Anthropic shifts hiring focus from coding tests to entrepreneurial prototypes

The end of the coding test

Felix Rieseberg, an engineering manager at Anthropic who led development for Claude.ai, is moving away from traditional coding interviews. He argues that these tests have become largely pointless because AI can now handle small programming tasks more effectively than human candidates do.

Rieseberg compares giving a candidate a small React challenge to quizzing someone on Assembly language. In his view, these tests no longer accurately predict who will become a top developer on a professional team. Instead, he is observing that ambitious new engineers are often able to overtake more experienced colleagues by leveraging these tools.

As reported by The Decoder, the focus is shifting from the ability to write code to the ability to use code as a means to an end. The goal is to find engineers who understand how to apply technical skills to solve actual problems rather than those who can simply perform manual programming tasks.

Why newcomers have the edge

Rieseberg suggests that new hires have a structural advantage because they are often the only ones who notice problems that long-time employees have accepted as normal. He references organizational researcher James March, who noted that companies rely on a steady stream of 'naive' newcomers to spot these issues.

Long-tenured employees often stop noticing friction points because they have become accustomed to the existing systems. Rieseberg points to his own experience at Notion, where he was asked to design an offline architecture that everyone else considered impossible because they had stopped questioning the status quo.

This perspective is why Rieseberg defines 'junior' developers as anyone without social status or a power network inside a company. These individuals are often the best positioned to flag things that seem strange or inefficient to an outsider.

Bypassing the corporate hierarchy

In the past, execution was expensive, which meant organizations were built with rigid hierarchies to approve every idea. You had to work your way up the ladder before management would trust you to actually build something. AI changes this dynamic because it makes execution significantly cheaper and faster.

Rieseberg advises engineers to focus on creating business value rather than just programming things. He suggests that developers should identify their manager's goals and pick one ambitious idea that would dramatically advance those objectives. By building a working prototype over a weekend, an engineer can skip the usual cycle of product specs and lobbying.

The goal of a prototype is to be so clear that leadership immediately sees both the opportunity and the right person to lead the project. Rieseberg warns that trying to build too much at once, such as a new language and mobile apps simultaneously, will lead to failure. The prototype should simply get the idea across with clarity and taste.

Empathy and product sense

Rieseberg distinguishes between two types of engineers: those working on user-facing products and those working on infrastructure. Newcomers often bring a natural empathy for outside users, whereas long-tenured colleagues may find it harder to see a product from an external perspective.

Conversely, engineers working on internal tools benefit from not knowing why certain things are supposed to be impossible. He highlights two relatively new engineers at Anthropic who built the graphical interface for Claude Code over a single weekend. That version was then shipped to the public only eight days later.

Because of this, Rieseberg now looks for small prototypes in job applications that demonstrate taste and principled thinking. He believes that while AI acts as an equalizer for getting past gatekeepers, it also raises the bar for entry-level roles. The requirement is moving away from diligence and toward a more entrepreneurial mindset.

The shift in requirements

This new approach suggests that the skills most universities teach may not be the ones that matter most in the current AI landscape. Companies have traditionally expected product sense and entrepreneurial thinking from experienced managers or founders, not necessarily from new engineers.

Rieseberg's philosophy assumes a 'hustle' mentality where developers build prototypes without being asked. While this is empowering for those who can do it, it places a high demand on newcomers to be proactive and self-directed. It moves the definition of a 'good' engineer toward someone who can identify a problem and execute a solution independently.

The classic coding interview is not expected to return in this environment. Instead, the most successful newcomers will be those who can navigate the gap between technical possibility and business value. They must be able to prove they can spot a problem and build a functional proof of concept quickly.

What to watch

How will other major tech companies adjust their hiring rubrics as AI tools become standard in the development workflow?

Will the 'hustle' requirement for building weekend prototypes create a barrier for talented developers who lack the time or resources to work outside of their primary jobs?

Try it

Identify a small friction point in a tool you use daily and try to build a functional prototype to solve it over a weekend.