AWS principal engineer says the best engineers are ‘paranoid and lazy’ when it comes to AI

• David Yanacek, a senior principal engineer at AWS, said the best engineers are "paranoid and lazy."
• He said paranoia keeps him alert about potential safety risks.
• Laziness motivates him to find ways to prevent those risks and make things work, he said.

Many think that judgment and taste are the traits that set apart the strongest engineers in the AI era.

David Yanacek, a senior principal engineer at AWS, thinks the ideal developer has some additional attributes.

He told Business Insider that the best engineers nowadays are also "paranoid and lazy." That, he said, makes for a good combination when working with agents.

"Paranoid means I'm very sensitive to what could possibly happen if I gave an agent permission to do a thing without supervision," said Yanacek, who works on agentic AI and safety.

Laziness, he said, means he doesn't want to perform repetitive tasks manually.

The new workflow

While AI has taken over most coding tasks, Yanacek said that wasn't the main bottleneck. In addition to writing the actual code, engineers used to spend days waiting for manual tests, code approvals, and design tweaks. Those processes required constant waiting for teammates' reviews.

Now, human intervention is reserved for design discussions and complex reviews, while agents handle much of the other work in a loop.

These loops involve generating code, running tests, reworking problems, and preparing updates that then get merged and deployed. While one loop can run for hours, Yanacek can shift to other projects or add additional features to a product in the meantime.

Setting the right guardrails

For agentic workflows to run efficiently and save time, agents need to operate autonomously, Yanacek said. So he spends much of his time deciding which tools agents can use without supervision, which actions need approval, and where those limits should be enforced.

To make that work, Yanacek makes sure his "auto-trust" file is sound. That's the set of rules that specifies which tools and commands an agent can use independently, and which ones require a human sign-off.

"I always think about what is in the auto-trust file, because that's how you get productivity: by having the agent be able to do things without you having to micromanage every single command that it runs," Yanacek said.

His goal is to have a fleet of agents making the right calls without constant review and approval. That requires clear policies for what an agent should and shouldn't do.

"Most of what I think about is AI safety," Yanacek said. "And because I'm doing that, I sleep pretty well."

Engineers need to have a layer of paranoia and laziness so they can be alert to what could go wrong when an agent acts without supervision, he said. However, they also need to build systems reliable enough that they don't have to supervise every action.

The end goal, he said, is to do enough work up front to make the agent trustworthy enough to run unsupervised so it can automate the tedious parts.