Google chief scientist

Jeff Dean’s statement wasn’t made lightly. As the Chief Scientist at Google and co-founder of Google Brain (now part of DeepMind), Dean has been at the forefront of AI innovation for years. His prediction that AI could perform at the level of a junior developer within a year was delivered at Sequoia Capital’s AI Ascent event—an influential gathering of top minds in artificial intelligence and venture capital.

Dean acknowledged that generative AI tools are progressing quickly, and developers are already using them for a wide range of tasks. However, he also clarified that writing code in an integrated development environment (IDE) is just one small part of what junior software engineers do.

“This imaginary virtual engineer likely requires a better understanding of many more things than simply writing code in an IDE,” Dean said.

He explained that a full-fledged junior engineer must know how to write and run tests, debug performance issues, understand software architecture, and collaborate effectively with a team. These are areas where current AI tools still fall short—but maybe not for long.

According to Dean, in about a year, AI’s skill level will be as good as an entry-level coder
Dean’s comments come as generative AI tools become increasingly common among developers
He also noted that writing code in an IDE is only one part of what junior engineers do
AI might be getting close to doing the job of a junior software engineer, according to Jeff Dean, Google’s chief scientist. During Sequoia Capital’s AI Ascent event, Dean said it may only be “about a year-ish” before artificial intelligence reaches the skill level of an entry-level coder. While that may sound futuristic, it’s a view that aligns with what many tech leaders have been saying as AI continues to improve rapidly, especially in tasks like coding. At a time when jobs in the field of tech are already scarce and competition is fierce, especially for fresh graduates, this could mean even more pressure on entry-level engineers.
Dean’s comments come as generative AI tools like ChatGPT, GitHub Copilot, and Google’s own Gemini become increasingly common among developers. These tools are already being used to automate repetitive coding tasks, offer real-time suggestions, and even generate full blocks of code. But Dean also noted that writing code in an IDE (integrated development environment) is only one part of what junior engineers do.

“This hypothetical virtual engineer probably needs a better sense of many more things than just writing code in an IDE,” Dean said in the Business Insider report. “It needs to know how to run tests, debug performance issues, and all those kinds of things.” When asked how AI might eventually learn these broader engineering skills, Dean compared it to how real people gain experience — by learning tools, studying documentation, and learning from more experienced colleagues. “We know how human engineers do those things,” he explained. “They learn how to use various tools that we have, and can make use of them to accomplish that. And they get that wisdom from more experienced engineers, typically, or reading lots of documentation.”

Dean believes AI could do something similar — trying out solutions in virtual environments, learning from documentation, and improving over time. “I feel like a junior virtual engineer is going to be pretty good at reading documentation and sort of trying things out in virtual environments,” he said. “That seems like a way to get better and better at some of these things.”

While he didn’t say how far this could eventually go, Dean thinks it’s going to make a meaningful difference. “I don’t know how far it will take us, but it seems like it’ll take us pretty far,” he said.

Google has not responded to Business Insider’s request for comment at the time of publication.