How to be become a a devleoper in the age of AI - Github CEO

he rapid rise of AI is reshaping the tech industry, sparking both excitement and apprehension among professionals.

In a recent discussion, GitHub CEO Thomas Dohmke shared his insights on how AI is transforming programming, the evolving roles of developers, and strategies for thriving in a tech landscape increasingly influenced by AI.

Below, we explore his perspectives on whether AI will replace developers, how to effectively leverage AI tools, and the best ways to future-proof a career in tech.

Will AI Take Developers' Jobs?

One of the most pressing questions in tech today is whether AI will render software developers obsolete, particularly as non-technical roles like product managers and designers begin using AI tools. Dohmke offers a nuanced view, emphasizing collaboration over replacement.

“I think the roles in a company will start overlapping more,” he says, envisioning a future where product managers, designers, and engineers work more fluidly together.

He describes a traditional “EPD” (engineering, product, design) squad—typically eight engineers, one product manager, and one designer—with distinct roles. However, AI is blurring these lines.

“The product manager might use something like a coding agent or agent mode to write a quick prototype… or the designer takes a Figma [design] and helps the product manager to implement some of that without distracting the engineers,” Dohmke explains.

Despite this overlap, he believes specialization remains critical. “There’s still specialization in each of these roles… a product manager by design needs to be interested in talking to customers… a designer obviously has a lot of skills that they learn in university… that are very different to what an engineer learns, understanding code, designing complex systems,” he notes.

For developers, understanding code is non-negotiable.

“It’s crucial for every professional software developer to understand code because… you can vibe code stuff without ever looking at the code… But as soon as you get into really complex scalable systems… you got to verify that code… whether the code is actually efficient, secure,”

Dohmke stresses, highlighting the importance of technical expertise to ensure systems are robust and secure.

How to Vibe Code Effectively with AI

“Vibe coding”—rapid, creative coding often for prototyping—has become a popular use case for AI tools. Dohmke sees it as a powerful way to spark creativity but cautions against using it for production systems without rigorous oversight.

“Vibe coding is great for the creative part of the job… when you have an idea in the morning and you want to quickly see if you can build a prototype and maybe in the evening… you have an app on your phone you can show all your friends,” he says, drawing from his own weekend hobby projects.

However, he warns, “You do that knowing that never becomes a product on that codebase that you just hacked together.”

For developers using AI tools, Dohmke emphasizes the importance of systems thinking and architectural decision-making.

“What we expect our systems engineer or software developers to do is… they need to be able to take a really big problem and break it down into smaller building blocks,” he explains.

While AI can suggest solutions, such as database or caching strategies, “Ultimately you need to make that design decision… you need to understand is that actually the right decision for my business.”

This blend of AI assistance and human judgment ensures that developers remain in control, using AI to enhance rather than replace their expertise.

Overcoming Resistance to AI in Development

Many experienced developers remain skeptical about AI, often due to early models’ limitations, such as hallucinations or unreliable outputs. Dohmke acknowledges this resistance but sees it waning as developers experience AI’s benefits.

“We saw a lot of skeptics in the early days… people were skeptical about just auto completions with Copilot… then three weeks later… they’re like this is magical,” he recalls.

He attributes this shift to familiarity and mindset evolution. “Evolving your mindset and learning new skills has always been part of the developer journey… those that didn’t want to jump on the next level of technology… often fell behind in the software industry,” he observes.

To integrate AI into teams, Dohmke advocates for cultural shifts. “If you’re running a software engineering team or building a startup… be very intentional about AI usage… make it part of your company culture,” he advises.

This includes incorporating AI into workflows, such as using AI tools before search engines or integrating AI into interview processes to assess candidates’ proficiency.

He also emphasizes early education: “Bring it into the schools and universities… kids are much more open to these technologies.” His anecdote about his children using Adobe Firefly to create puppy images illustrates how naturally younger generations adopt AI.

Future-Proofing a Career in Tech

For those entering the tech industry, Dohmke offers clear advice on staying relevant in an AI-driven future.

“Everybody should learn coding… programming languages, boolean logic, the binary system… if then,” he asserts, recommending starting with platforms like Scratch or code.org and using AI tools like GitHub Copilot or Claude to explore programming.

He encourages a curious, question-driven approach, likening it to a child’s relentless inquiries. “Encouraging yourself by learning with AI about AI about coding… you want to learn about systems thinking, systems architecture,” he advises.

Continuous skill development is key, as “the best artists, the best woodworkers… rehearse a lot, they practice a lot, they evolve in their skill sets to become the best of their craft.”

Landing Your First Software Engineering Job

Breaking into tech as a new graduate or bootcamp alum is challenging, but Dohmke sees open-source contributions as a powerful differentiator.

“Your GitHub profile… the number of open source projects you have contributed to or created yourself are equally important, if not more important than your LinkedIn profile,” he states.

A robust GitHub contribution graph showcases not just technical skills but also collaboration and cultural fit. “Learning how open source operates in a specific project… and then being able to contribute back and get the pull request merged… is a good sign that you also understand how to collaborate in a commercial environment,” he explains.

Dohmke advises aspiring developers to study open-source projects, understand their culture, and contribute meaningfully.

“It’s easy to create a pull request without doing all that and then getting shut down… you should go and learn about the project first and how we do test cases here,” he notes.

A strong GitHub profile with merged pull requests in projects like Ruby on Rails or React signals to employers that a candidate can thrive in team settings, whether in open-source or commercial environments.

Thomas Dohmke’s vision for the future of programming is one of collaboration, creativity, and continuous learning. AI is not a threat to developers but a tool to amplify their capabilities, enabling faster prototyping and overlapping roles while preserving the need for specialized skills.

By embracing AI, contributing to open source, and committing to lifelong learning, developers can not only secure their place in the industry but also shape its future. As Dohmke puts it, “Evolving your mindset and learning new skills has always been part of the developer journey.”

‍