David Ondrej describes his workflow for coding with AI agentsAn AI coding agent is a tool-using AI system that can inspect, modify, and validate software within a repository., from reproducing bugsDebugging is the process of finding why software behaved incorrectly and checking that a correction addresses the cause. and separating tasks into individual sessions to reviewing larger changesCode review examines proposed source changes to assess their correctness, clarity and suitability before or around integration. with a second model. His preferences about model choice and testing reflect his own experience and may change over time.
Ondrej distinguishes predictable parallel work from exploratory tasks, and compares local agents with isolated cloud environmentsAn AI agent sandbox is an isolated execution environment that limits which files, processes, networks, credentials, and external systems an agent can access.. He describes the automation loops he uses to triage feedback, investigate reports and involve a person in product decisionsHuman-in-the-loop AI keeps a person involved in reviewing, correcting, approving, or guiding an AI system's work..
The presenter also discusses local inference for sensitive work and the costs and trade-offs of hardware. He develops Cloud Room, a product shown in the video; its promotional invitations are omitted from this account, and the safety and productivity claims are his own.
Watch on YouTube
