn't Mount The Gutter It just gets stuck like basically luckily there was no uh touchwood hectic accidents so minor crashes skin knees that sort of stuff but nothing too outlandish yeah but that sounds like so fun just like a nice physical activity go out and enjoy the outdoors it sounds like you're quite the Outdoorsman that you've optimized for spending a lot of time Outdoors through your lifestyle 100% yeah I like uh yeah I kind of love the balance between uh being sort of a monk in the mountains and just exploring and doing nothing most of the time and then uh being like a tech nerd and in front of screens all day and concentrating and Building Things and making stuff so it's a yeah I wouldn't say it's 50/50 but it's uh sometimes the one one side builds up too much and then I'm like okay I'm sick of relaxing and being in the mountains I need to go do some computer stuff and then vice versa it's like okay I'm sick of being in front of a computer I need to get outside yeah and and you can just toggle between the two like really easily you got the beach there you've got that Hill that you can you know drift race down well it's no you're so right like it's not far it's about the the uh where I live is probably as 10-minute walk to the to the beach front and so yeah and or bike ride like I go for one of the two to almost every day so it's a very uh yeah it's a very cool place yeah well I'm excited to learn a little bit about like you know they say that like people up in really cold brutal climates like like Russia or Scandinavia they become like the best Chess Masters cuz they're forced to they they have the best authors because they're like forced to stay in all the time and they can't go out and have fun in the sun and you know and yet you're in a very like it sounds like a great place to just relax and enjoy the outdoors and yet you are also very productive staying inside working on your laptop a lot maybe you could talk a little bit about that like how growing up what got you interested in technology when you've got like this great big Outdoors like Australia famous for being like very sparsely populated having this huge kind of Outback area that takes up most of the continent uh like I mean why didn't you end up like uh Steve Irwin for example why why are you why are you more like Yan laon or somebody like that right yeah maybe I'm I'm a crossover between Steve irn and Yan laon I kind of got similar hair to Steve um that's a that's a fantastic question I I basically got into computers as early as I can remember I mean my parents I remember them being at my grandparents house basically talking about the fact that um talking my mom was talking with my grandparents that she was going to buy a computer computer and I was just so interested I didn't actually know what it was and then um uh this is maybe when I was 5 six or seven and then she bought a computer for her business and or our family business that is and it was one of those old school ones like big gray plastic box like monitor was probably I mean if we looked at the size of that monitor these days it would be like people really looked at these for hours on there uh whereas now we're kind of blessed with 30in High defin screens but um yeah I was just fascinated with it and then you just sort of you get into that habit of just as a kid just playing around clicking things going through I remember going through because this is I think almost pre- internet or the internet was like dial up and so it it kind of sucked and then um but I remember just clicking through the menus just being like I can move this thing on the desk and it like relates to the screen I'm actually still fascinated by that whenever I cuz I haven't delved into like how computer hardware works at a really deep level um it still fascinates me that there's analog Parts in computers and like you press a key and it somehow triggers a current and then the letter Q appears on your screen anyway um got into that naturally started to find the internet naturally started to find online games and it just like just went from there my dad started to get uh he was a teacher at the high school I went to and then he started to laptops started to get introduced but he he never learned to use a computer he never learned to use a smartphone um and that's also like really fascinating to me is thinking about cuz now like my life is basically computers and screens but he's lived his whole life without basically without computers and screens so it's like really like fascinating to me like that J position and I feel like that's kind of influenced me a lot into um not just being 100% involved in screen time um but he got a laptop and basically yeah never never really learned to use it so my brothers would like hijack it cuz he'd bring it home from school and we'd get this laptop and because it was much newer than our um computer I remember when we got Wi-Fi at home um that was like early early days of Wi-Fi and I'm pretty sure I may be wrong here I'm fairly sure the the technology for Wi-Fi was invented in Australia may be wrong there but um when my friends would come over they were like what the hell you have Wi-Fi in your house cuz it was like this thing that you would only get at school and so we'd hack um well not hack but I would get my Dad's laptop download a bunch of things back in the day was lime wire uh MSN Messenger and then within probably a month the computer would just be toast because it's just full of malware and that sort of stuff um would you like format it and like just try again well that's the thing like it was just like we would never reset said that it would just go to um the IT department and uh we don't know if it was like getting reset or if they just gave him a new one or whatnot but um he he he legitimately hard they must have hated him every month he brings in like his malware L computer that he doesn't even use exactly like the It people are like doing labor for his crazy kids just surfing the web being Reckless right that's funny oh and then then we kind of got a a bit um I guess you could say entrepreneurial or nefarious I guess it's been enough time you can tell these stories now is that um my dad had the teachers login right and so the teachers login had access to all the student hard drives and so um my friend and I in high school we actually had a little side business um like hacking into the smart kids and like selling their assignments to people who uh which in retrospect I don't condone this but it's like that's basically I wanted more time to play video games so I didn't have a job I would just like we ran a business uh hacking into other students assignments um wow which that's pretty pretty it's like CH before Che right yeah exactly yeah like like homework assignments but it was it was local to the school so yeah I mean that's industrious and I'm not going to travel back in time and condemn your kid even though obviously preo camp we have very high standards for academic honesty and I I am not going to condone any such thing but I will say that like I mean that is something you can do like I used to print out Mortal Kombat moves lists like back in in the arcade like nobody knew the moves you had to like learn the moves through trial and eror like oh forward down forward forward High punch we'll do this Subzero you know Fine Thing uh but like people didn't know that so I like collected all that stuff and I formatted this nice sheet and I would like take it to the arcades and I be like hey you you know you want all the moves from Mortal Kombat they be like sure be like all right five bucks and like just sell it and like exactly money just materializing out of nowhere right uh so similarly kind of like that early hustle well yes exactly I would I would much more recommend that route of actually creating materials in academic honesty but this is this is almost like pre this is I think I was like 14 or 15 so I kind of I'm like well the file's already there like we hadn't really been bombarded with like the importance of academic Integrity just yet and I'm just like well I have all these people who are like uh no one was doing their assignments cuz we were all I knew they weren't because we were all playing video games together and then I was like anyway um so basically into computers like the whole time um and then uh yeah Outdoors was just because it's I feel like it's in Australia it's like it's there so you're just um into it like we go to the the beach like for family holidays like we have so Brisbane uh North is the Sunshine Coast about an an hour drive 45 minutes and that's a whole bunch of beautiful beaches and then an hour south is the Gold Coast which I'm sure many people around the world have heard of and that's again just a collection of of some of the best beaches in the world so it's whereas like other cities like Perth there's a beach at the city so like you're uh you could have skyscrapers but then 10 minutes down the street it's like ocean and then of course Sydney there's Bondi Beach all that sort of stuff so it's like it's very yeah ingrained in Australian culture to to be beachy and Outdoors but and because most of the population is I think it's within 25 km of a coastline like 95% plus of the population is within 25 km of a of a coastline so it's kind of just ingrained in us to to go to the beach or to be outdoors and I feel like we're kind of a couple of years behind in terms of tech technology well that's what I noticed as a kid was like we'd always you'd see the new stuff get released in like the us or uh UK and we kind of it would trickle down to Australia in like 2 to 3 years I think the Gap as as like slightly Clos because of just globalization and whatnot but especially when I was hacking around on my Dad's laptop I'd be see all these new releases and be like oh I got to wait like 2 to 3 years I remember buying like my first Xbox accessory on the internet net and like having to convince my mom that it was safe and it took 6 weeks to arrive and then it was like it broke after 2 or 3 days and then it took 8 weeks to send back and get another one and like that me geographically you're very far from you know I know you're a lot closer to Asia than you are to North America that's what I mean right so a lot of things arrive late in Australia but not Wi-Fi it was indeed invented by the Australian science uh Bureau or or I can't can't it's it's called uh it's called CS i o the National Science agency so like as as a taxpaying American one of the one of the I guess many developments that that have come out of like the US government uh is what is GPS right like so all those GPS satellites that was technically like us taxpayers that like subsidize the development for that for and now everyone in the world gets a benefit from having like Apple Maps or Google Maps or um you know various tools that they can embed into their website or GIS systems things like that um and I want to thank the people of Australia for subsidizing development of Wi-Fi because I use it every day exactly same thing it's kind of yeah ubiquitous now you know yeah yeah I mean like those waves are passing through you as we speak they're everywhere yeah well maybe actually yeah maybe in a 100 years time we'll kind of go oh maybe it wasn't a good idea to pass electromagnetic waves through uh the entire world yeah but it's not ionizing I don't think it's too big a deal we'll see we'll see I could be wrong touchwood as they say in Australia or knock on wood as they say here in the US uh so I am really stoked to so it sounds like you had a great childhood um being uh an outdoorsy being a slacker you know the there's this saying about programmers that like the best programmers are lazy cuz they'll figure out like inventive ways to not have to do work um and it sounds like you're like at least like as a child you were someone lazy uh is that is that a mischaracterization well I was lazy towards the things that I was um told to do in a sense like so school like I was I was kind of uh naturally curious in terms of I would just I remember when I was like 10 I just would read the I read this Atlas like back to front like this, 1100 page Atlas and so I just I got an A in geography Without Really Trying but I was just fascinated by uh this this Atlas And so that Trend kind of continued through High School Xbox came out and I just would play Call of Duty and we we we ran the the number one this is back before e-gaming was like a a thing um Esports sorry um um we had this online forum and I would just we was like six of us at high school we had this team and our team was the number one in Australia so I would get home from school from like 3:30 cuz I liveed so close and we'd just plan strategies for Call of Duty until like midnight and then I'd go to sleep and wake up at 10: to 8: to be at school by 8:15 um so that was basically my senior years at high school but it was I would I would like just I don't know get passing grades through high school not really excelling but we we really did excel at Call of Duty because it was just fascinating to us you could put that as an extracurricular on your college application I guess like one Call of Duty team in Australia maybe the counselor would like recognize the value in that did you did you end up going to school like University or what did you do I did I did straight out of high school I went to University the University of Queensland which is a beautiful University one of the best in Australia and it's um it's uh uh for the first two years I was kind of just floundering around and um because it was just the thing that you did right um I didn't really know what I wanted to do out of high school because as I said I was just into Call of Duty but um or into computer games but back then 2010 that was not viable like to keep going down that path like computer video games it's still not really viable like like the biggest like Esports like the I my understanding is they got a lot of investment but they haven't been able to recoup that and a lot of them are not very solvent right now um well that's what I mean right like it's like yeah it was it was if it's a if it's a fake business now it was even more of a fake business um like in 2010 when I graduated high school but then yeah I I went to University kind of just um I was like I wanted to be a doctor for the I guess more so for the just The Prestige of saying I'm I'm a doctor rather than that was fundamentally it's a powerful thing to be able to say at a dinner party or on a plane when somebody's having Cardiac Arrest I'm a doctor everybody fors and you can walk right over and you can help the person right uh exactly I I can definitely see the appeal yeah so that's that's that's that was the main reason was more so for the status rather than for the the uh like actual will to or interest in studying human systems and helping people and so I basically fa failed almost every subject for the first two years of University before I got called in by the dean of Science and to his credit was basically like hey is everything okay like your grades are terrible and should you really be at University if you're if you're getting grades and um I kind of offered the um excuse that my dad was sick and like he he was and still is but it was really just because I was like well I'm lazy and I'm not really interested in what I'm studying and so on the side of that was the real reason the good reason was oh yeah my dad was sick please don't kick me out of University um but on the side while I was doing biom medicine and that sort of stuff I was studying food science and nutrition on my own uh because I was getting into the gym and I'm like I want to be fit I want to be healthy I want to lift weights I want to build muscle and so I was watching YouTube till 11:00 p.m. at night of people talking about how to work out bodybuilding nutrition that sort of stuff and so I was telling him that that's like he's like what are you interested in I'm like well I've been learning about this and he's like why don't you just change to food science and nutrition and I was like huh like a light bulb clicked I was like yeah you're right I can just change to what I'm interested in I kind of didn't I don't know it sounds like such a simple realization but um because I was so sort of stuck in that uh floundering mindset of like I don't really want to study this I'm failing this scks um I don't really know I'm kind of lost basically um that was just such a simple light bulb and I'm still really thankful to uh his name is Peter or it might have been pod I think pod's his nickname but Peter said he's like why don't you just study what you're interested in you're already learning about it and I'm like wow such a simple realization and then I I changed to food science Nutrition the next semester and basically got top of the class for the next 3 years cuz I I'd already many of the things in the first two years I'd already like learned on my own and so that was kind of like Premed was probably what you were studying do they have an equivalent Premed yeah yeah uh you probably already knew a lot about like biology right like all the different you know metabolism systems and stuff like that sort of I didn't even do biology in high school when I say I was really just like oh this girl is that I like is also doing um biom medicine I want to do biom medicine because I like this girl and I also like the idea of one day telling people I'm I'm a doctor but that's that was 17-year-old me and I guess with time you become a little wiser hopefully so um yeah that's what that's what happened and then yeah studied that and then passed with or actually didn't just pass like almost top of the class the whole way through and then uh graduated with that and I basically that was the more important lesson was learning how to learn like I I I realized that I was like oh if I'm interested in something and I just dedicate myself in a semi structured SL experimental fashion I can learn whatever I want basically like it felt that's how it felt that it felt like a superpower it was was like oh I'm just going to have to devote myself wholeheartedly to something that I'm fascinated and interested in yeah subsequently coding programming is just basically exactly the same thing yeah recurring uh character stepping onto the stage of the preo camp podcast each week learning how to learn uh and of course obligatory mention of Barbara Oakley's uh learn how to learn course which you know Leon Noel has mentioned I think Ben Awad mentioned it like a lot of different people who've been on the podcast have talked about the uh learning how to learn and just what it unlocks for you like I I'll I'll go on Soliloquy here uh developers the number one job descript is not sitting there and turning coffee into code or whatever it says on the cup uh but rather learning that is the killer app that is really what you're paid to do is learn what this Legacy code base is what it how it works and then going in and learning you know what you need to do to accomplish whatever goal has been put in front of you by your boss or by the open source Community or whoever you're beholden to um to get things done right so um yeah learn how to learn all right that that concludes the little the little ad read for learn how to learn as a mentality and yeah so so you were able to finish school you walked away with not just a degree from a a good school uh but also probably a lot of knowledge about health and nutrition but also just like this skill to learn new things what do you do from there yeah I would say that's yeah that was yeah the learning how to learn was just that's the Meta School right you can apply that to anything so that was 2015 I graduated and then uh I studied um uh Eastern languages for a year because I was working at Apple as a um a genius quote unquote that that was my actual title like youd help people when their uh their Hardware wouldn't turn on or something like that what was the most common thing that people would ask for help with uh broken screens or battery life of iPhones okay and so and so um they and then of course backups so uh it's kind of been ingrained into me to be um uh how would you say kind of obs obsessed with backing things up so um the amount of times I would have to have a very difficult conversation with someone to basically say everything on your laptop is gone because you didn't have a backup and you spilled coffee all over it and now it won't turn on and that we basically couldn't help you you may be able to go to like a data recovery center but again that there's no promises there because I mean it's all liquid damaged so they were very tough conversations to have and so yeah I've got goodness like half a dozen backups of almost everything everywhere okay what just off the cuff what is a good backup strategy I talked to like the CTO of my company he had a hard drive like a local hard drive that he had like time machine on uh time machine being the Apple backup thing I'm not sure if it's still called time machine but he had time machine and every day when he showed up with his laptop he plugg into that external hard drive at work and then he had one at home and he plugged into that and his reasoning is it's unlikely that both of these hard drives going to are going to fail concurrently or that there's going to fire be a fire that breaks out both at my home and my office concurrently so as long as I have one backup I'm good but like would that level of redundancy be enough for you or do you have you go even more hardcore than that I would say that yeah that would be like the minimum for me is uh so I've got my my personal setup is just um got the external hard drive uh got two of those actually I have uh iCloud on my personal Mac and then I have Dropbox and then I have cuz again these are just ease of use and I think as a developer I could probably set up my own cloud bucket these days but just the ease of use of these software companies are basically dedicated towards backups and then finally back Blaze um so they do basically just a um complete encapsulation of your hard drive and put that on their cloud storage so there's about five different versions of my just my local um MacBook Pro but then I have my server upstairs my deep learning PC um and so right now that's a bit more rudimentary I kind of just um rsync uh cuz it's Linux machine I just rsync yeah every every week or so um with an external hard drive and then uh I cuz I basically just use it to write code that's all just get committed um and then so if I was if that computer again touch W was to just completely fail I could um get a new one tomorrow plug in the hard drive R sync it across it would take a bit set up of course and then get clone the repos that I've been working on and then be off to the market awesome so like I mean that's got to be a very difficult conversation to have with a total stranger who just walked in like you know I always think like if you ever played like Friday the 13th on NES when you lose it doesn't just say game over it says you and all your friends are dead game [Laughter] over it's almost like that that kind of thing like the text appears every file that you loved is gone forever it it wasn't yeah thankfully it was like it was fairly rare but it was often enough to to yeah to be burned into my brain um but yeah it was it was that was really that job was actually really cool because it was like it was I could work on technology all day just be fascinated by that cuz I'm a nerd and um I would just talk to people all day so that's like kind of it built up my people skills and communication skills CU I would have to explain like a hardware repair or a file system structure in a in a way that uh a grandmother could understand because all your customers would be from different walks of life some of them wouldn't be able to speak English as well as I could um cuz we were in the city I so we would get tourists we would get uh elderly people we would get University students and so um I think over my three or so years there working part-time I I served 4 and a half 5,000 customers and so um it's just that repetition h of explaining something in something that someone could understand it was and that kind of really helped me um going into like programming and then subsequently um teaching myself that cuz I was write notes and code and then explain it to myself in a simple way that I understood and then later on a few years later would be like okay I've kind of built up this latent skill of talking to people talking to anyone um cuz I drove Uber for a year as well um and so that's kind of uh um like another where like you get in and you're just meeting a new person every 10 minutes and it's like okay some people don't want to talk that's fair I'm quite the I don't know animated talkative person uh just generally I think I get that from being the oldest brother and kind of got to be almost the leader of the pack you know and um that kind of gave me that yeah latent skill of just being able to have a conversation with anyone about almost anything so i' would be comfortable in walking into any room and starting a conversation with someone um from from zero and then in a technical aspect I would also be quite com but there was something I knew explaining or figuring out the where where they were um in terms of the the knowledge stack uh whether it be zero or whether it be okay we can just talk as if we were um both completely on the same level um so I don't really have to explain everything in depth if you know what I mean so that was that was really valuable I didn't quite realize how valuable that was at the time it was just like oh this is a cool job I'm working for Apple it's it's the biggest technology company in the world um it's not necessarily on the building the machines or the um the software that's what I wanted to get into um right but it was yeah that skill that you sort of just like at University it was that meta skill that I didn't quite realize until after the fact that was so important it's like it's very easy to go oh yeah you learned food science and nutrition like that's the important thing you you learned from there and it's like well to me I look back and I was like I was studying that for fun uh yes it is very helpful to know about health and nutrition and fitness and whatnot but it's also really good meta skill to have is being able to learn something that you're interested in and then another meta skill of like being able to talk to almost anyone about anything and that's that's certainly just in retrospect that i' sort of noticed that and so yeah after I left University Apple um had a program that would um if you want because there was so many International customers if you wanted to uh do a language course they would subsidize the the language course um and so I studied Japanese and Chinese for a year at um at the same University that I went to because I was just really interested in eastern languages and Eastern culture and we had a few um Chinese and Japanese speaking people that worked with us so it could practice there and then when customers would come in um I could I could talk to them like one of my favorite things to say was like if it was a Chinese speaking customer I would be like May and would just like instantly recognize what I would say which is do you have a backup or not um because like just saying do you have a backup if if someone hadn't really um been studying English for a number of years that concept of backup idiomatic expression what does that mean exactly right they're like they're like what does that mean and then I'd say ne May B they be like and it's like Mayo like as in okay no back up it's like okay well then we need to focus on that first yeah yeah and in Japanese do they have like a word for that or they just say back up yeah I didn't yeah I didn't quite think like the terms we didn't get to use it as much in Japanese I don't know it I don't know it if they have a specific term because um I'm not sure what yeah they just say back up a larger customer base was yeah Chinese speaking Yeah Yeah but um say back up yeah you're right I think that might be correct and like B fun is is kind of like uh like piece of of equipment like equipment yeah cuz play B is like show B like like gear almost um yeah I'm sorry I'm not going to like try to pick apart the emology on like there may be some people in the audience who do yeah yeah I spent 20 years like learning Chinese and I'm getting back in Japanese I'm going to go to Japan for summer uh so I've been like hardcore tooling up on my Japanese uh usage and and vocabulary and stuff but but yeah and if you hear my kids uh in the background they're very excited to be home after a long day of school so uh yeah that's super chill that you were able to like learn uh one thing that a lot of people may not understand is like the Australian economy and is like almost like closer and more interlined with like Japan and and China and and countries over in Asia than it is with like North America and Europe I mean would you say that's that's accurate just because of the location the time zones yeah I think so yeah we definitely have I I don't I'm not qu quite really well versed on um political relations between other countries but um yeah we definitely do a lot of trade with China I think a lot of our um and in Japan actually my brother is an accountant who um uh who audits large companies in Australia and um there was this one uh Meat Company so we have a lot of Outback and land so Farms to to raise cattle and whatnot and uh my brother was working on the their case and he said that they export they had one customer from Japan who who who runs like this really nice restaurant in Japan flew to Australia and bought $25 million worth of meat to ship back to Japan so that's just like one one customer so um I was like w um CU he had to come to Australia to see it in person to make sure it was high enough quality for the restaurant and so I'm just imagining if that's just if that's one customer who like running a high-end restaurant in Japan like imagine how many other deals are going on in other areas so uh yeah we definitely definitely have a lot of exports I know we we export a lot of wine and whatnot to to China and that sort of stuff so yeah and and I mean this podcast isn't about Australia but but I do want to note for a lot of people they they may not realize just the population of Australia is pretty small it's like maybe 25 30 million people like uh that's it yeah 25 it might not even be 25 I think it just under 25 I I but yeah it's definitely around that Brisbane the city that I live in is like 2 to 2 A2 million I think Sydney is number one at about 3 and 1/2 to 4 million Melbourne's around 3 to 3 and A2 um around that and so that's about over 50% of the population is just on the east coast in three major cities uh and then so it's yeah decreasing after that so yeah and to put that in context that's smaller than the population of Texas just one of the 56 one of the bigger yeah uh W just Texas yeah oh my goodness yeah uh so so it's like we hear all this stuff about Australia of course it's like a big pop cult like so many amazing musicians especially from the past 15 years or so have come out of Australia but it's actually not that many people especially when you compare to like China the big you know economic superpower of the region over there right we're small but Marty you know down here in Australia the way down yeah very cool so uh that's so cool so you got to learn some language and I just want to back up and talk about th those two meta skills you talk about learning how to learn and then learning how to take what You' just learned and communicate it to people like you may have not realized it at the time but you were basically learning everything you needed to know to be a successful tutorial Creator probably like would would you say that's the case yeah so I didn't even yeah it's it's again hindsight's 2020 right you can look back and go wow that's that's how you can apply that to this um and so that's that's a really hard thing when when people say oh you need experience for for jobs and whatnot and it's like it's sometimes the experience that you have isn't that quantifiable um whereas like yeah if I have a degree in in XY Z that's really quantifiable but then there's other things like I've talked to 5,000 people I'm not sure how many resumés um like you'll see that sort of as a extra curriculum kind of thing if you know what I mean yeah but also my my dad was a teacher and so I kind of I guess just that was just in nature and nurture as I was growing up and um we ran a family swim school like that was our family business and so I was a swimming coach from basically um 13 onwards growing up in the business and just interacting with with customers um of our of our business and learning how to teach teach swimming to to children and people the same age as me and so I feel like looking back it's it's yeah you you kind of realize it's like oh that actually helped for this that actually helped for that and you kind of piece the puzzle together yeah yeah well uh that's so cool it sounds like you had like a pretty industrious upbringing I didn't even realize you were like a swim instructor I guess all the water around it makes sense to get really good at swimming I don't I don't like pools very much anymore though because uh yeah I spent basically every day in a pool for 10 years wow so like one of the things that people enjoy unwinding with you're just dreading it like oh not the pool again they're R just as a kid I was like well as long as if it's fresh water like it's it's like okay but it's like um because chlorine like I was just I just always smelled like chlorine as a kid and I it's just like no I'm like if the pool's chlorine it's like I'll I'll happily skip that I'll just wait till we go to the beach or something yeah so let's talk about you know your progression into the workforce not just as like a Serv sector I mean like that's a really cool thing to be able to CL like I helped 5,000 not just talk to 5,000 people but like helped them solve their problems right at the Apple App Store over the course of like three years uh and sometimes even using foreign languages in the process uh can you talk about like how you ultimately transitioned into software development yeah so okay so 2016 I studied languages working parttime at Apple and then the start of 2017 was when I was like okay I go to I want to I always wanted to like basically code apps or build something that I was like I was using these computers every day and I'm like I want to build something that's like of my own um and so the start of 2017 I think it was about February I was just like I handed in uh my resignation to Apple and um my friend who worked there as well handed in his resonation at the same time and then we're like okay we're going to build we're going to work on a startup together and so this is where free code Camp came into play uh we were learning web development and we were going to the gym and we were like it really sucks that we can't work out together because you go to gym X and I go to gym Y and so we um started to build a website that we were like Airbnb was getting really big and we were like there should be an Airbnb for gyms and so that's what we started to build was any gym and we built out the website it was it wasn't coded from scratch so wasn't quite free code Camp spec it was just WordPress and um like a bunch of different plugins like maps and all these sort of things but we were we were piecing it together it kind of it looked pretty good from a design standpoint but then okay we're like okay now we have this website but now we need customers and so because in again in retrospect I didn't even know that it was a Marketplace um but that's what it was I was just like okay yeah every be like a two-sided Marketplace that's what I mean yeah sorry yeah you're right like yeah Airbnb has the marketplace of the the host of people who put up their homes or apartments and that sort of stuff and then it has the the customers who rent space in those and so in our case the host would be the gyms so they would offer their gym space that already exists and the customers would be people like me and my friend who didn't necessarily want a membership there uh like mbnb you don't want to rent the place for a year you just want to stay in there for or and then the same with with anyy was I just want to go for a session and I want that to be easy rather than to walk in sign all the the forms pay the fee get harassed by Gym Consultants or that sort of stuff I just want to pay whatever it is and do one session and be done and so we were thought yeah that's great in our heads yeah I mean I can tell you from having gone to like my friend's gym like I want to be a guest and then like you go you work out for like 20 minutes or something you're like all right let's Jet and then there's a sales guy who Corners you takes you in a room like hey you got to sign up and do like all this gives you the hard cell right like I don't want that I just want to go work out yeah yeah so I feel like there's still there's still room for for some I'm not working on this now like it the startup didn't work because um but I still feel like there's room for it if we could work if you could work out the right model and get the gyms on board and get safety all that sort of jazz ready but we were like okay we've got map we can find gy in a local area using Google Maps API all and then uh we're like well we actually need to get like we can't just have people rocking up to these gyms like we don't have uh like passes or anything and so we like okay we're just going to go on foot and so we just mapped out like a section of the town like near where his apartment was and we're like there's 10 gyms in this area let's just walk to them yeah today and talk to them and again young and naive just like going through like hey this this is this is our cool bit service we want to get you more customers and we started talking to them and then we kind of quickly realized that uh we didn't have a business unless we could get the gyms on board and um we then also realized that after we got pretty close with one of the gym managers and they were sort of like I like this idea but look at this and then they showed us their sort of numbers and they had 2,000 clients signed up but only 100 of them would come more than once a week and so he's he basically told us that their business model depends on people not showing up so I was like because if all 2,000 people came at one point no um the gym would obviously be overrun right and so so it's almost like you're solving a problem that is actually like a feature not a bug of the way the gy work yes and so we wanted more people to come to the gym but they're like we we we don't want more people we want more people sign up and and have those memberships that you forget about and just do the direct debit every week or month or whatever it is and you don't come to the gym and so we were just like oh and then we kind of went to another gym we kind of realized that that's just a repeating function that you get all these people who sign up for the gym at the start of the year and then go five times in January and then basically the the curve just drops off you go four times in February three times in March and then by June you you've forgotten about it and you're doing something else and then of course the hurdle to cancel it is not just like I'd like to cancel my gym membership it's like talk to our sales guy about why you'd like to to uh cancel it and so in theory it was a great business but then in practice it was like oh okay we the gyms depend on people not showing up and so if we were trying to charge a smaller fee for a one off it doesn't help their business yeah I mean like some of these like at least in the US sorry to go on like a long tangent about gyms you're but uh yeah in in the US there was like this big chain of like gyms was like I think it was like balles or something like that and they like signed people up for like these five 10e contracts or something like that and they they literally went out of business and people still had to pay for the next 5 years they're like no longer rendering services but the the contract was so Ironclad that people had to continue to pay for the next 5 years and they didn't get any gy in return it was just going to like the crors of the gym or whatever yeah so uh well that's what I mean it's I Ed the YCA the YMCA here in the US it's it's a charity they probably have them in Australia too it's a charity that was started 150 years ago and uh they may use some pushy techniques but like my experience has been like we just go there uh we we have like a family family membership whenever we travel like over the summer I'll often go to like Asia because my my kids live in Asia or I'm sorry my my my grandparents live in Asia my wife's from China uh we met in grad school and we might like leave and we can just go in and we can freeze the membership for two months and we don't have to pay anything it's not a big deal yeah see and so like I hope more gyms be like adopt that approach of like building up a reputation over like 150 years and like just being a chill wholesome place to work out rather than like being pushy and like using these techniques but I can definitely see based on everything you said that this would be like a non-starter of a business exactly right like it was so perfect in our heads and then yeah in in reality uh and I think that's like just another again in retrospect like an important thing of like okay well we test as soon as possible in the real world like is does this actually work because many great ideas are in theory but then in in do they translate to reality well yeah we got to test for it right well since we're talking earlier about China and Chinese I will take the opportunity to share one of those beautiful four character idioms in Chinese yeah Z which is close the doors and build the Chariot there was this like old inventor from like ancient China when of course all these like four character idioms came about and uh he had this Vision this kind of like platonic ideal of like what a car should be like the concept car like the uh the Cyber truck of his day right and so he goes in and he like builds it and everything and and he's making all these fine tuns based on his theory he's Theory crafting right but he never actually takes it out on the road and when his Masterpiece is finally done he takes it out on the road just falls apart it doesn't work at all and so he sunk all that time into it he he closed the doors and built the Chariot when he should have been doing what you did very early on which is go out and actually talk to people running the gyms like I could easily see you and your friends spending like months developing this prototype before you went out and got that user research done and so it sounds like you dodged a bullet by just quickly discovering oh this is not like this is not going to happen because the fundamental economics and the incentives of the people in this space dictate that Airbnb for gyms is not like a great idea and I I frequently said Airbnb is not a great idea and I'm amazed that people actually let strangers into their house uh to to stay like for short periods of time when they're not like professional hot tells they know what they're doing you know but um but yeah like like kudos to you for quickly iterating on your idea and and incorporating the feedback so did it become apparent pretty quickly that you should just Scupper this idea and move on to something else yeah it well my memory is probably a bit hazy on exact timelines but let's just say there a few months we probably spent uh a couple of weeks cuz this is we had no jobs it was kind of like um we're living off savings and my friend had an apartment at uh at um West End which is a suburb towards a city in Brisbane and I was just like basically sleeping on his couch and um studying through the day trying to build this website and then because he lived close to a few gyms we went around and we're like well we should probably test this quite quickly um because we we wanted to start making money uh because again we had no jobs and we were just living off savings um yeah so we found out I would say it was definely not longer than 6 months and probably closer to 3 to 4 that we were we were working on that um because the website truth be told was quite simplistic you had a a place to sign up for customers and a place to find gyms and then a place for um again we probably have to work out a lot of the payment system as it worked but it's like no point having a great payment system if no money is slowing through right so you you it sounds like this was like the mother of all kind of practice projects like you were getting to turn around and apply your nent coding skills to immediately build out this product and and you know yeah well it was a problem that we had like it was like there was a group of us at at working at the Apple Store that would um like we were all in the gym but because it the store was in the middle of the city you had people who lived everywhere around the city and so it was unfeasible to always just go One Direction we just wanted to stay in the city and go somewhere just local so it just made sense to us of like like as if you'd go to a restaurant you just walk in and buy a meal we're like we want that we want that workflow to to work for gyms instead of yeah walk in sign the form get 100 text messages from someone asking you if you want to sign up which is like no a simple transaction you know and like I think Apple pay was out by that stage I'm not entirely sure so it's like it could just be a past that you just walk in and scan yeah anyway that was maybe maybe it could work these days I think class pass is like another big startup that was doing something similar I think they they actually got funding and maybe made it to profitability I'm not entirely sure but uh that that was yeah my my dive into web development and then pretty quickly after that I uh realized that web development for me was far too tedious I'm like I want to do programming cuz I was like I have to build the workflow for that to go to there to go to there that to go to there like just every every block like again this is my laziness kicking in I was like how could I the whole point of this was to automate the stuff but right now I'm just building everything step by step and then of course machine learning was uh before it was really called AI it was was on the rise and then uh um basically every Tech podcast that I listen to um or YouTube video that I watched was about technology and then someone would talk about uh machine learning and then I was like what is this and then I started to of research into machine learning and then I was like oh so you just tell the computer here's my stuff you learn the patents and I was like yeah that's that's what I'm talking about that's what I want to use computers for so it's like that that laziness creeping back in like computer do the hard work for me you do the work yeah exactly um and so yeah that just began the fascination I think I probably discovered it properly in the middle of 2017 cuz I left Apple it started 2017 and then uh we we worked on the startup for a few months caned it because yeah it's like well we've spoken to enough gyms now I bet you this is the case around every gym um and also we're we're young and naive I'm like 24 or 25 or whatnot and have no money so it's like if we had millions of funding we could probably sway the gyms in some way but we were like bootstrapping the whole thing yeah anyway um we yeah so middle of 2017 start get into machine learning and on the weekends I was driving Uber so to support myself so I was still living at home with my parents and um so I had a a pretty strict routine for I guess 9 to 12 months of just Monday to Friday study study machine learning and Friday night Saturday afternoon tonight and Sunday morning is Drive Uber because they were the best they were the best days in terms of economics Friday night everyone's going out Sly so I was a Hermit for like a good a good solid year um and this is where I sort of middle of 2017 discovered like machine learning end of 2017 like September or close to the end I created my own AI master's degree which was just I'd kind of done enough two to three months of like Foundation stuff in terms of to realize what was the the the overarching um skills needed for it then I'm okay well I've got this skill of learning how to learn so I'm just going to create my own curriculum using online resources um that are kind of just a collation of I guess lists of of AI resources that I'm like these look like pretty good and I just put them in a just a blog post of basically I don't know let's say 10 resources the number is a bit arbitrary it's still on my website and I'm like okay over the next and I was publishing it publicly on M to basically no one but I wanted to just have a public record to keep me accountable so I'm like over the next X amount of time I'm going to go through these resources and get these skills and then I'll just get a job in machine learning that was my my thinking so you were using like your blog post and social media as a kind of an accountability mechanism to uh to like a commitment device essentially like I'm going to learn all this stuff and everybody's like good luck I'll look forward to hearing about the results of you learning all this stuff and then you're like oh damn I can't quit now everybody's waiting for me to share my results yeah well that's that's exact 100% that's exactly what I what I what I wanted to do and well one because yeah accountability two because I'm a very sort of I like interacting with other people who are interested in uh I think that's quite natural to just interact with people who are interested in what you're interested in yeah and uh this is 2017 so machine learning and Al is nowhere near as popular as it is now it was certainly gaining traction but C certainly when chat GPD came out everybody updated their LinkedIn profile HEI exactly or machine learning engine no longer yeah I had to update all my courses you know it's no longer teaching machine learning it's teaching Ai and machine learning um but uh but yeah the fundamental that isn't a marketing gimmick by the way that's the fun the fundamental Technologies are still the same between machine learning and ai ai is just a sort of yeah a bit more of a hyped up term at the moment um but yeah and it wasn't for a long time by the way a lot of people like so a little history lesson AI has like Winters and Summers right so like in the 1950s the US government's like we think we can create like human level intelligence within like 10 months or something with this team of like University researchers and of course that didn't work out right uh and then later they're like oh well now we've got an AI that can beat Gary Casp a chess you know blue IBM right or we've got Watson that can beat everybody at Jeopardy because it can hit the buzzer at superhuman rates before people can even hit their buzzer so they can't compete with it right like you know like like there all these like we we've had many yeah you're right I think it's almost 10 to 15 years we get yeah a new technology that's going to uh replace humans um so that's yeah that's very well you're right that's very well historically documented right back from like yeah the 20s almost um and then mechan now we can just get this to do it by itself without having a c underneath the the chess player right but uh sorry for like going off on a tangent I think it's that's important to have context yeah yeah cuz machine learning engineer for a while was like basically machine learning you'd say that instead of AI because I there was a stigma associated with AI because AI had disappointed so many people let people down you keep saying that like you know how 9000 is just around the corner the sky just around the corner and yet you know this AI can barely do anything you know so people like were oversold like and you could say to some extent there's so much hyper on AI uh that we being oversold right now but at the same time I use you know GPT 4 all the time like I use it several times a day there's something there there right uh it's not all just you know Sizzle and no steak this time it's there is real utility I totally agree yeah the same with me like I use AI tools uh I think now it's a lot more um customer facing as in like it's it's you can you can see what's happening with with AI tools now whereas in the background with predictive models so generative models the difference between predictive and AI uh sorry predictive Ai and generative AI predictive your sort of predicting is your email spam or not spam as in it's it's a set number of of classes um there's a a confined output whereas generative is kind of the out outut space is almost infinite it's still constra constrained to the distribution of of data that the model has seen but it's certainly a lot wider than just a a a unique set of of different labels so now um for the last I guess maybe 8 10 years we've had predictive AI running behind the scenes um for credit card transactions for battery optimization for email classification for uh audio quality optimization but all these things are sort of behind the scenes whereas now I feel like the reason why it's it's getting a lot more traction now is because it's generative we can all interact with it with language with image and soon many more modalities yeah yeah it's super duper exciting and I I don't mean to like uh denigrate like the effort of all the AI engineering and all the breakthroughs and everything uh I just wanted to point out that like the machine machine learning term it's synonymous with AI right would you say that's that's accurate yes well that's yeah that's that's how I use it I guess you'd probably have people who who want to sort of Define it a bit more but like to me it's like there's one big umbrella up the top and that's AI artificial intelligence actually funny story about AI I told my grandfather um a few years ago when I was first studying Ai and he he thought it was artificial insemination which is um yes something related to impregnate cattles yeah artificially but no it's like no it's it's to do with uh technology so that just goes to show like how quickly technology can change that's like only two generations like my granddad to my parents to me and it's like all of a sudden we have totally new nures for for different paradigms uh anyway AI is up top for me at the big umbrella and then under that is machine learning and then under that is deep learning and so really if we're being really specific all of the AI that's currently generative AI is been powered by Deep learning um and so that's a subset of machine learning and machine learning is a subset of the field of AR okay yeah a great little kind of like pyramid if you want to think about those different terms and how they relate to one another one question a lot of people ask me is data science how is that related to AI yes so data science would be to me is is in I guess a part of AI as well um cuz AI is just I mean what's intelligence it's yeah it's pretty broad right so data science to me is using data to predict things so you're you're trying to formulate hypothesis and seeing if the the data falsifies it or proves it uh somewhat correct and so whereas data analytics is you're analyzing the past so data science to me is predicting the future data analytics analyze the past so what has happened that's a little Nuance that I sort of that helps me get through that and then within data science if you're trying to predict the future Based on data uh one of the techniques you can use is machine learning so you you get you write an algorithm that is going to sift through much more data than any one person could ever look at and try to form a predictive model so that given a similar source of data You can predict something in the future yeah now again it's all probably istic so it's not it's not a magic box um but it can sort of give you some insight into what it might be yeah and so for example like uh one application of you know predictive models would be figuring out what the next good move is in a game of Go uh you know the game with the stones for example like the the the Deep Mind alphao um algorithm uh AI system whatever you want to call it that was able to master this ancient Chinese game that's like 5,000 years old that has that is you know many many orders of magnitude more complicated than chess in terms of the possible moves you can make and the complexity of different strategies and everything like that and my understanding is they took a huge data set of previously played go games that have been recorded step by step just like with chess they've got like their notation to record like who move where and put stones we and essentially they use that as the training data and so initially like that whole model is to predict what the next like the highest confidence interval or like the highest you know expected value uh move is when you're putting a piece on the go board right um and so so we're talking about superhuman data like you could be a go encyclopedic you know go player you could potentially even have like some sort of neuro Divergent Quirk where you could literally remember full games of go but you wouldn't be able to like consume the entire Corpus of all the games that these AI systems can consume and eventually they stopped even feeding it trading data and they just like had it play against itself a whole lot for hundreds of years s simulated right it does not matter how smart you are you cannot play against yourself for hundreds of years because you're going to die all right so so yeah so when we talk about superhuman you know models and like these can do useful things that humans simply cannot do there are real world limitations on the exent of like human intelligence right uh so that that's what's so exciting about it for me like watching like the the go uh or watching like the AI that plays Starcraft against the best you know Korean Starcraft players and stuff like that and like just the applications of these to see interesting things that come out uh that that I think is like that is what really gets me hyped up uh as somebody who enjoys games and stuff well the StarCraft one to me is is really fascinating cuz that in requires like team interaction versus like uh the go one is obviously very impressive but it's it's sort of you your your solo and your your board state is is influenced sure you're influenced by an opposition but you're largely your actions influence the board State as well but then you have like um uh what was it called or DOTA sorry or Starcraft what where you have a team of five versus another team of five and it's like okay now you have multiple people influencing multiple things and like these people if you if you're like a a newcomer you you could play Dota or Starcraft for 5 years but then someone who's an expert at it just comes and wipes the floor with you like and you have zero chance but then these AIS can can train for a month in computer simulator time and then they're going up against the world champions who are so far removed from the average player it's it's ridiculous but then the the AI is catching up in in a month of simulation yeah yeah I mean it really it it's probably a foregone conclusion that we have artificial general intelligence given like the constraints of just the human mind and the fact that like we're working with something that evolved to like find calories and like avoid predators and stuff and they're working with like basically the fruits of like the finest engineering Minds from the past you know 100 years and then all the mathematicians and philosophers and everybody else that was also kind of like seminal in in creating the uh the Corpus upon which like those Technologies sit upon right like yeah I mean I would be really amazed if we didn't have systems that were way smarter than human beings within the next 50 or 100 years in some Dimensions I'm I'm not sure that we'll have uh you know what are your thoughts on uh so so we can talk a lot more about AI cuz one of the things I want to talk about is how should somebody approach learning these things because frankly I think everybody who's interested in technology should learn about these things um and but the question is how do they go about doing it you just did that what would be your high level advice to people who want to just start let's say they just know a little bit about programming technology they know how to use you know a browser maybe they know a little bit of like a scripting language like python maybe they've used Microsoft Access or even SQL or something like like that so they have like basic knowledge of like how to get stuff done with computers how do they go from there to where you are in terms of building your own models and and uh you know getting things done with these tools well I feel like yeah there's almost an unlimited amount of approaches but I if you're just let's say you're just comfortable using a computer but you want to you know you want to get into using AI stuff I would say that the the the tri entry array the way that I did it was just relentlessly follow your curiosity and it sounds like trite but it's like that's just going to fuel you through like I had to do it in a I realized that my I needed a sort of structured curiosity following uh that's why I created my own uh rough path of like courses to follow to build a um a foundational knowledge and then after that I just once I'd had the foundational knowledge I just went and go I'm going to just relentlessly follow this so that was my path and I feel like that's obviously unbiased but I feel like if you're if you don't want to go to to college or back to school or whatnot the the resources on the internet are the best you're going to get especially in this new field well not new field but in the current climate of AI the universities and colleges just don't catch up as fast as the the field moves however that does not to say if you're already involved in a computer science or mathematic or some sort of technology related or any actually any related um uh degree or something at school because these Technologies are so wide ranging uh I am actually of a big uh like in big favor of we need people who are well trained in non-technological backgrounds to pick up these skills so that it's not just a bunch of computer nerds running the the AI world we need people who are well-versed in many different fields uh offering Insight because what what a lot of Engineers and I'm guilty of this and people who are in Tech uh sort of one downside is that they again they build Solutions because that's what they like to do they're Engineers but then the solutions don't map to the actual real world like um like say for example a healthcare solution um was built by Engineers but then you you start to talk to nurses and people who actually provide Healthcare and they say that's just not going to work because we have to I remember in my machine learning job perfect case of this is that we were before uh transcription got really quite good like as it is now you can download a model from hugging face and you can get almost Flawless transcription at least definitely usable transcription uh we were building models to transcribe doctor's notes um as in so you go to a doctor's appointment and instead of the doctor there typing on the keyboard uh while you're talking it would just automatically write down what you were speaking so you had a record there yeah so we're building this system the model worked pretty good um at least good for 2018 standards and then we went to deliver it to the the system and we kind of I was speaking to the business Dev or the business person that was running it and their whole goal was to they they found out that on average a doctor spends 5 minutes during each appointment to type notes and they wanted to make that basically zero and that way a doctor could see like three more patients a day if you add that up and just and just in my head I'm like sort of okay that works fantastic from an economic standpoint but I'm like why do people go to the doctor in the first place I actually asked that in the meeting and uh I kind of got looked at looked at like very like what are you talking about like what what of course people go to a doctor cuz they're sick and I'm like yeah they might be sick but they're also going for empathy and to to chat with another human being and if we're just basically going oh yeah you sa 5 minutes so now you can see three more patients a day it's just like a a factory line of like people in and out maybe that extra 5 minutes the doctor could actually talk to the person of like how's your family how's your how's your wife how's your kids um all that sort of stuff and like get to know the person and uh again it can't be proven in like medical terms but the spiritual sort of side of me feels like that that is almost as healing as like prescribing a a drug or a treatment of just someone caring about your problem the whole notion of bedside manner from your doctor essentially like they they that they care that they're not just that you're not just numbers on a chart that you're not just like this graph or this uh scan that they're looking at oh blah blah blah blah blah like they actually look at you they actually connect with you I mean that's like you know I've changed doctors a lot over the years and that's one of the most common reasons like it doesn't seem like they even give a damn about me like I'm just another you know appointment and they got to dash out the door to the next one they don't even bother sitting down you know like those are all flags for me red flags that I need to find a different you know primary care physician so yeah uh and and those are the kinds of things that you learn from being out in the field you know not building The Chariot behind closed doors but actually going out there but but was that the lesson of what you were trying to tell or or was it something different like well that's what I was saying it's like don't if you want to get into the yeah so the initial question was yeah how to best get into it and again I don't have the best I only really have my uh experience and my experience is uh there are a lot of things that you'll see in the in the tech field that are sort of almost too techy as in like you'll you'll you can yes there are fundamental skills that you you are going to have to learn which is uh programming in some sense and the language of AI at the moment is Python programming language you you basically can't go wrong if you start with that um and then writing C A lot of people tell me oh I'm scared of the math of AI and ML and uh it is all math I'm not going to talk away from that it is writing computer code to trigger math at scale that's that's like AI algorithms in a nutshell however these days I deal with 99% code rather than than math so I'm I I'm not an expert in math I don't have a background I don't have mathematics degree but I can write um computer code to to run that math and the computer is going to implement that for me and so I'm more interested in the Practical side of writing the code to to implement things then U doing the research into the mathematical foundations of the algorithm um to find new algorithms I'm more interested in applying existing techniques to things that I find interesting rather than Reinventing completely new techniques so there's two kind of sides to it a lot a lot gets uh in the there's kind of the the AI and machine learning research and I would say if you want to go into that and be a researcher um you're going to need to to learn the mathematics inside and out because that's what the the researchers doing and and let's say if you want to do research is that like you should just stay at a university generally or best yes that would be the best path for again you can there is no the internet there is no limits there's a website called archive which is ax X i.org which basically all the new machine learning research gets published to that and it's free so truly if you just wanted to sit at home and you you you have hours per day to read through archive you can just create your own self-study um like Facebook Google uh MIT Stanford all publish their research papers to Archive and then the code will be on GitHub so if you want to just really just find your own self way and then there's communities on Discord these days which are like basically all of those resources are completely free so if you have the discipline to go through that yourself um by all means but if you're just getting started you're at college already you're studying a technology thing mathematics whatnot you want to go to research uh I would stick with that path but then also just just always remember that um there are resources outside of college that are very ible um so that's a research side of things but practicality wise I would say if you come from a different background that's also really cool because now you have a different set of knowledge uh that you can sort of start to see what AI is capable of and go how can I apply this to my current field and that's kind of what what I I'm doing now as in I studied um food science and nutrition at University and now my brother and I are working on a startup called nutrify which is uh at the moment it's a like a Pokédex for food so or Shazam for food you take a photo of food and it will tell you about it so it uses computer vision and then maps that to nutrition data and then you can create a food diary get nutritional insights from from that so that's my practical standpoint is I have studied food science and nutrition and now I've kind of got versed in AI Technologies I'm trying to marry the two yeah um so that's kind of the between like research is is the Academic Way of like um you're just trying to push the field forward in terms of um studying new algorithms building new algorithms going through research papers and trying to collate the best things and again it's a very big overlap too with with the Practical side of things of building a solution that other people can actually use but I would say the the building the Practical solution that other people can use is a lot more code heavy versus the research is still Codey but also there's a lot more math involved yeah so like you know free camp for example we're teaching Applied Mathematics and it's always got to focus on okay how do you actually implement this in Python uh so if you look at the the algebra course that we published a few months ago and we've got a pre-calculus course coming in a couple months and we're going to have all the engineering math that you would get at like an you know at a big top computer science program here in the US would be like MIT Caltech Stanford uh but um um what we're doing is instead of focusing on like okay let's draw out all these formulas on the Whiteboard and stuff like that you still kind of learn that but it's all about okay let's implement this in code because any mathematical expression can be implemented in code uh right like you see that summation symbol the big scary I think it's like an Epsilon I can't remember all the Greek letters but like that's just a for right um yeah so like and like all these different uh variables you pull in and stuff like order of operations all stuff that's all code right you can Implement math in code and python as you said is like the the language dour people aren't using r as much or Julia Lang or these other computational languages I mean you can certainly use that but you can't go wrong with python there's like python libraries for pretty much everything right that you need to do in machine learning yeah and I feel like yeah the the language of AI these days is is or the framework of AI and deep learning is pi torch and the pi is python so um yeah pytorch is a framework for deep learning and it's all all accessible in Python you can WR you can write machine learning code in basically any programming language but uh python is is where you want to be if you're starting from zero and learning to to program like today or learning to code I would 100% recommend python yeah and you know pytorch like so that's like the Facebook open source project uh for machine learning and then there's also tensorflow which we have courses on both uh which is also another python uh machine learning library that's created by Google and you know all these tools are like open source they're free you don't have to pay for mat lab or you know like a site license or or like a lot of these other tools that are frankly pretty expensive like like uh and the python ecosystem is just amazing so yeah I I think it's cool that you do so much and you have tutorials on so many of these tools in Python uh you know your most popular video ever I think is a pytorch video and one of your other most popular videos is a tensor flow these are like yeah 10hour videos 25h hour videos that are literally in a day yeah literally a day like just drink drink drink your energy drink and and get your crack your knuckles and let's let's do some uh machine learning right so I want to fire off lot of questions so just to recap what you're were saying because you had so many insights there and I I want to try to like break it down for everybody listening so there's the applied side the engineering side and then there's the science side if you want to be an engineer start engineering right start grabbing tools off the shelf and start applying them to problems if you want to be a scientist there's something to be said for being in the Ivory Tower and going to just get a PhD but there are sites like AR archive I'm not sure exactly howc yeah like uh I think it's like from a Cornell or something like that like I'll link to that in the show notes we're going to put links to a lot of stuff we talk about in the show notes uh in the video description if you're watching this on YouTube but uh one of the interesting things that I think you said is whatever domain expertise you already have in your case nutrition you can bring that from wherever you are learning to code is kind of a skill that you layer on top of your existing domain expertise like I'm a teacher I was a school director right like uh learning the code for me was just a skill that would allow me to get new things done as a teacher right like free C Camp is partially like I'm I'm just one of many many many people who've worked on free C Camp but it's like an expression of my interest in education and you know apply like applying that using coding right same thing with like learning a foreign language you're just allowing yourself to go and do whatever it is you you do if you're a doctor and you learn Spanish bam now you can go to like Latin America and help people right right if there's like an earthquake there you can potentially go down there and be a relief worker or something like that right so it's just another tool through which you can express your expertise and and so I think it's really interesting that you have this domain expertise that you already built up and you're still in the space you're still building you know health and nutrition Wellness type tools but you're applying the machine learning skills that you have to get those things done that's so cool yeah so uh I want to just fire off a lot of really quick questions so thing that like people always say to me is like oh I want to do machine learning but I don't have the GPU or whatever like uh it used to be like coding I'm like oh cool like I used like a a netbook which was like 300 bucks it had 2 gigabytes of RAM and like I used that for like two or three years when I first started learning the code learning the code technically you don't really need a lot of computational power you're just you know writing loops and like building algorithms and data structures and stuff but with machine learning you do kind of need some Hardware to get somewhere right yeah so I would say yeah if you're okay let's let's just talk from from beginner level like if you're if you're just getting into programming uh any any modern laptop will do a MacBook Air or um I'm not quite I've been using Apple for a while now so I'm not quite familiar with the the the most recent Windows laptops but um anyone of those will do so you can you can get started learning all the basic sort of stuff but then as you no I don't want I don't want to downplay it and say that it's basic you can you can get quite far using uh a standard desktop machine but then if you want to start getting into more and more advanced um using bigger models training your own uh state-of-the-art AI models then that's where you're going to need more compute power um so for example I do 90% of my work day to-day on a M1 MacBook Pro but then when I train deep learning models and so sorry if there's a car starting that's my housemate leaving um when I train deep learning models I I write the code on my MacBook Pro but then it it sends it to I've got a a PC in my closet with a big GPU in there and it trains the model on there and then I take the model from that and I deploy it to a smartphone in the case of nutrify but then if you want to go even higher to build something like chat gbt you're going to need uh a warehouse for the of computers and so there's there's quite a large scale there however if you're I I I tend to towards like all my courses are beginner friendly right so I tend to I I like I'm a big advocate for the little guy you know the person hacking around on their their computer and trying to build something that can scale to a million people if you really wanted it to and so uh I don't have access to thousands of gpus like meta Facebook Microsoft oh meta and Facebook the same people anyway I don't have access to a warehouse of computers so I'm mainly interested in and the field this is a little bit kind of um it's something you kind I I don't realize it cuz I'm so embedded in the field but it's something that you it's a really big blessing uh when you get into the field is that a lot of there's such an open culture in uh machine learning like most as I said most of the research papers um are published free to read which is actually quite different to a lot of other fields and a lot of the the research papers come with the code so here's the code you can run on your own computer to implement it and then finally here is the trained model that you can uh use on your own computer so you may not necessarily and the benefit of this is that the trained model they may have been trained by researchers using uh let's say 10 or eight Nvidia gpus that are worth $220,000 each so it's $160,000 machine it's trained for 3 days on that machine they're going to release that model for you so you don't have to repeat the training you've already got the finished product and you can apply that on your machine now whether you'll be able to run that at scale is another question so there's kind of this is quite a bit of rabbit hole but I'll just keep going with it is that what I'm saying is you can have a a consumer level machine and then run these models that larger Enterprises have released and then you can start to build with things like that and so one of the biggest expenses in compute power for a machine learning model or deep learning model is the pre-training so training it and that the pre-training is the p in GPT so pre-training that can often take weeks and months for some models um literally weeks and months of time so you'll have in the case of GPT uh I don't think they release the exact numbers but I feel like having a computer like a th000 gpus or maybe more 10,000 running simultaneously for 3 months is not an understatement so that just goes to show you billions of dollars to train these models in some cases we're not yeah we'll be in the billions soon um I think the billion dollar training run will probably happen in the next 3 years or so but um definitely multi multi-millions like we're talk I think Lama 2 from Facebook was like a $10 million training run um and so then now we can use uh with Frameworks like mlx from from Apple we can run those models on MacBooks so that's that's really cool and then you have other tools like if you're running on a consumer machine you can go to uh Google collab which is a free resource basically like Google Docs but um for computing and you can run python code you can access uh free gpus so a GPU is a graphical processing unit that is very good at doing parallel computations which is what you need to do in machine learning you need to run numerical calculations and you need to do a lot of them to find patterns in data and so gpus just happen to be really good at doing many computations simultaneously so you can find those patterns faster and so um Frameworks like pytorch allow you to build models load models and you got hugging face Transformers free open source Library you can load pre-trained models into your Google coab instance and then start to build from there so as a practical example um um the other day like my friend works for an insurance company and um one of their big problems is H I'm getting fireworks on my thing your Apple gestures thing turned on for anybody who listening to the audio version you have to turn that off if you use an apple if you do double thumbs up and you have it turned on which it is by default you'll see like these fireworks so I didn't even know just triggered that that's hilarious it's a delightful way to learn about it in real time sorry just had fireworks behind me yeah so you were talking about your friend at the insurance C yes so um he was using uh Google collab to do some research on um the insurance document so as you can imagine insurance companies have a lot of files and a lot of sort of uh information cases to dig through and then he was using Google collab to download a pre-trained language model to um basically read through email threads to to find find like who was at fault who was um uh the person making the claim uh what was the amount what was the date and then structure that data from uh from just raw text of the email thread which is kind of very tedious to go through then would have the language model automatically turn that into Json which is a structured data pattern which is uh much easier to navigate so that's just like one one of many many examples of how someone from like an insurance like uh field not not exactly fully related to AI can start to use AI tools to to help them for their workflow yeah I mean just like software in general like anything you're doing that is tedious that you can describe really well historically you wouldn't necessarily be able to do that you might be able to describe in great detail exactly how you look at different photos and determine whether they need moderator attention or not right like if you let's say you work in one of Facebook's many like moderator kind of like data center or call center type places uh where you're like going through and like checking every photo that uh somebody has flagged on Facebook for example like you might be able to very accurately describe using words how you would go about determining whether a photo was inappropriate and needed to be removed but it was very hard to teach a computer to do that until now now it's like we've got like so much more powerful models and it's just like that that like kind of evolutionary leap and so now now like I would never have trusted a computer to like look at a thread of emails and figure out who was at fault or like pull out all these San details because human language is so complicated and everything like five years ago I wouldn't just thought oh yeah like I'm not even going to attempt to write that kind of program it would just not work yeah even even two years ago like this is this is really sort of new stuff that we're kind of um yeah we're still figuring out the the exact use cases that um all the the new tools generative AI there's like no real it's not an understatement to say that yeah in the last two years the it went from basically not working to working really really well so and of course it's still like there's it's not perfect but I mean that's the nature of machine learning it's it's probabilistic it's never going to be 100% perfect um but it can definitely be useful yeah and and that is why all the hoopla all the hubub about AI all of a sudden is because it's it's not just hype uh there have been some major breakthroughs that you and I are using like free Camp uses AI like everybody I know who's a software engineer is like learning how to work these into the workflows it's a new tool in your toolbox that you can pull out for novel types of problems that before you just had to you know use regx and pre essentially or something like that right I I feel like yeah I feel like now is I I get a question quite often is like should I even start now because of AI just like coming in and just changing everything and I I was I'm a big Advocate um of like now is again it sounds TR but it's like now is like one of the best times ever to learn to code because the tools are so helpful um and like there's a difference like there's a kind of a niche difference between like coding and programming like I would say uh like programming to me is just thinking like a lot of my time I think if if people get the um the idea of like if you're a software engineer you just spend all day um writing code I would say now majority like I do write a lot of code however uh I use AI tools such as GitHub co-pilot to write a lot of the boilerplate but for the stuff that I have to think out um that's mostly thought of like I've got a a text document here where I've just been collecting ideas on a workflow that i' I'd like to be implementing and I've been refining that workflow in my head not not writing any any code yet and then go okay that might work that might work and then I write the code to test the workflow and all the boilerplate stuff is is is helped with AI tools and what I mean by boiler plate that's another term in the industry of like really tedious stuff that you kind of um you're not really thinking about when you write it but you have it has to be there to get started so it's like the I guess the if you're writing an assignment it's like the the structure of the assignment it like needs to needs to be there to start but it doesn't really take much thought but it is tedious to get and so AI tools now are very good at that but for like specific workflows um that still requires programming which is very thoughtful um and a lot of the time is is a lot of the code actually you'll write 1,000 lines of code but then you'll delete 700 because um you don't actually need them and it's more efficient the way you're doing it the first time is really rough um that's really big thing in machine learning is you'll you'll write like a thousand different experiments and maybe 10 of them work quite well and you'll start to pursue those so a lot of my day is spent yes writing code but now ai tools write a lot of that Baseline stuff for me um I still know I could if I didn't have the AI tool this is another important point is that if I didn't have the AI tool I'd still be able to write it I would just it would just um take me longer uh if that makes sense so um to the people who yeah who ask who ask me and this is what I reply with is that it's it's actually a really great time to start learning to code because the the the really big hurdle when you're starting is that it's you have to learn you're learning a new language and then um like it's like learning Chinese learning Japanese but then once once you get over that that big hurdle of the momentum of like just learning the the syntax of like what goes where where it goes then you can start to okay you don't have to use such a big thought process just to get the syntax out you can write the thing and keep going uh and so AI tools are really really good like chat GPT Gemini I mean there's half a dozen of these chat Bots now um at writing fundamental python code and so if you have an error in Python which is the language of machine learning they're incredibly helpful for debugging it um and almost instantaneously whereas in the the past you'd have to really quite get good at searching and reading documentation and whatnot that's that's not to say that's not still a valuable skill but especially for that just that initial hurdle of learning something completely new they're really really good for that and then once you've got some momentum you can start to apply your skills there's always going to be that sort of um uh time where you need you you can't necessarily use a chat assistant or whatnot or an AI tool is that you're going to basically have to delve into the documentation of something that's new and hasn't been around for a while and you need to to just basically whack your head up against the wall and figure it out like take some time um yeah I mean there there aren't really any shortcuts like you said like all the stuff you're talking about doing you're still the problem solver you're just using these tools as a way to solve the problem but if if you look at the code is like very succinct instructions for what to do like yes the machine is very good at like figuring out like from its massive Corpus like co-pilot has I don't know hundreds of thousands millions of code examples from python for implementing similar types of functions or something that it can draw from to create uh similar type boiler plate uh but at the end of the day like figuring out what work needs to be done and like the the high level thinking you still need human intellect to do that and I don't think that's going to change dramatically anytime soon like I always say this like you watch Star Trek and the computer doesn't just do everything it doesn't just run the ship everything's done like the humans don't do anything no the humans are doing lots of things they're figuring out where to go they're figuring out what is important uh they're asking questions of the computer they're they're creating new programs within the computer to get things done and of course that's just a science fiction show written in like the' 60s 7s 90s um you know but I think it's it's a pretty accurate assessment of like where technology ultimately going it's just yet stronger tools yet longer levers for us to use to get more work done as empowered individuals so yeah I just want to quickly second what you just said Daniel I think it's more important than ever that people learn how to code and uh people worry like oh but like what if you know the jobs are taken by technology like the people the nutritionists who know how to use machine learning the nutritionists who know how to to get things done with computers are going to be much more competitive than the ones that just throw up their hands in Despair and say oh I don't need to learn that stuff or I'm not smart enough to learn that stuff right same thing with teachers who learn technology same thing with um you know doctors who learn how to use these Technologies like everybody in society can really benefit from layering this skill set on top of their domain expertise and on top of you know their profession their trade so um couple just fire questions man I just want to like make good use of your time first of all thanks again for waking up super duper early right what time is it over there uh it's 8:51 now but I was yeah it was uh I know we had a couple of technical delays on my end but it was um yeah starting nice and early 6:00 a.m. 6:30 so Google collab you mentioned that I just want to say if you use the free Camp curriculum we make heavy use of Google collab uh for our machine learning section amazing Tool uh it's basically just an implementation of jupyter notebook which is a popular open source python yeah that's correct um and then uh I've been taking detailed notes as you you talk one other thing that uh I think is is kind of cool that uh you know you've created this like I don't know have you done it every year but like the state of AI like what you need to learn in 2022 you know like like you create these comprehensive learning paths have you created one for 2024 yet not it's not every year because I didn't want to um one it took me a long time to do but another one is like it almost like because it was changing so fast the last couple of years I uh I've held off on my most recent one so I think um I don't want to make promises but there's there's like I've definitely got a note of like uh research and and whatnot of what I'd like to include in the next version and I think this may be the you the 2024 2025 and for the next couple of years and whatnot onwards um but yeah it was it was sort of a resource that I'd like to have so that was the 2020 machine learning road map and that sort of stayed valid like 95% valid up until about the launch of chat GPT and then it was like maybe 80% valid because it it and again when I say valid it doesn't mean that the material is invalid it's that it's it's missing a few of the newer things that are available if that makes sense so um the materials are still very valid for like the majority of the um uh pre generative AI sort of Boom but um the next one will sort of definitely have a lot more inform about generative AI tools cuz as I said it was up until about 2 years ago that they essentially didn't work very well at all and then it's like now they work really well so well we use them yeah exactly right yeah well I I think it goes without saying that if you do want to create like a 2025 Edition I always like to to joke that like you can release the 2025 Edition in like September whenever they release like September 2024 2025 Edition cuz that's when they released the 2025 version of cars things like that if you decide to do that that's what I think that's that's that's we'd be like super duper hyped to to post it or cross posted on free C Camp's YouTube channel for the community cuz like I enjoyed watching I I think I watched like two or three hours it was a long video uh and just like seeing you break down all the different fields and how all these disciplines interrelate because one thing you have is you don't have that academic pedigree of I've got a PhD in machine learning from XYZ you know Academy but instead you I mean you're a nutritionist who taught himself all this stuff and so you have that kind of like learning how to learn like in the in the trenches type mentality of like okay here's how I actually get things done here's how like the jobs that I've worked uh the first job you ever got by the way I you got to tell this story you got a your first developer job but you didn't actually apply for it like how did what happened exactly how did you uh how did you get your first job this is why I'm not good at answering the questions of like how to get a job it was um cuz when I got how do I get my first job because like I M was kind of and I just can only really speak to what I actually did was I was just again when I started studying machine learning at AI I created my online pathway and I was just posting on LinkedIn and my blog and medium uh in YouTube of like different different things that I was learning and then uh someone must have found a post on LinkedIn and I guess they were um they knew someone else and they knew that someone else was again this is 2018 so machine learning AI wasn't uh as big as as big as it is now and so they talked to someone else and then I got a a message saying hey do you want to I see you're into this machine learning and stuff do you want to um catch up for coffee and I'm like yeah sure and then I went caught up for coffee and I kind of again impostor syndrome kicks in cuz I'm like like oh wow do I actually have these skills or I'm just talking about them and I spoke with with this person named Mike and uh he he didn't actually know he was learning from me and so I was like oh wow okay um I am actually learning some stuff and then he's like I know someone who I want you to meet and I think like they're if your if your goal is to get a job in the field they can probably help you out too I'm like okay sweet so met this person cam uh two weeks later let's say and then I'm like yeah my goal is like I'm studying online uh just in my bedroom I'm driving Uber on the weekends and I think I'm going to go to the US um in a few months and just find a job like that was literally my plan I was sort of that's that's my sort of happy go lucky L Nature of like I'm just going to fly to the US and I'm going to make it work right and I kind of just had that rational confidence of like this is just going to work out and again that's not to say that there weren't moments of like oh am I doing the right thing am I am I actually going to make it what all this sort of stuff but nonetheless I still just had this deep down confidence that it's like no if I just keep dedicating myself something will happen lo and behold he replies to that with um well have you considered just staying in Australia and I'm like no all this technology stuff is based in the US and he's like no we're actually we've got a like a small company like we're starting to bring machine learning into a austalia and I'm like okay and he's like yeah do you want to come in next Monday and I'm like okay sweet and so I was uh the third machine learning engineer at this um well Machine learning engineer at that stage um at this technology company based in Brisbane and we were a consultancy firm partnering with other big businesses in Australia because machine learning was just taking off with this another writing this hype cycle and so we would go into big businesses like insurance companies Banks uh medical centers and figure out whether machine learning could be applied to their existing problems and so that's where uh that sort of ties back into what I was saying before if you're from other Industries the technology is so widespread now that you can kind of do some research I'm not going to say you're going to have an epiphany straight away but do some research and start to draw patterns uh between what you do day today and what these Technologies are capable of and so it was again luck I of stumbled into it but on one side of it I was also publishing what I was doing quite vigorously online um was it groundbreaking stuff no I was just sharing what I'd learned and so it was almost like leveraging the power of the internet uh to you you never really know what's ever going to happen um so none of my stuff like went super viral or whatnot but that's not I I've never banked on that it was just I was just like I want to uh um progress my skills and put that out there because I know I knew no one and I still know no one really well I know a few people now actually in real life in my like Hometown who was studying machine learning so I'm like I have to find people online um so that was sort of my mentality and that's why I kept like posting things um to sort of uh not necessarily like put my my word out there but also find other people who were interested in it yeah yeah that's really cool um you have said in the past that if if an employer on the on the notion of like chopping around for jobs and trying to find a good employer you said If an employer doesn't think that building projects counts as experience you probably don't want to work there can you tell me a little bit about how you decide like where you want to work and like what you look for in an employer well I've only ever had one machine learning job and then I started to work for myself myself so that and that's actually been my goal the whole time is to start my own company um so I was uh I thought I wanted to work for a big tech company like Facebook or um meta or Google or something like that but then I realized that once I learned more about the lifestyle I was like no that's I I want want to sort of work for myself have my own hours have uh my own schedule work on what I'm interested in kind of just a very like selfish point of view but it was just a that's just what I want as long as I'm following my own curiosity and sharing that with the world I figured like that's that's how I want to live and I want to be able to spend time with my family um look after my dad that sort of stuff and I figured it's it would be really hard to do that if I was in the US and working uh a super big hour tech company sure the compensation all in comping lifestyle basically working for mean yeah like the sure the compensation the salaries everyone's heard of those but I I kind of learned that the economic sort of um category wasn't as appealing to me as it once was um compared to other things yeah and so obviously you need some baseline level of Economic Security so I wasn't going to just do nothing and just sporadically follow interest without actually turning that into economic value um and so um back to yeah I left I left that job started working for myself and um what I'd learned there is how machine learning gets applied in many different Industries or at least how it attempts to be applied um and so what I was like I was like okay I now have a balance of I've learned machine learning myself and now I've applied it in several different Industries I um again I published an article online of like things I've learned in my first machine learning job and that kind of went semi viral I guess in the tech space on medium and then someone from Canada read it Andre who's my business partner now uh who was teaching web development um on on udemy and his own website he's like we're looking for um machine learning curriculum all of our students looking for machine learning and I'm like okay I've never really taught a course online but I have applied these skills and I have learned them and so I started to to create a course on machine learning of like all the things that I'd actually worked on but I didn't quite learn online so it's very practical Hands-On um so there's no there's really no math in my course it's all just code and like this is the code that you you would write in day-to-day as a machine learning engineer and so that turned out to be a way that I could uh make a living by teaching skills I could also follow my curiosity by researching and then putting that um into the courses and then now because of the the flexibility that gives me of I don't have necessarily scheduled hours to make resources I I just make them at will um I can now use time to build nutrify which is and fund like self-fund uh a startup with my brother that combines what I'm so it's kind of like a uh a loop of like I learn new things to include in the courses and I also there a fly above my head learn new things to included in the courses and then I use those things to build in a practical application nutrify which anyone with an iPhone can download and use and so then I integrate that back the new stuff in the Practical world into updated materials it's a virtual Circle it's it's like an engine that you built for yourself where like that you've got this kind of like feedback loop that benefits you that's yeah well that's that's it but it's all sort of just stemmed from just having very real conversation with myself how I'd like to how I'd like to live and again there's there's many different ways to I had a physics teacher who had a funny saying who was like there's many different ways to skin a cat and I have no idea where that saying came from but it just comes into my head every time I think about like someone's like what's the what's the right path to learn or what's the right path to to live and I'm like there's there's so many different ways that is just it's just and if you you take advice from me take advice from anyone else it's it's all going to just add to basically sum to zero because some people are will be like um who have the opposite point of view to me is like well you should go work at a big tech company and earn $300,000 a year for five years and do and then after that do XY Z yeah I me that's that's reasonable advice too by the way like that's what I'm saying but if you can do what you did you skipped that you kind of leap frogged over the because you're able to sell fund by creating courses by having a successful YouTube channel and uh you know and you're able to apply your knowledge so that you can sell fund your startup which self funding your startup is a hundred times better than having to like answer to investors all the time and I say that as like well I've talked to so many people that uh prefer it it depends on like you can't have a self-funded satellite company but yeah yeah for some businesses yeah fortunately um our business is yeah we don't have any expenses except for our time oh and database storage which is quite cheap but um yeah we've been rejected a few times from uh from like things like this AI Grant and we've been rejected half a doz there was even a yeah my friend applied for that uh yeah everybody gets rejected like I think pretty much everybody I've had on the podcast has been rejected from black commentator for example well that's what I mean so that's that's what I mean it's like it's the sort of the the nature of um the tech field in general is that it's just it's it's l literally just experimenting nonstop and like most of those experiments won't work and so whether it's like you're beginning to uh learn whatever it is it's just like if you are curious about something it's just like follow that Curiosity but just beware that basically the most of what you do won't necessarily work out so but then what you can you do with that is apply what doesn't work the the quicker you figure out what doesn't work the quicker you figure out what does and so that's one of the reasons uh like I have a saying in my all my courses is experiment experiment experiment is because uh whether it's with learning how to learn whether it's with communic ating with customers or investors or something else like that it's like you're just it's just all one big experiment um and so the more experiments you run the the quicker you figure out what doesn't work the quicker you figure out what does yeah well uh that that is such a profound simple thing like just keep experimenting keep iterating on your ideas don't build the Chariot behind closed doors hopefully none of us are ever in a post-apocalyptic situation where it makes sense to skin a cat and eat it I hope we never again that's another one of those things from high school that I'm now remembering that's like I don't condone like maybe he grew up during the Depression or something I I don't know I don't know but it's just a saying he would do he would he would like do sign and cl mathematics and his head on the on the Whiteboard and then like write an algorithm and be like cuz he's a really smart like he's genius physicist and then he'd write like several different things and be like oh we don't have to do that formula we can do this one thousand ways to skin a cat so now it's just like whenever I think of whenever I think of something with multiple paths it's just ingrained to me of like why would someone say that but I don't know it makes sense to me well it's a very literary expression and I want to draw attention to uh for those of you not watching the video who are listening to the podcast which is how I listen I don't actually watch the videos I watch I listen to the podcast you can subscribe to the podcast distributed through the RSS feed the open Airwaves uh of this open protocol real simple syndication uh you can listen in apple podast pod uh Spotify wherever you listen to podcasts but if you're watching on YouTube you will see behind Daniel quite a stocked bookshelf you've got probably like I don't know two or 300 books on there uh yeah I think that's right yeah so you actually read physical books if you're like a technologist I love physical books yeah talk talk about that's my that's my addiction physical books so you just like take a book to the beach or where do you get most of your reading done uh yeah basically anywhere in the in the backyard um I have a little like just wooden table and chairs and I every morning I I usually out there um early morning and just um look at my garden and read some read some pages um coffee shops basically anywhere in bed before um going to sleep physical books like I'll often fall asleep holding a physical book um I just find it yeah very soothing very calming like in this again like my my life is mostly built around screen so I feel like um a physical book is just a it's not trying to sell you anything you've already bought it it's like it's like one of the last places on Earth where there's no ads um so hopefully thats the case yeah exactly right and it's um I just I just love it there's something very satisfying about the tangibility of like flipping through pages and like I guess I don't know completing a chapter or something like that and then like just the smell it's just very nostalgic to me I can't I can't really explain it but it's like it's I just know that I love it and one of the reasons I bring up books I'm not just like doing this to comment on your beautiful bookshelf back there uh you wrote a book and published it a novel if I recall correctly uh about maybe a year year and a half ago uh can you can you talk about that experience and what drove you to write fiction yeah well um so for those who aren't for listening this is my novel it's on the screen called Charlie walks but Daniel Burke it's in the flesh just to say that like to prove that Quincy is not making anything up this is but every word in this is by me uh some some help from uh uh an editor of course to to fix the typos and mistakes but um no I just I basically I like I'm a copycat if I see someone doing something that I find inspiring I want to do that and um so some of my best mentors in life are authors and I've never met I've never met anyone I can't think of any maybe in real life maybe I don't think I've met anyone who's written any of the books on the shelves in real life I mean a lot of them are passed on um but I read them and it's almost like yeah you you have this connection and it's just like that's inspire spiring to me and so a lot of my heroes are authors and so I'm like well I want to do that and I kind of didn't read much at all in high school or even just outside of university cuz I was it was a lot of books that you quote unquote have to read and that's kind of a theme with a lot of my learning is like I forget anything that I had to learn whereas I remember everything that I wanted remember every page of the atlas but nothing of you know high school chemistry class or something like that right exactly you're forced to learn that but but when you're pursuing your own interest it feels so fresh and interesting and like you've discovered this and yeah personal and I started reading yeah fiction novels the first proper fiction novel like obviously uh as a child I think I read a little bit of Harry Potter with with uh my parents and whatnot but then the first fiction novel that really got me into like wow writing can be this was uh post office by Charles macowski which is uh worked in a post office years like that was his fulltime job right that was his full-time job was working like excuse me um Charles macowski for people who don't know is a a very prominent American German author um but he for the majority of his life just worked dead beat jobs which is what he describes them as um and then he has a bunch of books that are basically just writing about this life and um that he lived and so post office was his first novel and it's basically just a story of how he worked at a post office for 12 years and hated it and uh a lot of people can read Bow's work and be sort of um I guess find it sad or like he had this very grim look on life but I I kind of and this is maybe my rosec colored glasses I kind of looked at it and it was just like you can make artwork out of the mundane that was so inspiring to me I was like this is I didn't want to put the book down I and I'd never really had that feeling with a book and like ever since I was a child when I was reading another um A Series of Unfortunate Events by Lemony Snicket that's like another series that I read as a kid um it Fly going around um and I read this book and I'm like oh my gosh you can just write a story about anything you want and it's amazing and so I'm like I want to do that and then that's where like I started writing like about probably a year or two after I read post office I started I started writing my own novel and um finished it basically through Co um when it was lockdowns and all that sort of stuff and uh yeah wanted to get it go right to the end of not just um not just writing behind closed doors but similar to sort of how I like to publish a lot of what I work on is I wanted to get it um out in the world I wanted to to to have my name basically amongst the the other people on the shelves and just be like that was fun I'm I'm glad I did that that was fun I didn't and if you had told me it's another thing it's like you told me when I was a young that I would be writing a my own book or fiction book I would be like no that's that's stupid but then again things change over time right yeah absolutely and uh I think that's really cool that you pursued that uh that dream of publishing a book and having your book on the Shelf uh a living author among many dead authors on your bookshelf back there uh I'm going to put a link to check out Charlie yeah one day but hopefully you got like another 80 years ahead of you man who knows what kind of medical advancements are going to be brought about by gener yeah that's a good point and of course I'm sure you're taking good care of your nutrition that being your domain of expertise so uh Daniel it's been such a blast having you on man it's been so great talking with you learning more about you because I feel like I couldn't find many good interviews Ken G's Ken's nearest neighbors podcast that was the only podcast interview I could find with you I'm like man this guy has so much to say let's put him in front of people so uh yeah I feel like I say enough on my YouTube channel and blog so I'm like a lot of I'm very glad you asked me an interview cuz I I say I do get requests but I I um yeah I I I definitely do decline a fair few because I kind of yeah I'm I'm a little bit of a Hermit despite what it sounds like yeah man well uh you're very well adjusted hermit well thanks again and hey everybody who's listened this far I just want to thank you for tuning in again I be more like Daniel check like follow your curiosity experiment experiment experiment and check out the show notes for lots of insights and links from Daniel and uh I hope you all have a fantastic fun productive week until next week happy coding