- #ai
- #llm
- #openai
- #research
Inside OpenAI: AGI, RSI Day, and the Rise of Coding Agents
Simon Willison offers a rare glimpse into OpenAI's internal workings, revealing "RSI day" for Recursive Self-Improvement and the pervasive use of coding agents.

Simon Willison offers a rare glimpse into OpenAI's internal workings, revealing "RSI day" for Recursive Self-Improvement and the pervasive use of coding agents.

Learn to create interactive 3D face masks using MediaPipe's face landmarks, Google's canonical face model, and the power of Three.js. This tutorial guides you through integrating these technologies for real-time web applications.
Simon Willison shares Terence Eden's stark view that DNS has become a primary vector for scams, citing an Interisle report revealing that 10-20% of new gTLD registrations are abusive. This piece highlights a critical web security issue.
Jakub Pachocki of OpenAI shares his reflections on the increasing capabilities of AI and the critical challenge of alignment. He advocates for stronger safeguards and international cooperation to manage advanced AI.
Simon Willison shares his candid thoughts on why rewriting software systems from scratch rarely works as intended. He argues that it often results in two incomplete systems and persistent technical debt, rather than a clean slate.
This minor pnpm release introduces workspace dependency resolution via `workspace:` protocol and new `trust-lockfile` options for `remove` and `update`.
Discover how OpenAI is leveraging coding agents internally to reshape its AI research process. This article shares early data on agent usage, experiment velocity, and the impact on task complexity.
Simon Willison shares a powerful quote from Zach Kehs, highlighting that unlike physical structures, software has no inherent limit to how bad its code can become, always allowing for more indirection or performance reduction.