Full Workshop: Setting Yourself Up for Success —Jason Liu, OpenAI Codex
Jason Liu from OpenAI shares advanced strategies for using Codex to automate complex workflows and manage personal information. He introduces key concepts like "compaction," dictation for input, and "appshots" for contextual awareness. The talk covers bringing context into the system via plugins and a personal memory vault, working with AI through automations, goals, and interconnected threads, and taking actions out in the real world, emphasizing the transformative power of "Computer Use" for broad system control.
Jason Liu, from OpenAI’s Developer Experience team, provides a detailed guide on leveraging Codex as a powerful tool for computer control and workflow automation. He begins by highlighting that many traditional AI best practices, such as starting new threads for long conversations, are now obsolete due to advancements like “compaction.” Liu emphasizes that the goal is to avoid wasting tokens by adopting effective strategies.
Core Principles for Maximizing Codex Productivity
- Compaction Works: Codex threads can maintain long, complex conversations (e.g., five weeks, 400 sub-agents) while retaining context, making the old advice of starting new threads for every task unnecessary.
- Talk to Your Computer: Dictation is significantly faster (three times) than typing and allows users to input