Anthropic engineering manager Felix Rieseberg says his team has stopped using traditional live coding interviews, arguing that AI tools have weakened their ability to predict on-the-job performance. In a post on X and a blog post, Rieseberg said the company’s Claude model now outperforms candidates on many whiteboard-style tasks, prompting him to remove coding interviews from hiring loops.
“The coding interview is dead, at least for my team at Anthropic,” Rieseberg wrote in a post on X on October 27, 2026. In his blog, he added that there is “no point” in assigning candidates contrived coding challenges or quizzing them on low-level languages such as assembly, because, he wrote, “Claude is now better at both.”
Rieseberg described coding challenges as once a “weak but useful” signal, but said that the engineers who make the biggest impact on his teams are not always the ones who excel under traditional interview conditions. “I’m seeing engineers who are comparatively new fly past people who are probably better coders and have been around much longer,” he wrote. “Hunger beats tenure.”
A manager at Anthropic says coding interviews are obsolete and advises junior developers
Keisha Lance Bottoms confident despite being heavily outspent in close Ga. governor race
Rieseberg’s move comes as other large tech companies adapt hiring and evaluation to a world where AI can write and debug substantial portions of code. Business Insider’s Hugh Langley reported last week that Google is expanding pilot interviews that let candidates use its Gemini model to demonstrate AI fluency and prompting skills. According to Business Insider, Amazon has told engineers to emphasize customer and business outcomes rather than raw AI usage, calling the moment the rise of the “Renaissance Developer,” in which human judgment and machine intelligence are meant to work together.
How junior engineers can stand out
Rieseberg used his blog to offer practical advice for early-career engineers and others without formal authority inside organizations. Rather than focusing on clearing routine backlog tickets, he recommended identifying the concrete business problems the company is trying to solve and proposing bold, narrowly scoped solutions that materially advance those goals.
He urged newcomers to build a working prototype quickly — ideally over a weekend — and to ensure that a project communicates one compelling idea rather than burying several promising concepts in an over-ambitious product. He also encouraged junior engineers to use their outsider perspective to spot broken user experiences and to question assumptions that more senior staff may take for granted.
Rieseberg pointed to a recent example inside Anthropic: two relatively new engineers who built a graphical interface for Claude Code over a weekend, with a first public version shipping roughly eight days later. He said the episode illustrated how rapid, focused prototyping can surface practical value faster than conventional measures of seniority.
At the same time, Rieseberg acknowledged that companies still lack a reliable way to assess candidates who excel at building with AI tooling. For now, he said he is examining resumes and portfolios for small prototypes that reveal what he called “taste, empathy, and principled thinking.” He framed that shortlist of qualities as the attributes he and other AI leaders are increasingly valuing when evaluating candidates in an era of powerful code-writing models.
Rieseberg’s changes reflect an ongoing industry debate about how to evaluate engineering talent as AI tools shift what it means to write and ship software. He emphasized that his team has removed the coding interview for now, but also admitted the broader question of fair, reliable assessment in an AI-augmented workforce remains unresolved.
