With AI writing the code, entrepreneurial thinking is now what counts most for newcomers at Anthropic
Felix Rieseberg, an engineering manager at Anthropic, has stopped using traditional coding interviews. He says these tests no longer make sense because AI can write code, and ambitious young developers can use AI to get ahead of more experienced coworkers.
Rieseberg has led the development of Claude.ai and Claude Code Desktop at Anthropic. He says there is no longer a good reason to ask job candidates to complete small programming tasks. According to him, Claude can now do these tasks better than the candidates can. He believes giving someone a small React programming challenge today makes about as much sense as testing their knowledge of Assembly, an older programming language.
On his teams, high test scores no longer predict who will become a great developer. Instead, he sees engineers with less experience passing more experienced coworkers. "Hunger beats tenure," he says. Rieseberg's advice is aimed at "junior" developers, but he uses this word in a different way. He does not mean only people who are new to programming. He means any engineer who lacks status, power, or connections inside their company.
Why newcomers spot the problems veterans stopped seeing
Rieseberg says newcomers have an advantage because they notice problems that longtime employees accepted as normal years ago. He points to organizational researcher James March, who wrote in 1991 that organizations need a steady flow of newcomers who do not yet know how things are usually done. Stripe founder Patrick Collison has also directly encouraged new employees to point out anything that seems strange. Experienced employees, he noted, stopped noticing these everyday problems long ago.
Rieseberg gives an example from his time at Notion. He was asked to design a system that would let the product work offline, even though people inside the company believed this was impossible. They thought the data model was too complicated and that CEO Ivan Zhao would never accept the only solution that could work. The technical problem was real, but the belief that the founder would reject the solution was not based on fact. Nobody had asked him because everyone treated this belief as something they already knew to be true.
Cheap execution weakens the old hierarchy
Rieseberg says the main problem for newcomers is that people inside the company do not listen to them. In the past, turning an idea into a working product took a lot of time and money. Companies were therefore organized so that a small group of decision-makers could approve or reject ideas. Employees had to climb a strict hierarchy before managers trusted them to build something. Rieseberg writes that AI changes this situation because building things is no longer as costly.
His first piece of advice is based on a statement Patrick McKenzie made in 2011: "engineers are hired to create business value, not to program things." Developers should understand what their company is really trying to achieve and focus on work that helps the business make money. A good place to start is to learn what their manager's manager wants to accomplish.
Next, Rieseberg suggests choosing a bold idea that could make a big difference to those goals and building a working example over a weekend. This lets developers skip the usual steps of writing product plans, getting approvals, and persuading designers or product managers. The important thing is to focus on one strong idea instead of mixing three good ideas with 20 bad ones. Trying to build a prototype that includes a new programming language, a backend system, and mobile apps all at once will lead to failure. The prototype should explain the idea so clearly that company leaders can quickly see its value and recognize the right person to develop it further.
Newcomers have the edge on empathy
Rieseberg describes two types of engineers. Newcomers who build products for users often have a natural ability to understand people outside their company. At companies such as Slack and Notion, some employees who had been there for a long time struggled to see things from the point of view of an ordinary outside user. By contrast, engineers who build infrastructure or internal tools can benefit from not knowing why other people believe something cannot be done.
He supports this point with an example from Anthropic. Two engineers who were relatively new to the company at the time, Anthony Morris and Alice Zhao, built the graphical user interface for Claude Code over a weekend. The first public version was released about eight days later.
Rieseberg now looks for small working prototypes in job applications. These projects should show good judgment, an understanding of users, and clear principles behind the decisions made. He says traditional coding interviews are not coming back.
On one hand, Rieseberg's advice gives newcomers more power. AI can help them get past people who control access to opportunities but have stopped noticing problems inside their own companies. On the other hand, this approach changes what companies expect from people applying for entry-level jobs. Instead of mainly looking for careful work and technical skill, employers may place more value on business thinking, understanding products, and good judgment about what to build. Most universities do little to teach these skills, and companies have traditionally expected them from experienced product managers or founders.
The engineers Rieseberg praises also spend their weekends building prototypes without being asked. "Hunger beats tenure" sounds inspiring, but it assumes that every new developer can bring the same strong drive to work and put in extra effort outside normal working hours.