# [YTalks] Siraj Raval - Stories about YouTube, Plagiarism, and the Dangers of Fame (Interview)

## Метаданные

- **Канал:** Yannic Kilcher
- **YouTube:** https://www.youtube.com/watch?v=kEhEbVZQwjM
- **Дата:** 31.10.2021
- **Длительность:** 1:06:44
- **Просмотры:** 31,415
- **Источник:** https://ekstraktznaniy.ru/video/12709

## Описание

#ytalks #siraj #plagiarism

A conversation with Siraj Raval about his journey on YouTube, and the perils of fame.

OUTLINE:
0:00 - Intro
1:30 - Welcome
3:15 - Starting out: From Economics to YouTube
13:00 - More Views: Plagiarizing Video Content
23:30 - One Step Up: Copying A Research Paper
29:15 - Was there another way?
39:00 - Clickbait Course: Make Money with Machine Learning
50:30 - Rock Bottom and the Way Forward
1:01:30 - Advice for Future Generations

Siraj's Channel: https://www.youtube.com/c/SirajRaval

Links:
TabNine Code Completion (Referral): http://bit.ly/tabnine-yannick
YouTube: https://www.youtube.com/c/yannickilcher
Twitter: https://twitter.com/ykilcher
Discord: https://discord.gg/4H8xxDF
BitChute: https://www.bitchute.com/channel/yannic-kilcher
Minds: https://www.minds.com/ykilcher
Parler: https://parler.com/profile/YannicKilcher
LinkedIn: https://www.linkedin.com/in/ykilcher
BiliBili: https://space.bilibili.com/1824646584

If you want to support me, the best thing to 

## Транскрипт

### Intro []

the following is a conversation with siraj raval siraj has one of the largest channels in the machine learning youtube space over 700 000 people are subscribed to him as of this state uh siraj pumped out lots and lots of videos on topics such as coding tutorials explaining beginner's concept in machine learning and in other topics like blockchain or other computer science things now his rise came to an abrupt stop when a series of scandals hit him at the end of 2019 and there were a lot of articles written back then twitter posts made and even siraj himself made an apology video but i was wondering how did he feel like during all of this what did he think back then how did he come to this how did he feel during the highs and the lows of his career and how does he look back on things now i was struck by how straightforward siraj was in this conversation i was sure there was going to be wisdom in there for the rest of us be that youtubers or machine learners and i was not disappointed he was definitely honest looking back with a different view and we touched on many things in this conversation i hope you enjoy it i hope you find something in there that helps you and yeah let us know what you think

### Welcome [1:30]

well hello everyone um today is a special day um in many ways uh siraj who is my guest today is one of the pioneers of the field of ml youtube now i'm pretty sure uh pretty much every single person in the field has heard of siraj has seen him watched one of his videos or something like this and if i can maybe frame it a little bit there is that you were one of the first machine learning youtubers he became really popular quickly things went uphill more views and so on and then i think it's fair to say it kind of all came crashing down in like a very short period of time and yeah then it's it just sort of crumbled if i don't i can't really frame it any differently there seem to be like things one on top of another that just all came in like a month or so the same month it seemed crazy this time at the end of 2019 so yeah i'm i'm happy to host siraj today uh thanks so much for being here and uh you know talking and you agreed to talk a little bit about your side of things of what happened and what you're doing now so yeah welcome thanks it's great to be here i love your videos they're uh they're definitely they've got a personality and character to them that i definitely admire and i'd like to see more of thank you and yeah so

### Starting out: From Economics to YouTube [3:15]

i thank you well since you're since you're the og youtuber of this you know that i guess character is a little bit of what it takes i want to go back a little bit to the beginning though you if i recall correctly you started studying economics is that correct correct at columbia that was my freshman year i was an economics major yeah and for some reason you switched over to computer science because it what took you there um well i was um i took a semester to travel around europe using couchsurfing i was cow surfing for three and a half months and the first person that i couch surfed with in london his name was alex mccall he showed me um his terminal window he had a hackintosh that he made and he really inspired me to get into computer science it turned out you know several years later that alex uh wrote the book the o'reilly book on javascript and he has this really cool startup called clearbit that he already sold by now um but i got to meet him before all that happened and once i saw alex terminal and all the cool things he was doing i knew that once i got back to columbia i needed to like switch over to computer science because that was how you really made an impact in the world yeah so i i guess you saw pretty early that the impact was to be made right i think a lot of people go into economics and they think like they maybe think a little bit of money if they go into economics because it's kind of close to it but i guess computer science especially you know nowadays is really the impactful field or one of the impactful fields little known fact i also didn't i started out in medicine and then switched over to computer science so much of the same journey there and then did you finish computer science or no i dropped out um my senior year of all times to drop out wow yeah and i and that was because of youtube or no no folks so i dropped out because i had a robotic startup at the time we were making a six degree of freedom robot that would pick things up off the floor for older people with something called als because they can't bend over and uh we built a prototype raised money but it turns out like nobody would buy it and also there were some software problems at the time this wasn't this was like 2012. so um yeah i just um moved to san francisco from there from new york and then that's when i really started to feel like i was around my people like texas yeah you're american originally but from smaller town or big city or i'm from houston texas so i was born here my parents are from india um definitely have a deep connection with india i still dream about india um cool and then you were you're in san francisco and how did you get into youtube so i worked at a several contract jobs in san francisco for companies like cbs interactive doing mobile development i worked at meetup for a year um just as a general software engineer i was i started off as an intern and then um eventually the last job i had w2 job was at twilio the api company and i worked there as a developer educator for about eight months and then i was fired because i i think it was just a performance thing i that that's what they said so i don't know but i remember wanting i learned a lot at twilio about developer education and how innovative it could be um to give you an example we were learning about different ways of getting developers who use the twilio api and you know as i was writing documentation across nine different programming languages like ruby and php and python one thing that i was told by my mentor was that we don't want to use too many exclamation points inside of our documentation because if you have more than three what developers do is that they subconsciously think of not equals from code and that gives them a negative impression of the text and i was like that level of detail i never thought about that but it really is an art and so i started wanting to make videos on the side and actually my first three youtube videos i made while i was at twilio at the conference room at midnight when nobody was there and i showed it to my colleagues there and they were like my boss was like you know that's great that's cool we don't think developers are going to use videos as a learning tool they want something static like documentation and so that's when i thought well maybe there's something here and so once i got fired i got a severance and i had enough to live in san francisco for about six to eight months and that really gave me the impetus i remember i had all my stuff in a box um that they gave to me from my desk and i literally the day i was let go i walked across the street to a hair salon and then i got my hair dyed and i was like all right i'm all in on this youtube thing now like i have to figure out how to make this work did you just the hair did you consciously do that did you think i need some sort of a thing yeah i mean i was always inspired by um a guy named bill nye the science guy and how he was very unique character for general science and i thought what is my thing um i didn't know what exactly i wanted but i remember a roommate of mine at the time who was a matchmaker she was like you know you'd look really cool with like a silver streak in your hair i just tried it out i mean you chose better than me the sunglasses now i have to code with sunglasses which is annoying you get it you're very recognized with the sunglasses in person i get i get recognized within and without i think the hairline is gives it away but uh yeah that's how our branding works i guess so but yeah so then you just you started creating videos was it always machine learning or did you also like get into that somehow no so we started out my first few videos were all on bitcoin in fact my first video was called what is bitcoin yeah that's really i think a bitcoin is the soul of the hacker community um everything comes from bitcoin and emerges outwards from there if i i'm not religious but mike the closest thing to a religion would be bitcoin um but i started making machine learning videos just because it seemed really interesting and i was really interested alphago really was the catalyst for me like oh there's something here let me start making videos on this with no credentials no um phd or anything like that yeah also also i felt like um this is kind of weird to say out loud but like i'd spent six months in india traveling across the entire subcontinent before i started working at twilio and one thing that i saw was like you know i was living in such a box my whole life in the united states and india's such a beautiful country however there's a lot of issues there it is developing country kind of ascending country i like to say um but you know we can't just solve all of these problems in our lifetime and some of them are just they're going to take many generations of salt perhaps if we created some sort of super intelligence digital organism god it could solve everything for us and the thing that i personally could do was use my specific knowledge um to help make that happen in the form of funny interesting videos that would raise awareness around these technologies to as many people as possible and that would somehow increase the amount of research happening in the field and all of this together would accelerate development of a super intelligence yeah i mean that's i have one socialist like borderline communist friend and whenever i make fun of communism has never worked he always says like but we haven't tried with an ai supermind planner right and then i'm like yeah okay that's god he's got a point right but yeah so when did you so you had this plan of doing videos when did you really see that this could be something like was there a moment where you saw like wait you know views go up and was there like a particular moment or did it come you know slowly or when did you really feel like yeah i could make this work well i think it was three months into making videos once a week because back then i could only do once a week it took about 40 to 50 hours for a single video eventually i got up to three a week at my peak but after three months of one video a week i got um someone emailed me from this company called big ml which was a machine learning platform it was my first person who ever reached out to me they wanted to pay me for a series of videos and i was elated because ad revenue was like you know nothing really um i did have patreon that definitely helped for sure but that was my first i think they paid me 2k usd for six videos um which was huge and yeah and that was really like oh this is something and then of course udacity reached out to me and that was the biggest catalyst like for it to help um make their deep learning course nanodegree yep so yeah udacity but that also fell through if i recall correctly

### More Views: Plagiarizing Video Content [13:00]

and this is so maybe for people who don't know and you have made you've made an extensive like apology videos about this but it some of your videos or you know to the degree were plagiarized not exactly the videos but you would sort of write or show some code and then you would say like either like oh look at this code or watch me build a trading bot or something like this and you know just be very vague about the origins of the code and then you would you put attribution maybe really small at the bottom of the code but essentially it'd be other people's code that you presented is that about a fair framing of things so a lot of times you took other people's codes didn't fork it on github but just kind of downloaded it re-uploaded it and then changed the like the readme or maybe some wrapper and things so when yeah when was that do you was this always your mode of operating or did you like did you at some point start did it increase because that's what i'm wondering like i right you started out saying you know i could do raise awareness and so on and you ended by or ended you at some point you found yourself in a mode where you would a new video would just be like i take someone else's code i make a video claiming essentially inferring that i made it right how did you get from a to b so if it was a process it didn't happen all at once i mean if you look at my first few videos they were like i really did write the code for the first few videos um they were like 10 to 20 lines using the skills that i learned at twilio of like making something really basic a skeleton app that a developer could just download it hit compile and it runs make it as simple as possible i would look at these very complex repositories for the initial versions of tensorflow and um you know a neural conversational model by oriole vignelles who's my favorite researcher still to this day and just try to condense it into you know 10 20 lines as a wrapper um but over time i just it was like a gradual process of you know instead of just raising awareness it became more like chasing clout rape making the number go up for views likes and there was also like almost no accountability i was a lone actor i wasn't working with anybody um so that definitely made it easier to do something like that and um eventually like once i moved from san francisco to los angeles um and that was the last year and a half that i worked on youtube um so from 2018 to 2019 that's when um i think that was a bad move like i not really an la person but that's when i really started to um really chase the clout and pursue fame for the sake of it because i'd already gotten these opportunities and it seemed like i just needed to get to a million subscribers no matter what yeah a million is was that your personal goal or i mean for me a million was always the point a little bit where you could live off of ad revenue was it like this or was it just a number you liked or no it was just a number it was just like a fun little goal in my head yeah yeah it so and did you did you at any point feel like uh maybe i shouldn't do this maybe at the beginning and did it become easier for you or how did you think about yourself or did you just think you know everyone else is doing it or yeah i mean i guess i you know everybody is a protagonist of their own story right i felt like i was doing you know just having the little name in the very bottom of the github not forking the code but just putting it down there that made me you know feel guilt-free um yeah at the time but obviously that wasn't how i should have done it i mean obviously what you did was very public and therefore the backlash i felt was also very public i mean a lot of people got angry and you know once it all let's say came crashing down a lot of people came forward and said oh yeah me too i was also my code was plagiarized and so on i i feel like i have seen exactly stuff like this in research it like tons of times people essentially copying papers mildly attributing like once but essentially the entire page would be like taken from usually it's their or earlier papers so what authors will do is they will have like one new equation and then they'll write an eight page paper where seven and a half pages are essentially their old paper right and uh so i mean but that is never it's never as public right it's never as as big i guess the more public one is the the worse it gets when something like this really happens uh did you so i know i've read your udacity course that you said that became an issue there right people try to tell you can't plagiarize stuff is that correct or so i i've seen it like a tweet from someone at udacity saying you know the course fell through essentially because they try to tell you that that's not how they do things or what is or maybe you can tell a little bit what the udacity course you said that was a big thing for you why did it fall through yeah so you know the what happened with udacity was we had a 16 week course that i essentially designed and then udacity helped me build a team around that to help me one issue that one of the people at udacity had that i was working with he was also in the initial trailer video matt leonard was that i was not writing the code from scratch i was using existing examples and he didn't like that um we also didn't have that good a working relationship during the course but i think in terms of falling through um that happened like you know everybody made money from that course including udacity and there were several cohorts of students that didn't just run once i think it ran like three or four times you actually udacity actually approached me two years after that course was over to do another version of it and i did help get that too um in terms of following through yeah when all of this happened then you know people came out and said this stuff yeah i don't know what happened with the courts honestly i haven't okay i think maybe uh maybe i got this one wrong one wrong um yes and so i've seen like i've looked at your social blade and so on you're at about 700 k subscribers and i've seen also an interview with lex friedman and you where essentially you also told him like you know what matters to me is views i'm attuned to views to more subscribers and so on uh is it fair to say a little bit that you might have lost sight of you know the bigger picture or other things just in pursuit of this goal it is i was definitely disillusioned with um agi and the initial goals that i had at the start i definitely also had a you know an issue with uh i had like a drug problem near the end i was doing too much of a certain drug that makes you really um up and have a lot of energy and there was a point where i pretty much almost overdosed on it and that's when i knew like i even like you know called the cops on myself to because i thought i was going to die i don't i've never really said this out loud before but that was near the end this is basically like a month or two before um you know that scandal happened and i was just you know i just felt like i was unfallible like i was untouchable like i could do no wrong and yeah i'd never had that level of fame before as well like that was pretty that was it was quite a drug of its own as well on top of that but yeah it was a gradual process i think of going from uplifting developers and like that being the primary concern to also then chasing clout chasing fame wanting more opportunity more views um more recognition and just making uh stupid decisions yeah i can i mean i'm you know as a as a another youtuber i i get the draw of this like i unders i can i get this feeling of being sucked into these metrics and it's not only the metrics right the metrics are correlated with money correlated with fame and so on uh i like yeah i see the and so many youtubers fall into this right and uh your mistake was also a little bit that you your setting was in an maybe like an academic or a professional setting where people actually care about you know not stealing stuff and things like this so maybe you know you unluckily for you chose the wrong field to do something like this and because in many other fields i think this would have just you know been completely fine so in addition to let's say making videos

### One Step Up: Copying A Research Paper [23:30]

and you were making insane number of videos like two a week or three a week as you said and that certainly also you had a schedule that certainly must have also pressured you but then you also there is this there's the issue with your paper right and that to me that to me was really something where i thought this is someone who is almost like blinded by either the speed or the fame or as you said you felt infallible or something like this so for people who don't know you had written a number of research papers but this particular one you even made a video about it i think like i wrote a paper in a week or something like and uh it was about the neural qubit and one of your viewers then uh went public and claimed and could show that this was copied from largely from two other papers copied together the diagrams copied and the text copied and you changed some of the wording which was the most puzzling thing to me so instead of a quantum gate which is equivalent to a logic gate you changed it to a quantum door which makes no i like this is a meme until today right and instead of complex uh numbers or complex hilbert spaces i think it was uh complicated hilbert spaces which also is kind of if you so maybe if you just if you look back now what is your reaction now to past you in with respect to that paper yeah um yeah that was hilarious that's eternally a meme now um what i yeah i mean i used ai to generate some words and like make things different i what so this was automated the replacement yeah okay yeah yeah i think there's a tool called like um i think it's called like it's a web tool i forgot it's like ai writer or something like that you like paste in a paragraph and then it like rewrites it um yeah like what a super decision that was i but there i mean at this point it's really it's not it's not this it's not quite it's a step up from copying code and attributing someone at the bottom right because there you can still say you know i attributed them i'm you know i can sleep at night this is really i go i take paper i put it deliberately into a tool that rewords it and then i say here's my paper right this is what made you or how did you find yourself making that step that you know like they're really from i can justify this to myself to i guess i don't know what maybe you explain better than me yeah i you know it's just like ego it's like i'm untouchable and i can just do anything and i um i guess i didn't really understand what it's like before i plagiarized that paper i talked to an actual quantum researcher um who works at in santa barbara for google and um you know he's like we should write this you know i was paper together he's like yeah let's do it it's going to take a year and i remember thinking like that's way too long for me like i'm not doing that in a year i'm going to do this in three days and just thinking like you know i guess i didn't respect the scientific process enough to yeah it was just um to me i just thought of it as like a another link in the video description just adding it i should have just linked to the seven papers i just instead i put my name on it and just made it into one and like oh people are going to like me more because of this i'll have more credibility because of this instead of the opposite and i don't know it was just in general it's just you know really um drugged out honestly like that i don't know why i made a lot of decisions that i did um i'm so over now about the way yeah at no point it did it ever because that's the baffling thing to me a little bit and that that shows me or at least seems a little bit like someone who has really lost touch a bit is that when some when a it's like a an experienced researcher tells me it's gonna take a year to write a paper and sure if i think i'm fast i can i think i can do it in three months right but three days is a like is a different thing so clearly your idea was already you know i'm gonna take a shortcut it's not like i'm gonna write the same paper in three days it's just um how can i make a video out of this in the shortest possible time yeah i was like what's my next video i wrote a research paper and just thinking about that that's really the angle like i want to make a video that shows or tells people that i wrote a research paper yeah

### Was there another way? [29:15]

yeah so a lot of i've seen a lot of commentary saying things like you know it's a shame you have a good platform you're charismatic and you could have they say something along the lines of you might have just as well credited all these people and just had the same effect like implying you know there would be another way of doing this you could just say you know here is a bunch of code by some cool people i'm going to show you how it works uh and their implication is you would be just as famous liked and so on did you first of all do you think that's true and second of all did like or was it really your conviction no if i did that i would be way less popular i do think that that's true now i did not think that was true that i thought that i would have to be the guy with who is behind all of this in order for my brand and channel to grow because yeah because it's just hard like in the youtube game to like differentiate yourself and i felt like this was a way i could do that yeah i mean it's it's it is true right um i'm not sure that these people are correct like it's for sure good advice to credit the people whose work you present but i myself i'm not sure if they are correct when they say you would have been just as popular and and just as you know well respected by the people who think you really did do these things right i'm not sure as you say how youtube works is it's a it's a tough game and you at some point this all came and together also with your course which we can talk about in a second but specifically with respect to the code and to the paper you made an apology video which was fairly lengthy it was not your usual style it was just kind of you standing there you essentially said straightforwardly you know here's what i did i credit i didn't credit these people enough just took their code and so on and um then people noticed that only like a few days later in your next videos essentially you did the same thing like there were slides where you took from somewhere and so on is it i don't know is it fair to say and so you made these videos you made the apology videos then you immediately started uploading videos and before you really quit and you quit for a long time after that what was what were sort of the last videos like for you or you know like after let's say the apology video and so but before you quit what was that like you're asking about the time between when i quit to the apology video what that was like no from the apology video to the point where you it didn't upload for months after that or uploaded very infrequently was how did you feel at the point like of the apology video and a little after that yeah well i mean it felt pretty bad generally i'm a pretty happy guy as you can surma but um i can say that's the only time in my life where i've ever felt somewhat suicidal like uh just for a little bit and uh yeah i didn't know how to deal with that level of sadness so i tried about a bunch of different things like i um moved from la i uh got a dog i just i don't know did some soul searching some meditation just tried out a bunch of i tried virtual reality like escapism as well um it was a pretty tough time as you can imagine but in terms of like doing the same thing again i guess i did but i didn't think that i was like maybe there's something wrong with me like i just i don't know like i need i need some kind of mentor to be like here is how you credit people in a youtube video about machine learning and here is what people are going to find acceptable yeah did you think at some point maybe i can turn this around you know maybe i can because you were at the beginning when people brought these things up you were i saw just a bunch of twitter posts and so on sort of discrediting them denying them like no i never never did anything like this was there a point where you thought you know people are getting iffy maybe i can turn it around yeah there was um i mean i tried everything i was like maybe i don't need to apologize maybe i do that would make it better or worse maybe i should just deny didn't i deny like politicians do maybe i should you know make fun of you know make like uh um reply videos to other youtubers who made videos about me there's a lot of things that i thought i could do um eventually i decided and i don't even know if that was the best thing for my brand i know it was the right thing to do to make an apology video morally but i don't know if that actually helped me or hurt me i still don't know to this day um yeah was it so i think if i hear this a little bit out of you that there was a time where you were still mainly thinking brand mainly thinking you know which actions are gonna let me still reach like the million subscribers or continue on and then was there a particular point where you thought no actually you know let's do an apology let's tone it down um was there a time when you thought when you consciously let go maybe of the million subscriber goal there was i think it just came from introspection and seeing how like the um the amount of um i don't even know what you want to call it uh feedback negative feedback or um criticism he just wouldn't go away he was just there and it didn't really die down i thought i mean there's really nothing else i can do here i need to just accept defeat wave the white flag um part of my brand is just like you know super confidence and always being um okay with being um like haters or whatever not even yes but you know what i mean and like there was a point where i was like ah you know i'll just apologize and then i also felt you know near the end i did feel i started to feel like guilty because you know some people said that it wasn't just that i plagiarized but that i was actually doing the opposite of like accelerating um researching the space like this sets a bad example for people and this actually gets in the way of research and it's going to slow it down and that's what i was like okay that's if that's true that's really bad and honestly i like i was reading too many comments as well um but yeah i mean i still don't know to this day like whether or not um the apology video helped or hurt my brand in fact if i had to ben i would say probably hurt my brand but you know at least i felt better afterwards and i guess that's what mattered in the end yeah i mean i think few people really understand what it's like to get youtube comments on a on a bit of a scale and and people there will always be people criticizing and hating especially i guess you with very little credentials in the field i guess you have always had people saying you know this is a maybe this is a clown uh has no credentials what not and it didn't help that you copied code because then you not authoring the code also meant you knew less about the code which might also be sometimes shine through a bit in your videos but i think you with time you sort of learn to tune out the haters because you're gonna get them anyway but then sometimes they're right and i think it's i think you know i don't think and uh i don't think many people in the like public sphere get like have a good understanding of when should i listen to the bad comments and when not because usually it's no right so right yeah um so then this was this was very shortly people really complaining about

### Clickbait Course: Make Money with Machine Learning [39:00]

plagiarized code and uh this paper which was one of the sort of big points raised and then in a very sure like within a month or so there was also the issue of a course you offered right so you you maybe can you tell a bit how this course even came to be you made videos at an insane rate how did you think you could also offer a course and why yeah i think it comes down to two things one i felt like i could do more than what i actually was capable of doing because i my ego was so inflated at the time so i that's one the other is just looking at the metrics generally the videos that were about making money were the ones that did the best and so i started to follow that trend and tailor my content in that direction as opposed to what i would have done years ago which is like how do we solve them you know millennium problems like poverty reduction and water cleanliness and um environmental sustainability things that you know actually matter the course was around that like well if people want to make money let me make a course around making money with machine learning that was what is called right it was called make money with machine learning literally that is a hell of a clickbait yeah like the most clickbaity exactly what's going to get the views title and it was supposed to be a paid course it was i think about 200 per student and the issue the first issue was that you claimed it was like a limited entry course with personal supervision now both of these things didn't really turn out to be accurate as you promised so there was an issue of you said i only let in 500 people but then you let in twice 500 people so you had two different slack work workspaces with twice the five some i think one even had 700 but there's a few extra ones i guess and then also there was apparently not really like you can't personally supervise a thousand two hundred like it's impossible did you plan on these things already or did they just sort of how did they happen i didn't plan on them i did think that i would have 500 um but when i put the course out there were so many sign ups so fast and i got greedy i was like i'm just gonna let this keep on going let's see how many people can sign up for this and i thought yeah i can just have two different uh cohorts and um you know i had people volunteer to help at the time you helped me like as i guess you called them teaching assistants and yeah but they how many roughly how many tas did you have do you remember um there was at least one there might have been written that there's at least one yeah but they sort of did they quit after a while or did they stick with you well no they actually they were amazing they stuck the whole time yeah okay but they were volunteers yeah okay so it was 200 bucks and like one two three maybe volunteer tas for a thousand two hundred students and you did you plan on ramp did you realize at some point i can't provide personal feedback to all of these students or did you just think you know whatever i'll just i can do this or i did realize i was in over my head i think it was like week two or week three that it really started to dawn on me um and then i think it was week four that some of the students started you're going to social media um and then everything came crashing down in the middle of the course and then i had to give out a bunch of refunds but still had to finish the course to the end it was a 10-week course so we still have to keep going for five weeks after that um but yeah i mean there were still you know hundreds of students who stayed in the course i don't know that like the register made an article on this but they didn't say like it's not like everybody just dropped out all of a sudden yeah still people in the course i still had some responsibility yeah so i maybe briefly summarize these articles and you know they're written from a certain angle right and uh that's exactly why i also wanted to get your just your side of this story so these articles they claim for example that you know people started noticing there was no personal supervision they complained um you you never essentially showed up in the slack workspaces well you know or infrequently they all got the same feedback on their exercise so that was the sort of like a copy paste of like good job um it was like that then people started demanding refunds but were some claim they were even banned like for demanding refunds then it was also claimed that you eventually said there was a refund period which was for 14 days but the article claimed you quietly introduced the refund period 30 days after the course started so it was essentially impossible for anyone to have known because there was no refund policy at the beginning you introduced a 14-day refund period 30 days after the co the course started you then and then you know once people discovered that there were two different cohorts and so on or how what of these articles is is true and what is overdone um so there are also several tweets of students that said yeah people claiming refunds were banned um or or that the fact that you introduced this refund period how did this go down from your perspective so paul that is true um what i don't i think was overblown is the banning part i never personally banned anybody um but i can't speak to whether or not one of the tas may or may not have done that i love you but yeah everything else like definitely um on point like it's all part of the the story yeah can't refute any of that yeah and did you get did you get scared at any point or did you were you still in this you because all of a sudden people and their money are involved right it's not i mean 200 bucks is not that much for maybe an american but it is a lot for maybe someone in india or something you know at some place like this did you get at some point you know scared because like wow there's actual money here that i may have to pay back or yeah i mean i got scared for a lot of reasons i was scared that um yeah i would like have to go through some kind of lawsuits people were saying like oh i'm gonna there's gonna be a lawsuit you're lucky you're not in jail and stuff and um yeah about the refund stuff like that 30 day versus sneaking it in and i'm sure i did that i honestly don't remember it now like i'm sure like that's probably what happened but i mean when i look at it now i'm like i mean when you charge money you need to be very upfront with people and like that's how you make a sustainable product i wasn't thinking very sustainably and long-term it was a very short-term thing um but i was scared yeah i was here did you but your thought was still i can educate these people even if i can't give them personal supervision or was it all like you know like i'm gonna get their 200 bucks i'm gonna tell them something so they can't complain or did you still think you know i can't like the course has value for the people who are in it no i did think the course had value i mean it's weird because it's like i'm conflating my bias against academia and the traditional learning path with this course that is yeah it's got a super clickbait title but you know i guess i didn't fully appreciate what online learning and i'm still learning what online learning really can be in the future i thought well you know you don't need to be in a physical classroom to learn like i think we can all agree to that now like you can watch videos online but also you know what is um personal supervision and does there need to be x y and z for someone to be able to say i learned a lot of learning comes from self motivation and um you know education is not a scarce resource it's abundant it's the desire to learn that is scarce and perhaps that alone i felt justified like if i could get them to want to learn these things that would be enough um at the time i felt that way now i know like what would i change differently besides the obvious part like the 30 day refund from the start is to just hire help like if i were to give advice to anybody doing anything like this like any youtuber who wants to make a course like hire help step one hire help then figure everything else out don't plan it out yourself it's too big at scale for one person to do what happened did you end up giving refunds to people or i did did you still have enough money to give refunds um i yeah i did what happened to the money like i can imagine you get 200 bucks a thousand people that's like 200k um how where did that go did you end up plus or minus or did you spend on refunds did any lawsuit result or there were no lawsuits everybody who wanted a refund got a refund there were still a bunch of students who completed the course to the end like and i'm very thankful like despite all the drama they were loyal to the thing and so was it wasn't negative it was positive it wasn't nearly like probably like 10 percent what i mean at the start

### Rock Bottom and the Way Forward [50:30]

and then and then you know i think this as i said this was within like a month of of every everything down you were making lots videos the paper the course all at the same time and then everything comes crashing and i think it's one thing when you feel bad because life is crap right because something happened to you that's bad and you know but it's an entirely different thing when you're you know you're responsible for it right like that is worse that is like my life is bad and i'm to blame and it you know like it's my doing right like was this i guess this was your experience right you know whether you thought it was good or bad it was like my life is crap and i'm responsible how did you what did you do at that point you said bit of soul searching and so on how did you decide to go forward um so i moved back to san francisco um i was there for a few months i basically invested in my friends and family talked to them that helped um got really into virtual reality that helped as well like disassociating from this reality bring it to a virtual world um where i was anonymous and uh logged off of all social media as well so that helped as well and kind of just gave up with the whole like you know million subscriber path that i was on and what else yeah just um oh yeah focus on my health as well like i was like i'm just gonna like try to focus on being healthy because i can control that i can't control what people think about i can control my health so that helps you made a you made a quite astounding uh body fitness transformation as well you were at the end you were like in 2019 when it all crashed you were kind of a like a chubster yeah like right and i saw like a before after picture was this a conscious effort by you or it was yeah because like part of like what you know having a desire to live is to like be able to look in the mirror and you know say like for me at least like hey this is an attractive guy so that you know it's kind of vain but it definitely helped for sure like that yeah and so you eventually you got let's say back up on your feet after all of this what was your or what is your current plan or what are you doing right now you've posted a few videos again here and there but um i'm not so maybe you know what's what are you doing essentially so um yeah making videos along this series called alpha care about health care and ai which has kind of always been like my the industry i'm most excited about for ai like applicability like oh we can make people healthier so i'm doing that i'm almost done with the book i've been writing for the past three months um which it's gonna be a free e-book i'm not gonna charge for it um so that's been interesting that's also on like uh deep learning for healthcare apps for begetters a bunch of examples in there and once i release that all of this will be done in like three weeks probably from now um like the siri the video series in the book then i have to figure out what the next thing i'm going to do is um what i'm most excited about currently is paying people to be healthy um there's this app called sweat coin it's out of the united kingdom it pays people in cryptocurrency to walk i find that really interesting because you know two of the most meaningful things to me are um keeping people healthy and reducing poverty and this kind of does both at the same time so i'm wondering if there's a way to create what's called a dao a distributed autonomous organization around um health care and health data and keeping people healthy paying them somehow with cryptocurrency to stay healthy i just use a service called inside tracker which cost me like 500 bucks way too expensive a service for most people to use um but i got a blood test done two weeks ago using the service they took 43 biomarkers of mine and that now i have a bunch of health data like my cholesterol level is apparently way too high because i way too much red meat um so i've got to cut down on that but something like this if we could turn into um like a free service that keeps people healthy and actually not just free but pay them money and then somehow turn it into a business where also the service makes money that'd be really cool so i'm kind of like thinking like i'm going to start some kind of company around that or a dow i should say i'm not exactly sure what it looks like though i mean there this is happening in part already with i don't know we have like high taxes on cigarettes right so essentially the smokers they finance a little bit the non-smokers via taxes some health insurances they already give discounts if you do like regularly go to it to a gym or something so i'm like something like this is definitely in the realm of possibilities now with respect to cryptocurrency is this a meme or was there actually a siraj coin at some point i haven't found anything like what was that yeah that was a real thing i launched a cryptocurrency i think two years ago or something three i don't know uh called siraj floyd and uh people really didn't like it so i took down the video i'm like there's still you could find it if you really search siraj coin okay but it was just it was more like for a video or did you think you know maybe i could make some money with launching my own cryptocurrency yeah both i mean this was at the height of the um ico crate yeah and everybody was doing it and i felt like wow i'm gonna do it too here we go that's right and the idea was that you can with siraj coin you can uh get a meeting like buy a meeting with me or like make a music video with me just you know i am the scarce resource like in these cryptos there is a scarce resource created token the token is how you access the scarce resource yeah and uh yeah i mean i'm glad i did it still like nobody got hurt from that it was just like a fun experiment and i learned a lot from it as well like i still think it's an interesting idea like i do think that we're gonna see more individuals create tokens around themselves and um yeah i mean yeah a couple of nfts work this way right that there's some kind of like a meeting with a famous person tagged onto it or something like this yeah so with respect to your book and your new set of videos and you know i guess the question everyone asks is there still how do you handle citations plagiarism things like this are you toning it down or are you like extra super duper careful or what is your sort of how do you approach this topic i guess you're in a bit of a special situation not only are you held to the same standards but now you know people read your name they're probably the first thing they do is put something into a plagiarism checker yeah i'm super careful i put it in the video description not just like the github i say it verbally um yeah i just try to be more careful um yeah and the what's the book about can you is there is it something you can disclose already or yeah it's on bioinformatics for beginners i'm also a beginner to bio informatics i'm really interested in multicomics like all the homics genomics hepa genomics transcriptomics and just thinking about how we can integrate all of these different types of data to make both diagnostic and prognostic predictions for people and i think that's the future i'm really interested in reversing the aging process um david sinclair pat harbert has a great book on this called why we age and why we don't have to he has a podcast that he's going to release next year on this topic and i just think that there's a great space for data science and data analyst enthusiasts to make a contribution in this field because i do think the future of health care isn't going to be targeting individual diseases like alzheimer's or heart disease but rather that this the disease that is upstream of everything else aging itself that's a i mean it's a tough task um but yeah it's a i guess it's a cool cool outlook i it seems like a little bit of a rebirth it you know you told how you were at the beginning of your video career thinking if i could just you know make video about these cool topics and so on and uh it almost feels or at least to me it sounds like it's got a little bit of that same spirit again i'd like to think so i mean i don't have the same i don't know level of or maybe i just feel this way i don't have the same like energy that i did back then um where it's just like a i have to do this or else like the world is gonna end like that level of conviction i just feel like hey i mean i'm really interested in biology in general i don't think i'm gonna get i honestly don't think this is gonna give me the level of fame or opportunity that talking about deep learning from 316 to 2020 did it's just something i'm interested in uh and i'm okay like not reaching a million up i mean probably never going to reach a million subscribers i just want to be interested in this and even if and you know if this like company doesn't work out i'm happy to like take a job somewhere and just like learn about bioinformatics full-time as a bioinformatician halos or something um yeah well in yeah i mean in many ways i i've

### Advice for Future Generations [1:01:30]

told you that this privately but in many ways you were you're sort of with all of this happening you were still sort of a the pioneer of what many of us other ml youtubers essentially that the path we go is you made it kind of like i remember when i started making videos there was like nothing and when you started there must have been like really nothing right and um you know that for for all the things i think it took balls to go that way and you certainly hustled even if it led in into like a wrong direction um do you have i don't know do you have because i know there are quite a number of people who look at maybe you also me other youtubers a lot of people are starting their podcasts nowadays also start channels like mine or similar to mine any advice you have for people starting out in the in the sphere of online education or what might what we might call being an influencer anything like this um yeah i would say that you this is not something you do as a side job like a lot of people you know kind of have to because they need a source of income from their day job but um i would say like the only way to be successful in this is to pick hits to be your one fate and do that all day and it's got to feel like play to you but it's got to look like work to other people like to me this whole time i've just been playing like really enjoying myself like it's not work and that's honestly why i think i grew as much as i did i genuinely enjoy the topics video production process editing lighting thinking about metrics all of that stuff just felt like play to me and that's how you're going to be successful it's not going to be if you feel like it's hard work um you should pivot or think of some other content to talk about or maybe a different medium like you know i had a podcast as well um i did i think five interviews and then i stopped because it didn't feel like play to me like i don't actually yeah for some reason i just don't enjoy being a podcast host like i enjoy monologues and that kind of thing so i stopped um whereas someone like you or you know joe rogan or other podcasters they actually enjoy it so they're gonna they're actually gonna be successful so that's my best advice is like make sure that it feels like play to you and then i you will be you'll probably be successful and when someone finds themselves a bit successful and finds themselves to be sucked and drawn by the metrics by the cloud by because i already said it but i'm gonna say it again like this is a thing i feel it i like other youtubers feel it for sure this suck it's like a thing drawing you right and you know leading to the kinds of decisions you made and uh and uh what is do you have any i don't know you know other than don't do it do you have any you know best the mindset that creates in a person do you have any maybe recognition of what could help someone to get out of it or to resist or you know what do you tell yourself when there's like a really easy opportunity to get a lot of views or clicks i would say the best thing you can do is google sir roger ball and happen to this guy and uh yeah just be afraid you don't want that to happen to you for sure luckily it happened to me first so you've got an example in front of you now of what can go wrong when you follow views and likes too much you chase clout too much in the education space um the internet gives everybody a voice you will be held accountable um there is no um we are moving into a world that is much more transparent every day less and less privacy um yeah the internet gives everybody a voice and power so um yeah that's what i can say use it wisely i guess use it wisely well siraj raval this was a pleasure really truly i i thank you very much for being here with me today um thanks for coming on thanks for being so open and forward and and honest i think it's very valuable the world also hears from you and you know in it not just from articles and and you know reviews and things like this absolutely thank you yannick awesome