Early adopters are abandoning the 'fire-and-forget' model of autonomous agents running on dedicated hardware. The new paradigm uses tools like Codex as integrated 'operating systems' for work. This approach favors closer, semi-synchronous collaboration across multiple devices over the high-latency, low-control model of full autonomy.
Optimal AI workflow involves humans acting as the "bread" on either side of the AI's work. A human first sets the frame and defines "good," the AI then executes the core task (drafting, coding), and finally, a human judges the output and decides the next steps. This structure ensures quality and strategic direction.
Every initially gave each employee a personal AI agent but found this created a massive maintenance burden and knowledge silos. They shifted to shared agents focused on team functions (e.g., analytics). This centralizes maintenance, improves continuity when employees leave, and scales benefits across the entire team.
Instead of freeing up time, AI agents expand the scope of possible work, creating an endless queue of tasks. The key human skill becomes managing this "infinite backlog" and deciding what agents should do next, rather than executing the work itself. This introduces a novel form of professional overwhelm.
AI models are trained on past human work (code, articles, designs), making those skills cheap and accessible. This abundance creates homogenous, default outputs or "slop." Consequently, the market develops an urgent demand for human experts who can create something novel and differentiated, moving beyond the model's defaults.
While layoffs often cause a stock bump, there's growing evidence that markets are more interested in long-term, AI-driven growth. Atlassian's stock soared after reporting strong sales for its AI-enhanced products, not after its layoffs. This suggests investors see AI's true value in expanding capabilities, not just cutting costs.