The Post-Prompt Era: Alex Graveley on Delegating Work to AI

AI is moving from something we prompt to something we increasingly delegate real work to. In this episode of the ODSC Ai X Podcast, Sheamus McGovern speaks with Alex Graveley, co-creator of GitHub Copilot and a key contributor to Perplexity Computer and Comet, who is now building UFO AI. Alex has had a front-row seat to several major shifts in how people interact with AI. He worked closely with OpenAI during the early development of GitHub Copilot, before ChatGPT, and admits that at the time he was skeptical of some of the more ambitious predictions around AGI. His view has changed considerably as models have become capable of working for longer periods, using computers, recovering from errors, and completing increasingly complex tasks. The conversation explores what happens as AI moves beyond copilots and chat interfaces toward persistent systems that can maintain organizational context, run many tasks in parallel, and take on meaningful responsibility. Alex also explains how UFO is approaching shared memory, agent orchestration, security, model routing, and the possibility of much smaller AI-first companies. Key Topics Covered From Copilot to delegation: How Alex’s view of AI evolved from GitHub Copilot to systems that can take on larger pieces of work. Long-running AI: Why better reasoning, computer use, and self-recovery are changing how much work we can hand over to models. Shared memory and context: How persistent company knowledge could become a core part of AI systems. Parallel AI workers: Alex’s experience running many agents at once and shifting from prompting to supervising and redirecting. The changing role of engineers: Why AI developers may increasingly focus on objectives, constraints, review, and orchestration rather than writing code directly. Trust, permissions, and sandboxes: What it takes to safely let AI act across software, data, and business systems. The future of software: Why Alex expects more bespoke, AI-generated software and fewer rigid SaaS workflows. AI-first companies: How smaller teams could delegate large parts of engineering, operations, support, and communication to AI. Open-weight models and routing: Why cheaper open models are becoming more useful and how choosing the right model for each task affects cost and performance. What UFO AI is building: A connected work environment combining shared memory, multiple interfaces, company data, and autonomous execution. Visit our website and choose the nearest ODSC event to attend and experience all our training and workshops: https://odsc.ai To watch more videos like this, visit https://aiplus.training Sign up for the newsletter to stay up to date with the latest trends in data science: https://opendatascience.com/newsletter/ Follow us online! • Facebook: https://www.facebook.com/OPENDATASCI • Instagram: https://www.instagram.com/odsc/ • Blog: https://opendatascience.com/ • LinkedIn: https://www.linkedin.com/company/open-data-science/ • X (twitter): https://x.com/_odsc