The Stacked Podcast discusses social reports of AI coding tools deleting files and production data, explicitly acknowledging that the underlying incidents cannot be verified from the reports alone. The hosts compare cost, design quality and oversight needs before considering AI-assisted model improvement and public participation in AI wealth.
The conversation explores control over powerful models, cross-model adversarial review and the economic value of agent platforms. Later sections examine routing services and video-generation interfaces, arguing that usability and choosing the right model can matter alongside raw model capability. Far-reaching forecasts remain speculative rather than guaranteed outcomes.
Watch on YouTube
