Chapter 2 - A Brief History of Talking to Machines
Loop Engineering by Fabio Correa, the Applied AI foundation volume on working with AI through framing, feedback, verification, memory, and human judgment.
Last updated 2026-10-25. The original publications Chapter 2 condenses, plus the two sources the chapter itself cites: Willison on the shift from prompting to loop design, and Sweller's cognitive-load work, which supplies the per-era load framework.
The question this chapter takes up
What did each era of talking to machines make possible, and what did each leave unsolved?
Put it to work
Choose one interface shift from this chapter and write what moved from the human to the machine: translation work, active memory pressure, feedback latency and clarity, error recovery complexity, abstraction burden, composition burden, or verification burden.
Key sources
Chapter 2's own citations are Willison on the shift from prompting to loop design [D11] and Sweller's cognitive-load work [D17], which supplies the per-era load framework. The full source list lives in the print book's Appendix D.
- D11 - Willison, S. (2025). Designing agentic loops. simonwillison.net/2025/Sep/30/designing-agentic-loops - the shift from prompting to loop design that frames the chapter's final era.
- D17 - Sweller, J. (1988). Cognitive load during problem solving. doi.org/10.1207/s15516709cog1202_4 - the load framework used to score each interface era. The chapter is explicit that no primary source quantifies per-era allocation at this granularity; the trajectory follows the theory and the per-era discussion.
Primary-source links
- Correa, "Dialog Engineering: AI as Your Research Assistant" (October 2024) - the public artifact this chapter's dialog-era treatment draws on directly.