Clemens Rawert connects AI capability to the slower work of achieving economic productivity. Historical examples from steam engines, electrification and computing show why investment, complementary tools, new skills and organizational redesign can matter as much as an invention itself.
Kitze traces the shift from autocomplete to coding agentsAn AI coding agent is a tool-using AI system that can inspect, modify, and validate software within a repository. and software factoriesAn autonomous software factory is a repeatable development system that coordinates agents, tools, checks and human review to turn specifications into verified software changes.. His practical emphasis is on isolated task environments, repeatable pull-request checks, portable tools and skills, and human review rather than layers of AI-generated work passed between colleagues.
Lélio Renard Lavaud describes a shared agent harnessAn AI agent harness is the software framework that packages a model with tools, instructions, context management, execution controls, and user interaction. for developers and business users, with connectors, asynchronous tasks and sandboxesAn AI agent sandbox is an isolated execution environment that limits which files, processes, networks, credentials, and external systems an agent can access.. His examples illustrate incident investigation and argue for dynamic permissionsAn agent permission boundary limits the information, tools and actions an AI agent can use during a task. and runtime controls that resist prompt-injection-driven data exfiltration while preserving useful workflows.
Lélio Renard Lavaud connects enterprise deployment to data, compute and operational sovereignty, alongside specialized speech and document models. These are a speaker's plans and demonstrations, not independent proof that every proposed capability is generally available.
Watch on YouTube
