"Don’t Learn to Code" Is WRONG | GitHub CEO Thomas Dohmke
Resumo
O vídeo apresenta a perspectiva de Thomas, CEO do GitHub, sobre a importância do ensino de programação para crianças. Ele argumenta que, assim como disciplinas fundamentais como física e matemática, a programação deve ser ensinada nas escolas, já que o software domina nosso cotidiano. Thomas destaca que entender como criar ou modificar software é essencial para não ficarmos apenas como consumidores passivos da tecnologia.
Thomas compartilha sua trajetória profissional, começando pelo interesse em tecnologia após a queda do muro de Berlim, que levou à compra de seu primeiro computador e uma carreira inicial na indústria automotiva. Posteriormente, ele cofundou a Hockey App, uma plataforma para desenvolvedores de aplicativos móveis, que foi adquirida pela Microsoft. Esse percurso o levou ao cargo de CEO do GitHub, onde ele utiliza suas habilidades de desenvolvimento para liderar a maior plataforma de desenvolvedores do mundo.
Ele também discute o impacto da inteligência artificial (IA) no desenvolvimento de software, afirmando que a IA torna o aprendizado de programação mais acessível e aumenta a produtividade dos desenvolvedores. Thomas vê a IA como uma aliada para realizar ideias rapidamente, apesar de acreditar que ainda estamos longe de substituir a complexidade do desenvolvimento humano com agentes de IA. Ele destaca que, embora a IA facilite muitos aspectos do desenvolvimento, a necessidade de engenheiros qualificados para tomar decisões complexas ainda é vital.
Por fim, Thomas reflete sobre a cultura remota do GitHub e o entusiasmo contínuo dos desenvolvedores pela plataforma. Ele menciona como ferramentas de IA, como o GitHub Copilot, têm transformado sua rotina, permitindo maior eficiência em tarefas diárias e aprendizado contínuo. Thomas enfatiza a importância de nunca parar de aprender e de explorar novas tecnologias para se manter relevante no campo do desenvolvimento de software.
Chat
Entre para conversar com o vídeo.
EntrarPrincipais Pontos
Thomas, CEO do GitHub, defende o ensino de programação nas escolas como essencial para o entendimento da tecnologia.
A trajetória profissional de Thomas começou após a queda do muro de Berlim e inclui a cofundação da Hockey App.
Thomas acredita que a inteligência artificial facilita o aprendizado de programação e aumenta a produtividade dos desenvolvedores.
Ele ressalta a importância de engenheiros qualificados para decisões complexas, apesar do avanço da IA.
Thomas menciona como o GitHub Copilot transformou sua rotina de trabalho e destaca a importância do aprendizado contínuo.
Mindmap
Palavras-chaves
Transcrição
i first of all I strongly believe that every kid uh every child should learn coding We should actually teach them coding in school in the same way that we teach them physics and geography and uh you know literacy and math and whatn not Those are all fundamental skills Coding is one such skill because software is everywhere and in fact our day-to-day is dominated by software already You can't really live your life travel um wake up in the morning without software anymore And I think as as humans it is crucial to not only be in readonly mode but also be able to create uh things ourselves or at least understand how how creation is done on these devices My name is Thomas and I'm a developer I've been developing uh software since the early 1990s and today my role is mostly being the GitHub CEO leading uh the largest developer platform uh on this planet I was born in uh 1978 in East Berlin when Germany was divided into two countries So I was on the on the eastern side uh in Berlin in in a suburb called Matsan I grew you know had a normal childhood um but was fascinated by technology remote control cars uh uh you know little computer games Uh in 1989 the wall fell and and that opened up you know a whole new world to me as an 11year-old um obviously from from a toy perspective um getting access to to Lego and and Mickey Mouse but also uh to uh to computers and so in the early 90s um I bought my first computer uh 64 and after um university um I went to technical university in Berlin I actually uh went into the automotive industry uh I started working at uh Mercedes back then it was called Dameler Chrysler working on driver assistance system uh for for the S-Class and I did that uh for a while and then switched to the supplier side to Bosch uh uh which is you know a large automotive supplier amongst amongst other things and and there uh worked you know on parking systems and it was in 2008 uh that two things happened One is I finished my PhD thesis but also Steve Jobs showed the iPhone SDK Everybody wanted to build apps um and and at the time only Apple could build native apps and I thought this is this is cool I got to you know build apps myself and I got to get into that space And so I quit my job at Bosch uh at the height of the financial crisis amongst all things uh in late 2008 and um reconnected with a friend um from from university and um we just became you know two freelance developers building apps uh for the German market So mostly agency work or or subcontractor for for larger app projects I think in in 2009 and 10 we built around 30 apps That work on building apps um for for for mostly German customers um German enterprise customers media and and automotive and so on got us the idea of of building a platform for mobile app developers And so we um together with three friends Stefan Andreas and and Michael we founded a company called Hockey App which was a platform for mobile app developers you could distribute your beta builds and um collect uh error reports crash reports and feedback and those kind of things And so we built that startup you know the platform for our own you know freelance business because we had that painoint that we wanted to solve uh for ourselves Um because it was before that you would send you know a build uh over email you know to to a project manager um in Berlin or so and then the project manager would take that build and send it to the customer and explain to them or you have to you know drag and drop that into iTunes and connect you know uh uh the dock cable and and all these kind of things since we made that way easier which made our own life easier and I think those are some of the greatest startup ideas when you're using it yourself uh uh day in day out and then improving it um based on the feedback that you're getting from your own customers So we basically started building out the the hockey app business while we had the the contracting business and they were companies um both about the same size and um when Microsoft came in in 2014 we thought they would only buy the hockey air platform and the product business and and leave the contracting business alone uh given there were existing customer uh contracts and and what have you uh but in fact Microsoft bought both companies uh the the subsidiary the product company and the the mothership the the contracting company because the contracting company had a number of iOS and Android developers uh that were a hot commodity uh in 2014 and and hard to hire um certainly you know at at startup rates and so Microsoft actually took over both of those companies and uh a fun funny story is today of these 11 employees seven including myself work for GitHub and so different paths Um uh they all got to Microsoft and then from Microsoft into GitHub You know coming from a small company of of our size based in Germany in Stutka Germany southwest Germany Uh moving you know halfway around the world I moved with my wife and back then two very young kids uh to Seattle that alone was a big change And looking back now um 10 years later and we moved we moved in in early 2015 it feels all like a blur like things moved so fast and sometimes I don't even realize how we did all that Our startup was very small and bootstrapped um for the whole time So we never took any outside investment uh until Microsoft acquired us and at that point in time I think it was about 11 or or 12 employees a still very small team all of all of engineers you know building building the product together and in some ways you know my role today as GitHub CEO um is very similar uh in that you know um my developer skills uh my understanding of code my um you know empathy for for how software developers work helped me both with my you know internal team of about a thousand uh engineers and with our customer base which are also all software developers or those that aspire to become software developer And so I think you know a lot of the you know my passion for software development is actually perfect you know for me being being the GitHub CEO I don't think I've seen anything more exciting and and and changing how uh we think about software development in my in my 30 plus year career as a software developer You know when I started coding in the early 1990s uh there wasn't even the internet or I I certainly had no internet access and so I had to figure it all out all out all by myself with with books with magazines um going to computer club in the community center kind of hoping that somebody will be there Um if we fast forward to where we are now is that it's so much easier to get into into software development you can you know just write a prompt into Copilot or or Chat GPT or or similar tools and it uh will likely write you um um you know a basic web page um or a small uh application a game uh in Python And so AI makes software development uh so much more accessible um for anyone who wants to learn coding And on the other side of the spectrum it makes developers so much more productive Most uh developers that uh work on a project have way uh too much work to do uh they have long backlogs um of their own ideas of customer feedback you know uh things that they're hearing from their managers or from the market or what they're seeing at competitors And so almost any software project that has a certain you know age um has way too much work on the innovation side but they also have what we call technical debt Uh you know legacy code things that have you know been created over over months or years that need a cleanup that need what we call refactoring And so engineers constantly balance those two backlogs So having something that brings the effort down and makes them you know 10% 20% maybe even 50% more productive is is completely changing uh how software developers work the role of GitHub you know in this in the first uh let's say you know five years of the age of AI um given that we started working on GitHub copilot in uh June 2020 right after uh GP3 uh uh was was first shown to the world is that we want to be you know on the forefront um of uh AI code generation we want to provide um tools to developer to be more productive and more happy uh when when writing code because the reality is the dream I think for most developers that start their their journey um uh as a programmer is that they have an idea in their head and they're trying to find a way how they can can get the fastest from that idea to an app or or web page or service right like the challenge is not that developers uh uh don't have enough ideas The challenge is that you take that big idea and you have to break it down into small building blocks And as you're working on the first block or the first module or the first class or uh micros service whatever it is you're you're realizing that this this idea that you have is so much more complex to implement than than you thought And so from what became a weekend project it becomes a month-long or sometimes year-long project And many you know apps that uh I wrote as a teenager and I know you know many of my friends and and my employees wrote never get anywhere because you you ultimately realize it's much more complex than you thought and it's not worth spending the time on it And in in the world that we live today you can always you know download an app from the app store or find you know some online service that does the same thing So I think AI helps us you know to realize the dream of taking an idea and implementing it much faster And you see um some of the early signs of that where very small startups sometimes you know five developers um and some of them actually only one developers uh uh believe they can become million uh if not billion dollar businesses by leveraging all the AI agents uh that are available to them and maybe building their own uh to to write code uh to write software much faster Now the flip side of that is that you know uh I don't think we're anywhere close to a world where you can just write a single prompt and say build GitHub and then an AI agent builds all of the features of GitHub um or even just the very basic primitives like repository storage you know git storage and uh issue tracking um because the the decisions um that we as developers as engineers as product managers have to make to build a complex system like GitHub you know thousands if not tens of thousands decisions um there's the simple ones right like which programming language which which open source framework which cloud to use or do we even use a cloud which operating system and and so on But there's the much more complex decisions of how you architect the system you know are you building a monolith and or you building microservices and getting to a point where agents can make all these decisions and write an app that actually is a viable business you know finds product market fit has a great user experience and and ultimately generates both revenue and profit uh because any business at some point has to get to the place where they're making profit and and return that profit to the to the founders or shareholders that I think we're we're quite far away and so we need uh engineers uh to do engineering stuff They they need to exercise their craft and uh apply systems thinking and design and and and build really great applications I think the unique thing about GitHub is its size and both the love uh that developers have for you know our brand um for our mascot um the octoat or we call it internally Mona um and um the reputation uh that GitHub has created for itself um since the very early days since the I remember the early launch of GitHub and and meeting um or seeing um Chris one of the founders speaking at RailsCon in in Las Vegas in in 2009 and then signing up for my own account and and started using it I I was excited about using GitHub and now we are in in 2025 and there are still many people that uh love GitHub but you know what what comes with love is also that you're not holding back your criticism and that we have 150 million users and uh on the platform and so there's at least you know a million opinions of what are the things we should invest on and what's working well and and what's not working well and what's the one feature that is important you know to that set of users while it's not important uh to to me and and my product leadership team because we have you know our own uh strategy and decisions to make So filtering out the signal from the noise and I don't mean noise in any negative way just so much feedback that we're getting I remember when we when we did the acquisition in 2018 and I joined GitHub and afterwards um we sent an email to 10 GitHub users and say hey we we're looking into a new project we would love your feedback I think we got nine responses of exciting users saying if you want to I want to provide feedback If you do that in many other companies and startups you get one response and the one response is kind of like well I have like 10 minutes time to to give you some feedback There's just so much um uh uh information that's that's coming back to us um on on social media and our platform in email and support tickets and so on The second piece you know that comes to mind is that GitHub um for the longest time has been a company with a very strong remote culture like long before co uh the GitHub founders started hiring um developers uh sales folks support folks all over the world and today I think we're one of the largest remote only companies where all work you know from our homes or from you know hotel in in soul um or from from wherever we are and so a lot of um our culture is focused around GitHub as a platform form which obviously through open source encourages um uh asynchronous uh collaboration uh and then tools like Slack um um and and video video calls um uh that we use much more heavily internally than we're using email like like old school uh companies do And so when I wake up in the morning especially here you know in Seoul uh which is like lots of time zones away uh from from the US uh where about 80% of our employee population is is that I wake up to 30 40 Slack messages plus you know uh uh hundreds of channels uh with with conversation and then figuring out you know what is actually important for me as CEO what what to react to what can I snooze uh for a while and what can I just ignore that that's the you know a big part of my job but it's also so exciting because you know I can be here in South Korea um at at this event and and still still run the company And a lot of you know what we do on a day-to-day basis doesn't actually matter whether I'm here in Seoul or whether I'm in Berlin or whether in I'm in New York or anywhere else in the world And and I think that's for for many hubbers how we call our employees A really strong part of our culture that we are remote company Uh it's not related to the pandemic It's a choice that we made of how we want to run the company you know how we select GitHub as an employer and ultimately how we believe we can be successful I first of all I strongly believe that every kid uh every child should learn coding and we should actually teach them coding in school in the same way that we teach them physics and geography and uh you know literacy and math and whatnot Those are all fundamental skills Coding is one such skill and it just has taken us too long to actually realize that because software is everywhere Um hardware is also everywhere We carry we carry both software and hardware with us uh through our day And in fact our day-to-day is dominated by software already You can't really uh live your life uh travel um wake up in the morning without software anymore And I think as as humans it is crucial to not only be in read only mode but also be able to create uh things ourselves or at least understand how how creation is done on these devices That doesn't mean that every you know 18 19year-old when they leave high school become a software developer In the same way that not every uh uh kid that learns physics or or chemistry in school becomes a physicist right just because you learn those fundamental skills doesn't mean that you decide for yourself that that's the career path to take So that's number one You you got to you got to learn coding I think number number two is you got to use AI to do that And and you know whether it's here in Korea or in Germany uh most kids um uh and in fact you know most most people don't speak fluent English um which is the primary language of software development And so it democratizes access to technology and that's true for many other uh things in the world And so having uh an agent available uh that answers you any question but also lets you you know realize your dream building your dream uh is incredibly exciting And then the third thing you know for anyone who is already a software developer or wants to you know develop their craft is you got to keep you got to keep rehearsing You got to keep training Yeah You got to keep learning There's you're never done with learning If I look back 30 years of what development looked like back then and what it looks like now um I would have been you know um very behind if I hadn't constantly read blog posts um uh literature and and tried out things myself So I think those are uh as as crucial as they were in the ' 90s they are still crucial in 2025 You just have so much more access to information uh to become you know top of the field The obvious answer is that I most enjoy using GitHub Copi right like that's our product Uh it's our baby Uh we're working on this day in day out I often you know uh see features long before the world sees that And and the flip side is also true that I often don't actually know what's shipped versus what is you know in in preview or in in internal ships And so I'm daily excited about what we're building there and I'm using a lot of that myself as I am at heart a developer And sometimes it's very simple you know asking it to to write me a quick script that that downloads you know ids of all our repositories um from our API And in the past I would have gone you know to our API documentation and figure that out all myself and probably would have taken me like half an hour to to get to you know a shell script that does that Today I just ask Copil and it writes me the script and it works within within minutes Um and I think actually that's the one of the true superpowers of AI Whether that's learning to code or exploring the world you have um an assistant available to you that has infinite patience You know it doesn't judge you Um it will you know chat GPT or co-pilot never tells you what what a stupid question is It always gives you an answer and it even accepts when you tell it it's wrong or it needs to you know explore the topic a little bit further You know you have seen uh the prompt examples where by telling it to outline its thought process it actually gets to a better answer I love using it for my blog posts and PowerPoint presentation to just uh generate some images and and play with that Um I'm I'm really bad at using Photoshop and and and drawing myself but I'm I'm really creative and I can write prompts and and figure out how to rewrite the prompt to make the image look more than uh more more closely to what I had in my head There's you know tools that we're using like teams copilot to summarize meetings especially when I I'm on business trips like that Um where I miss a lot of the meetings that happen uh in on the west coast and west coast time zone Just getting a summary figuring out other action items for me Uh same for summarizing emails Um uh you know using something like reclaim AI to to manage my calendar Those things are making me more productive And I think the really exciting thing is that there's always a new tool to try out and see is that how far along that journey is AI and how much more do we still have to do as an industry to actually get to that to that dream of of having an orchestra of agent uh that that we're controlling during our personal and in our professional lives
