Anthropic engineer declares the coding interview dead
Anthropic engineering manager Felix Rieseberg says he has stopped using traditional coding interviews for his team, arguing that AI tools such as Claude can now outperform candidates on small coding challenges. He believes those tests have become a poor way to identify the builders who will make the greatest contribution.
The article describes a wider debate over how to assess candidates as coding agents improve. Google engineer Jaana Dogan says she would be interested in seeing how applicants use coding agents, but a ZipRecruiter survey of 1,500 employers found that 62.5 per cent of those in the Information sector still require formal skills tests at the initial application stage. Labour economist Nicole Bachaud suggests employers may put more weight on human skills: 61.5 per cent of Information employers say critical thinking is lacking among candidates.
In short
- Anthropic’s Felix Rieseberg has dropped coding challenges for his team.
- Google’s Jaana Dogan sees value in assessing how candidates use coding agents.
- 62.5 per cent of surveyed Information employers still require formal skills tests.
Steelman: both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Traditional coding interviews may now measure how well candidates perform a narrow exercise that AI tools can complete, rather than their ability to build useful systems at work. Assessing how applicants use coding agents, reason through problems and exercise judgement could better reveal the skills teams need as software development changes.
The case against
A timed skills test can provide a consistent, comparable way to check whether applicants have core technical abilities, especially when hiring teams face many candidates. Since employers also report gaps in critical thinking, abandoning tests before reliable alternatives are established could make it harder to distinguish sound engineering judgement from effective use of AI tools.