OpenAI’s internal RSI day highlights rapid research boost from GPT-6 Astra
OpenAI marked a “Recursive Self‑Improvement” (RSI) day, signaling what the company calls its new AGI milestone. Chief Scientist Jakub Pachocki released an essay titled An Alien Mind that, together with internal reports, describes how coding agents are now central to the lab’s research workflow. The shift mirrors a broader 2026 surge in “agentic engineering” at OpenAI, illustrated by a steep rise…
Key points
- OpenAI held an internal “RSI day” to showcase its new AGI‑focused efforts
- Coding agents powered by the soon‑to‑be‑released GPT‑6 Astra drove a sharp rise in AI spend per researcher
- Chief Scientist Jakub Pachocki published the essay “An Alien Mind” outlining the new approach
The spike appears linked to internal staff gaining early access to a model later announced publicly as GPT‑6 Astra. While details remain sparse, the model is positioned as a next‑generation large language model that powers the new coding agents and accelerates experimentation across the organization. The internal rollout suggests OpenAI is moving quickly toward more autonomous, self‑improving systems, a development that could reshape competitive dynamics in the AI field.
Observers note that the rapid deployment of GPT‑6 Astra may foreshadow broader industry adoption of advanced agentic tools, raising questions about research speed, safety oversight, and the timeline for achieving general AI capabilities.
The story so far
4 episodes →- OpenAI’s internal RSI day highlights rapid research boost from GPT-6 Astra this story
Research acceleration: The view inside OpenAI
Simon Willison · 6 September 2026
6th September 2026 - Link Blog
Research acceleration: The view inside OpenAI. Apparently today is RSI day at OpenAI, for Recursive Self-Improvement - I think it's their new AGI. Both this piece and the new essay An Alien Mind (by Chief Scientist Jakub Pachocki) talk about it, and this one doesn't even bother to expand the acronym.
Included are details on how OpenAI's own research team are using coding agents. Like pretty much everyone else 2026 has been the year that agentic engineering really took off at OpenAI, best illustrated by this chart:
I'm intrigued at what caused that significant acceleration in AI spend per researcher in late July - my best guess is that's when internal employees gained access to the model later released as GPT-6 Astra.
Recent articles
This text was published by Simon Willison and written by Simon Willison. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
Coverage and discussion
2 sources- OpenAI chief scientist warns no-one is prepared for consequences of AI Press · BBC Technology ·
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us.
More in Research
All →- AI Researchers Fear Machines Could Kill Us All · 7 src
- Study finds AI agents lack creativity needed for recursive self-improvement · 4 src
- OpenArch launches PyTorch repo for readable LLM architecture implementations · 1 src
- R2VC Boosts Fact‑Checking Accuracy by 13.74% on FEVER Using Modular Retrieval and Calibration · 1 src
- Study Finds Deictic Ambiguity Can Undermine Draft‑Verify‑Revise LLM Pipelines · 1 src
Comments
via GitHub Discussions