# Live from DevDay — the OpenAI Podcast Ep. 7 Source: https://www.youtube.com/watch?v=QIdUllqmuls Recap page: https://rapidrecap.app/video/QIdUllqmuls Generated: 2025-10-06T20:32:17.14+00:00 --- ## Quick Overview The OpenAI Podcast live from DevDay featured interviews with developers from School AI, Jam.dev, and Abridge, focusing on how recent OpenAI announcements, particularly the Agent Builder and SDKs, accelerate their work in education, web development tooling, and healthcare documentation by reducing orchestration overhead and enabling faster iteration. **Key Points:** - Caleb Hicks from School AI emphasized that model progression provided leaps in intelligence and cost improvements, enabling them to offer safe, managed AI tutors to students, moving educators from banning AI to recognizing that every student must know how to use it. - School AI utilizes a three-part product: a basic AI assistant tuned for schools (Dot), tools for creating structured outputs like lesson plans, and the special feature of creating guard-railed, safe AI tutors with real-time teacher dashboards. - Danny Grant from Jam.dev announced 'Please Fix,' a tool allowing non-engineers to edit a live site like a Google Doc via a browser extension and submit clean, design-system-compliant pull requests, aiming to eliminate bottlenecks in development. - Both School AI and Jam.dev highlighted the value of using OpenAI's tools internally to build their own products faster, echoing the sentiment that internal tool creation often yields the best products, contrasting with traditional startup advice to 'do things that don't scale.' - Zach Lipton of Abridge, an AI platform for doctor-patient conversations, shared that their tool saves doctors up to an hour or more daily, alleviating 'pajama time' documentation and leading to profound emotional relief, with one doctor citing the tool was 'saving my marriage.' - Zach expressed excitement for the Agent Developer Kit creating a common platform for orchestration, allowing Abridge to focus more on content and specialized models, especially concerning hallucination management where they achieve about 97% recall in error detection. - Lee Robinson from Cursor noted that while initial hopes were that code completion would 'solve code,' the reality requires more complex agents, and their internal culture prioritizes dogfooding, measuring feature success by internal adoption rates before external release. **Context:** The OpenAI Podcast episode was recorded live from DevDay, featuring interviews with developers from various domains: Caleb Hicks from School AI (education technology), Danny Grant from Jam.dev (developer tooling), and Zach Lipton from Abridge (medical AI). The core theme revolved around how the new tools announced at DevDay, such as the Agent Builder and SDKs, specifically impact their ability to build specialized, high-value applications by abstracting away complex orchestration layers and improving development velocity across education, web maintenance, and clinical documentation. ## Detailed Analysis The discussion opened with Caleb Hicks of School AI, whose work focuses on creating safe, personal AI tutors for students; he noted that recent model improvements offered necessary intelligence leaps alongside crucial cost reductions for their industry. He detailed their product stack, which includes an AI assistant named Dot, utility tools for content generation, and specialized AI tutors that provide teachers with formative quizzes and social-emotional check-ins via real-time dashboards, addressing the impossible choice teachers face when supporting 300 students. Danny Grant introduced Jam.dev's new tool, 'Please Fix,' which lets non-developers edit live websites and instantly generate clean PRs, illustrating a potential 'Web Four' where experiences are less mechanical. Zach Lipton discussed Abridge's mission to end physician burnout by automating clinical documentation, saving doctors up to an hour daily and leading to significant personal benefits for users. All guests expressed strong excitement for the Agent Developer Kit, viewing it as a means to stop building their own orchestration tools and instead focus their subject matter expertise—education, design, or medicine—on application-specific problems. Lee Robinson of Cursor emphasized the necessity of using their own tool to build the product, asserting that 'the vibes are important' alongside quantitative evals, and noted that internal feature adoption metrics drive external releases. A recurring theme was the high stakes in their respective fields (education, medicine) requiring rigorous trust-building, evaluation suites, and bespoke approaches to issues like hallucination. ### DevDay Impact on Development Velocity - Agent Builder excitement - Reduced orchestration overhead allows focus on subject matter expertise (School AI, Abridge) - Jam.dev leveraging new tools to iterate on their pricing page instantly. ### School AI's Educational Focus - AI tutors provide personalized learning - Teachers receive real-time dashboards on student AI interaction - Educators are shifting focus to teaching AI literacy. ### Jam.dev & The Future of Web Editing - 'Please Fix' extension allows non-engineers to create submission-ready PRs - This enables entire creative teams to move at the speed of the fastest developer, removing bottlenecks. ### Abridge in High-Stakes Healthcare - Tool automates documentation, saving doctors 'pajama time' and reducing burnout - Trust building is continuous, focusing on security and consistent delivery in partnership with hospital systems. ### Addressing AI Reliability and Trust - Abridge defines hallucination bespoke to medical context, achieving 97% recall in error detection - Cursor prioritizes internal adoption metrics ('vibes') for feature prioritization over developer suggestions alone. ### Developer Advice and Startup Philosophy - Lee Robinson noted that startups often pivot from being internal tools that solved massive internal friction - Founders must love the problem because they will work on it for a decade.