As
AI tools become increasingly powerful, enabling everyone to quickly generate complete applications, websites, or data analysis processes, traditional manual programming logic seems to be becoming redundant? Are future engineers more like curators and editors, rather than keyboard laborers?
Similar discussions have been increasingly prevalent online recently. In this regard, Thomas Dohmke, the current CEO of GitHub, the world's largest developer platform, recently gave an interview sharing his views on programming education, how AI is reshaping the future of software development, and emphasizing the critical role of engineers. Here are the key points compiled by BlockTempo:
Dohmke Background Introduction
As GitHub's CEO, Thomas Dohmke's programming journey began in the early 1990s when he was a technology-obsessed teenager in East Berlin. After the fall of the Berlin Wall, he was exposed to computers and, after graduating from university, entered the automotive industry, developing driver assistance systems for Mercedes-Benz.
In 2008, inspired by the emergence of the iPhone SDK, he resigned and co-founded HockeyApp with friends, a platform providing beta distribution and error reporting for mobile application developers. This startup, with only about 11 employees and never receiving external investment, was acquired by Microsoft in 2014. Interestingly, of the 11 employees at HockeyApp at the time, including Dohmke, 7 are still working at GitHub today.
Dohmke believes that his development background and empathy for engineers' work are crucial to leading approximately 1,000 engineers internally at GitHub and serving 150 million developer users globally.
AI: Developers' Super Booster
Dohmke emphasizes that AI is the most transformative technology he has seen in his over 30-year developer career. He recalls that in the early 90s, programming learning resources were scarce, requiring exploration through books, magazines, and communities. Today, AI tools like GitHub Copilot or ChatGPT have significantly lowered the entry barrier.
"You can directly input a prompt to Copilot or ChatGPT, and it will likely write a basic webpage or a small Python application for you,"
AI not only makes learning programming easier but also achieves technological democratization, especially for non-native English speakers. On the other hand, AI has significantly improved the productivity of existing developers, expected to bring 10% to 50% efficiency improvements, helping them address backlogs, technical debt, and accelerate innovation.
"Almost all software projects with a certain history have too much work to do," Dohmke points out, "AI tools can alleviate the burden and fundamentally change developers' work methods." GitHub has been developing Copilot since GPT-3 emerged in June 2020, striving to be a leader in AI code generation, with the goal of helping developers write code more efficiently and happily.
Future of Human Engineers: Creativity and Systems Thinking
Despite AI's powerful capabilities, Dohmke does not believe engineers will be replaced. "I believe we are far from a world where inputting a command can make AI build a complex system like GitHub."
He explains that constructing complex systems requires thousands of decisions, from choosing programming languages, frameworks, and cloud platforms to more complex system architecture design, all of which require human engineers' professional judgment, systems thinking, design abilities, and creativity.
"We need engineers to do engineering work," Dohmke emphasizes, "They need to apply their craft, systems thinking, and design to build truly excellent applications."
He views AI as a powerful assistant that can more quickly transform developers' ideas into reality, even enabling small startups with just a few or even a single person to build businesses worth millions or billions of dollars with AI's help. However, the ultimate product-market fit, user experience, and business model success still depend on human intelligence and decision-making.
Advice for Learners and Developers
Dohmke firmly believes that programming, like physics, geography, and mathematics, is a fundamental skill that should be taught in schools. "Software is everywhere, and our daily lives are already dominated by software," he believes understanding how software works is crucial.
He offers three pieces of advice to learners and developers: first, learn programming; second, use AI as a learning and development tool to overcome barriers like language; third, continue learning and practicing.
"Learning is endless," he states, "Compared to the 90s, there are now more diverse channels for obtaining information, and continuous improvement is key to staying ahead." He practices what he preaches, using GitHub Copilot daily to write scripts, generate images with AI, summarize meetings and emails, and is eager to try various emerging AI tools, viewing them as infinitely patient personal assistants.
