AI can do a lot of work, but it still can't replace human taste and critical judgment, which are the qualities that make engineering candidates exceptional today, says Coinbase CTO Rob Wittoff.
AI can do a lot of work, but it still can't replace human taste and critical judgment, which are the qualities that make engineering candidates exceptional today, says Coinbase CTO Rob Wittoff.
As AI takes over routine tasks, developers are increasingly expected to solve much more complex strategic problems. Companies are no longer looking for specialists who work within the narrow framework of one specialization.
"We want to see people who are able to take a step back and really understand why they are doing something," Wittoff said in an interview with Business Insider.
A Coinbase blog post explains the concept of “taste and judgment” as: “The ability to understand what is truly worth building, the ability to distinguish a truly good solution from one that only seems right, and the sense of when to reject an AI model’s response.”
According to Wittoff, sophisticated taste doesn’t necessarily come from decades of industry experience. For many candidates, it’s formed through active use of the company’s products or participation in the crypto market.
Instead of testing how quickly they can write code or solve algorithmic problems, Coinbase is assessing how engineers find compromises and when to reject an AI-generated result. The company has completely redesigned its interview process to test how candidates navigate AI and apply their own dissenting opinions where the model makes mistakes.
At the same time, deep expertise in any one AI tool becomes less important. Because models are evolving so quickly, Wittoff advises not to get tied to a specific technology, but to be flexible and confident in using a wide range of tools.
The view that taste and judgment are becoming the most valuable skills in hiring is shared by other industry leaders, including Y Combinator co-founder Paul Graham and OpenAI president Greg Brockman.
At Coinbase, engineers make up about half of the workforce, and the company now has more Senior and Principal developers than ever before. Thanks to AI tools, they can complete tasks in days that used to take months or quarters.
Since one strong engineer today can do the work of an entire team, and each employee has a fleet of AI agents at their disposal, coordination and interdisciplinary collaboration skills become key.
Humans still have an undeniable advantage over AI in understanding other people, so a focus on customer needs becomes crucial.
Today, even the strongest teams spend only 20–30% of their time on strategic issues (what and when to build).
In the future, this share is expected to increase to 95%, as artificial intelligence will completely take over the entire routine.
