Silicon Valley is undergoing a fundamental transformation as AI coding agents automate technical execution, forcing software engineers and founders to prioritize “agency”—the ability to direct and oversee autonomous systems—as the ultimate competitive advantage. As tools like Claude Code and Codex handle the “how” of production, the human role is pivoting exclusively toward the “what” and “why,” signaling a shift that experts believe will soon disrupt finance, law, and the creative sectors.
From Coding to Orchestration: The New Developer Workflow
The traditional image of the lone developer writing lines of code for hours is fading. Industry leaders, including Notion co-founder and COO Akshay Kothari, suggest that modern AI agents may already possess more technical capability than seasoned professionals. In this new landscape, human “taste” and “agency” remain the final frontiers of value. While some find the transition unsettling, others are leaning into the role of the “Super IC” (Individual Contributor).
Simon Last, another co-founder at Notion, exemplifies this evolution. After two decades of manual coding, Last has transitioned entirely to managing up to four AI agents simultaneously. His workflow is defined by “token anxiety”—the urge to keep agents working in the background 24/7—tempered only by the “context overload” that occurs when a human brain attempts to supervise too many autonomous streams at once. For Last, the ability to harness these agents is now the world’s most indispensable skill, one that rewards curiosity and open-mindedness over rote technical proficiency.
The Death of Raw Execution in Tech Hiring
The shift is fundamentally altering how venture-backed startups build their teams. Traditional “raw execution” is no longer the primary metric for success. Instead, companies are hunting for individuals who can build the “machine” that allows agents to handle end-to-end feature development. Yoni Rechtman, a partner at Slow Ventures, highlights this trend through a job description for Phoebe, an AI healthcare startup, which explicitly states it is not looking for manual execution but for people who can automate their own roles from day one.
The “No Slop” Rule and Human Accountability
Despite the autonomy of these tools, the burden of responsibility remains firmly with the human operator. Jennifer Li, a general partner at Andreessen Horowitz, notes that ignoring AI tools is now a “red flag” for investors, yet she warns against “slop”—low-quality or incorrect code generated by AI and left unchecked. Under the “no slop rule,” the person who submits the work is held strictly accountable for its integrity, regardless of whether a human or an agent wrote the initial draft.
Beyond the Buzzword: The Cultural Divide of “Agentic”
While “high agency” has become the defining trait of the AI era, the term itself is facing a cultural backlash. Critics argue the label carries a “cringe” factor, implying a binary worldview where the population is divided into “main characters” (the agentic) and “NPCs” (those who follow scripts). However, the economic reality remains: as AI continues to commoditize technical skills, the premium on those who can proactively navigate and bend these systems to their will continues to skyrocket.
As Notion’s Kothari observes, the company isn’t necessarily downsizing, but it is hiring differently. The market now values a handful of high-agency “orchestrators” over thousands of traditional engineers. This “new way of working” is currently a Silicon Valley phenomenon, but as AI agents become more sophisticated, the question “Am I agentic enough?” will likely become a central theme for the global workforce.
