Building OpenCode with Dax Raad

Gergely Orosz, host of The Pragmatic Engineer podcast, interviews Dax Raad, co-founder of OpenCode, about why faster code generation has not removed product judgment as the bottleneck. Raad describes an open-source, provider-neutral harness strategy while warning that agents can make weak features and architectural hacks cheaper to ship; OpenCode's growth, revenue, and speed observations are company-reported rather than independently measured ([00:03:33]-[00:06:27], [00:17:46]-[00:20:38], [00:49:05]-[00:54:35]).

Key Points Covered

• Product judgment remains scarce: Shipping every user request or competitor feature creates an incoherent product and an expanding support burden, even when implementation is cheap ([00:03:33]-[00:06:27]).
• Neutrality is a competitive strategy: OpenCode claimed the open-source harness position so rival model providers would have reasons to support it; continued neutrality still depends on provider competition ([00:17:46]-[00:19:51], [00:25:16]-[00:27:12]).
• Initial UX came before harness superiority: The team invested in terminal rendering and a fast first prompt, while Raad says the early harness itself was only good enough ([00:28:15]-[00:30:21]).
• Provider access is not permanent: Anthropic's subscription restriction showed how vendor incentives can change; OpenCode responded by adding official support from competitors ([00:20:38]-[00:27:12]).
• Agents hide the felt cost of hacks: Easy implementation mutes the discomfort that once encouraged an engineer to reconsider a poor design, while leaving future maintenance risk intact ([00:49:05]-[00:51:52]).
• Cheap generation can fund cleanup too: Raad advocates migrating improved patterns across old code to keep the repository workable, while acknowledging that cleanup has no immediate revenue line ([00:52:46]-[00:54:35]).
• Builders receive feedback directly: OpenCode moves work to a demonstrable foundation, then has the implementing engineer process GitHub and social feedback without a summarizing intermediary ([01:02:39]-[01:04:47]).
• Guardrails are familiar engineering work: Tests, conventions, and stronger domain structure help fallible agents make safer changes; the episode does not measure their comparative defect impact ([01:12:40]-[01:15:27]).
• Speed and business claims need qualification: Raad says aggressive AI use did not obviously make OpenCode faster, and the episode's user, run-rate, and margin figures are self-reported snapshots rather than durable outcomes ([00:19:51]-[00:20:38], [00:33:26]-[00:38:55], [00:54:35]).

Full video: https://www.youtube.com/watch?v=1VqKUrxR2C8(opens in a new tab)