AI’s so-called second moment isn’t about making tasks faster - it’s about replacing the worker entirely. Nathaniel Whittemore on The AI Daily Brief marks Q1 2026 as the inflection point where autonomous agents began executing complete workflows, not just assisting with them. The result is a “SAS apocalypse” for traditional enterprise software, where valuation is decimated by the obsolescence of the per-seat revenue model. Companies like Pulsia now generate $6 million in annualized revenue with a single founder and zero employees.
“Software is being eaten from within. If an agent can write its own code and generate its own apps, the traditional software-as-a-service seat model is obsolete.”
- Nathaniel Whittemore, The AI Daily Brief
The collapse of technical barriers is immediate and personal. Jesse Genet, a former venture-backed founder, went from never opening a terminal to building her own software agents in six months. She credits the threshold where natural language became a viable programming interface, a shift she calls a personal Cambrian explosion. Her story illustrates a broader trend: the motivation to automate is now the only real bottleneck, not technical skill.
This agentic shift is structurally rewriting company design. Keith Rabois of Khosla Ventures argues the traditional product manager role is now incoherent, as AI capabilities shift every quarter. The human role narrows to pure business acumen - deciding what to build and why - while AI agents act as the execution team. In top organizations, Rabois notes the Chief Marketing Officer is often the single largest consumer of AI tokens, bypassing entire layers of staff to produce work directly.
The underlying architecture is converging into a commodity. As noted on The AI Daily Brief, whether it’s Linear building coding agents or Notion building work agents, the core design is a universal looping harness where the model calls tools until a goal is met. The competitive moat shifts from architecture to proprietary data and distribution. Performance data supports the harness: Blitzy’s agent scored 66.5% on a professional coding benchmark, beating raw models like GPT-5.4 by leveraging deep contextual knowledge a single-pass model would miss.
The displacement is not a future risk - it’s a present reality. Block recently cut 40% of staff, a move read as a portent for aggressive AI-first recalibration. As Whittemore notes, investors are no longer worried AI won’t work; they’re terrified it works too well. The question is no longer if agents will replace roles, but how many, and how fast the organizational chart rebuilds itself around them.
