cd /news/artificial-intelligence/the-traits-that-make-engineers-stand… · home topics artificial-intelligence article
[ARTICLE · art-92918] src=businessinsider.com ↗ pub= topic=artificial-intelligence verified=true sentiment=· neutral

The traits that make engineers stand out, according to Coinbase's CTO

Coinbase CTO Rob Witoff told Business Insider that human taste and judgment are the traits that make engineering candidates stand out as AI takes over routine tasks, with engineers expected to tackle 'much harder strategy decisions.' Coinbase has rebuilt its engineering interview loop to test how candidates direct AI, evaluate its output, and apply judgment, following the company's May layoffs of 14% of its workforce (about 700 employees) as CEO Brian Armstrong said AI was helping engineers 'ship in days what used to take a team weeks.'

read3 min views1 publishedAug 12, 2026
The traits that make engineers stand out, according to Coinbase's CTO
Image: Businessinsider (auto-discovered)

AI can do a lot, but it can't replace human taste and judgment.

Coinbase CTO Rob Witoff told Business Insider that those are the traits that make engineering candidates stand out nowadays. As AI takes over routine tasks, Witoff said engineers will increasingly be expected to tackle "much harder strategy decisions."

That means the company wants developers who aren't "working in a single lane," he said.

"We want people that can step back and really understand why they're doing something," Witoff said.

What it means to have 'taste and judgment' #

In a July blog post on the topic, Coinbase describes taste and judgement as "Knowing what's actually worth building, telling a genuinely good solution from one that only looks right, and sensing when to override the model."

Witoff said strong taste doesn't necessarily come from years of industry experience. For many candidates, it can come from using the company's products or participating in the crypto market.

"You can develop that judgment a lot of ways, but having that as a capability inside is what's really important," Witoff said.

Rather than grading how quickly an engineer can write raw syntax or complete algorithmic puzzles, Witoff said that judgement comes into play when workers have to determine which trade-offs to make and when to reject what an AI model produces.

The company has rebuilt its engineering interview loop to keep pace with the change by testing how candidates direct AI, evaluate its output, and apply judgment where models fall short.

What matters less, Witoff said, is expertise with any single AI tool. With AI models evolving so rapidly, he said he avoids over-indexing on a single tool and instead prefers software engineers who are comfortable using a range of models.

The CTO's comments come after Coinbase laid off 14% of its workforce in May, or about 700 employees, saying AI is reshaping how the company operates.

At the time of the job cuts, CEO Brian Armstrong said the company would have "no pure managers" as it pivots to tiny teams, some of which will operate with engineers, designers, and product managers in a single role.

Armstrong also said in the layoff memo that AI was helping engineers "ship in days what used to take a team weeks."

According to Coinbase's most recent regulatory filing disclosing headcount, the company employed 4,951 employees as of December 2025, before the layoffs.

Why judgement and taste are becoming more important

Witoff isn't alone in arguing that judgment and taste are becoming the most valuable skills in technical hiring.** **Tech leaders, including Y Combinator cofounder Paul Graham and OpenAI President Greg Brockman, have said strong judgment is increasingly essential.

Witoff said engineers make up roughly half of the company's workforce, with more senior and principal engineers on staff than ever before. The company said it doesn't disclose specific employment numbers.

Coinbase is increasingly bringing those engineers into higher-level conversations, Witoff said, and AI tools are enabling many of them to solve problems in days that once took months or even quarters.

Now that one strong engineer can now do the work of a full team, and everyone has a fleet of agents working for them, engineers need to know how to direct teams and collaborate with agents and people across disciplines, Witoff said.

Witoff said humans still have an edge over AI in understanding other humans, making it increasingly important to focus on customer needs, which products to build, and when to build them. Today, even the strongest teams may spend just 20% to 30% of their time on those strategic questions, he said. Over time, he expects that share to climb to 95% as AI handles more repetitive work.

── more in #artificial-intelligence 4 stories · sorted by recency
── more on @coinbase 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/the-traits-that-make…] indexed:0 read:3min 2026-08-12 ·