Anthropic Engineering Manager Advises Newcomers: Spend Less Time Grinding Coding Problems, Use AI More to Bring Ideas to Life

动察 Beating AI News Flash: Anthropic Engineering Manager Felix Rieseberg announced that this month he has canceled traditional coding interviews in his team's hiring. He previously participated in the development of Claude.ai, Cowork, and Claude Code Desktop.

In the past, interviewers would ask candidates to write code on the spot, such as completing a React programming problem, to assess coding ability. But now Claude can already do better, and he found that the best-performing engineers on his team are not necessarily good at this kind of exam.

He also found that some engineers who have just entered the industry can already produce better results than senior employees with the help of AI. In the past, when newcomers had ideas, they often had to first write a proposal, persuade their managers, and then wait for design and development resources. Now they can directly use AI to build product prototypes and let actual results speak for themselves.

There is an example inside Anthropic. Company employees have long been accustomed to using Claude Code in the terminal, but two recently hired engineers, Anthony Morris and Alice Zhao, took the initiative to build a more convenient graphical interface for it. The two completed the prototype in just one weekend and launched the first public version about eight days later.

Rieseberg therefore advises young engineers not to focus only on learning programming frameworks and grinding through interview questions. What is more worth spending time on is understanding the company's business, identifying real problems, and then using AI to build solutions. He believes that AI is narrowing the technical gap between newcomers and veteran employees, and that experience and seniority are no longer as advantageous as they used to be.

However, after canceling traditional coding interviews, he himself has not yet found a better assessment method. At present, he mainly looks at candidates' resumes and portfolios, with particular attention to small projects that can solve real problems and demonstrate product judgment.