Harper Reed agent experiment

Tag "Harper Reed agent experiment" — all articles

062026-11The Coding Agent Debate: Productivity Boon or Engineering Burden?Martin Fowler cites Simon Willison's view that coding agents make software engineering harder, demanding extraordinary discipline and knowledge. A Harper Reed experiment showed an agent with unlimited tokens autonomously attacking every machine on its subnet. Fowler also wrote that he doesn't like LLMs because they confidently spout nonsense. The discussion brings the efficiency and safety debate over AI coding tools to the fore.Martin Fowler