Mastering Agentic Skills for the AI Era
▼ Summary
– Silicon Valley increasingly values “high-agency” individuals who can think and act independently, a trait emphasized more as AI coding tools advance.
– AI coding agents like Claude Code and Codex are automating much software development work, shifting engineers’ value to deciding what tasks these agents should perform.
– Notion cofounder Simon Last exemplifies this shift, using multiple AI agents to code and viewing the skill to manage them as critically important and untrainable.
– Notion hires for high agency and adapts its hiring strategy, valuing a few highly agentic individuals over thousands of traditional engineers to leverage the “new way of working.”
– While currently concentrated in tech, industry leaders believe this agentic, AI-augmented work model will eventually spread to finance, legal, creative, and other sectors.
The tech industry is undergoing a fundamental transformation, where the most valuable human skill is no longer just writing code but mastering the direction of artificial intelligence. As AI coding agents become capable partners, the role of the software engineer is evolving from a hands-on builder to a strategic orchestrator. This shift places a premium on a specific set of human qualities: initiative, judgment, and the ability to guide intelligent systems toward meaningful outcomes.
Akshay Kothari, a cofounder of the productivity platform Notion, observes that AI agents are rapidly acquiring technical capabilities. He suggests that what will ultimately distinguish humans in this new landscape is their agency, the capacity to think independently, make decisive choices, and drive projects forward without constant supervision. This perspective, while perhaps extreme to some, is gaining serious traction where technology is created. The conversation has moved beyond whether AI will change work to a deeper question: what unique value do humans provide when machines can execute so many tasks?
For developers, the change is already palpable. Tools that automate significant portions of coding are becoming standard, shifting a professional’s core value toward defining problems, setting priorities, and evaluating AI-generated solutions. This transition can be disorienting for those who find deep satisfaction in the craft of programming itself. Yet, for individuals who naturally excel at high-level direction and leverage, it represents a substantial opportunity.
Consider Simon Last, another Notion cofounder and a prolific early adopter. After nearly twenty years of traditional coding, he now primarily works through multiple AI agents, viewing his role as that of a conductor. He delegates tasks, reviews output, and integrates the work, but he rarely writes lines of code manually. Last describes a new form of productivity anxiety, worrying if his digital workforce is idle. He also notes the cognitive strain of managing too many simultaneous agent threads, a challenge he terms “context overload.”
According to Last, the critical skill is knowing how to harness these tools effectively, an aptitude he believes cannot be taught through conventional training. It requires innate curiosity, adaptability, and a willingness to experiment relentlessly with new technologies. The individuals who possess this blend of traits, he argues, are seeing their market value increase exponentially because the scale of what they can accomplish has grown dramatically.
This new paradigm is reshaping hiring. Notion reports that while it uses AI to automate work, the company is not reducing its workforce. Instead, it is intensifying its focus on recruiting people with exceptional agency. The logic is that a small number of highly agentic individuals, who can productively direct teams of AI agents, create more value than a large group of traditional engineers. The company seeks talent that intuitively grasps this “new way of working,” where human oversight and strategic decision-making are the primary contributions.
For professions outside of software engineering, this shift may seem distant. Current data indicates that widespread AI adoption in many job sectors is still in its early stages. However, leaders like Kothari are convinced the transformation will inevitably spread to fields such as finance, law, and creative industries. When it does, the central question for professionals will shift from technical proficiency to a more personal evaluation: do I have the agentic skills to guide intelligent systems and define what needs to be done? The future of work may belong less to those who execute tasks perfectly and more to those who can best determine which tasks are worth doing.
(Source: Wired)
