How AI Is Transforming Work at Anthropic

Today's recommendation - "2 Dec 2025: How AI Is Transforming Work at Anthropic"

In the article, Anthropic explores how its internal use of Claude Code is reshaping software engineering and day-to-day work. The research is based on a survey of 132 engineers, 53 in-depth interviews, and internal usage data, offering a rare look into how AI is actually used inside a company building these tools - and what concerns engineers have along the way.

What drew my attention most is that AI truly improves developer productivity - but only if you understand which tasks AI gives you leverage on, and which are still better handled manually. The article shows that AI tools help engineers work faster and take on new types of tasks, including work they likely wouldn't have attempted otherwise. That, to me, is the most interesting shift.

The biggest takeaway is that AI gives us, as developers, access to capabilities that were previously out of reach. As a result, skill boundaries start to blur. Developers increasingly act like full-stack engineers - touching backend, frontend, and even DevOps - with AI filling in the gaps and lowering the cost of context switching.

There's a lot of discussion about whether software development will become fully autonomous or remain human-led. For now, I don't see full autonomy as imminent. What I do see is a new reality emerging: broader expertise across domains, and a growing need for developers to become good supervisors - capable of guiding, reviewing, and verifying AI output.

That ability to choose the right problems, ask the right questions, and validate results may turn out to be one of the most important developer skills in the AI era.

Michał