For decades, digital computation was structured around the bit—the smallest unit of data. But with the rise of artificial intelligence, a new paradigm is emerging: one in which the model becomes the core computational engine. This shift, as I propose in The New Bit, repositions models—particularly foundation models such as LLMs—as active participants in reasoning, decision-making, and autonomous action.
This white paper explores the rise of AI agents as the embodiment of this transformation. These agents represent a leap beyond traditional bots, crawlers, or RPA. They reason, plan, and adapt. They leverage large models, orchestrate tools, and pursue objectives across time. In doing so, they signal not just a technological evolution, but a deeper conceptual shift in how we think about computation, agency, and collaboration between humans and machines. Unlike traditional bots, crawlers, or RPA, agentic systems reason, plan, and adapt. They leverage large models, orchestrate tools, and pursue objectives across time. In doing so, they signal not just a technological evolution, but a deeper conceptual shift in how we think about computation, agency, and collaboration between humans and machines.