OpenAI’s GPT-6 and RSI rumors get a source map with internal compute and self-play clues
imjustnewatai · x · 2026-07-26
A second thread compiles the evidence behind the GPT-6 and RSI speculation
This follow-up post gathers links behind the claim that OpenAI is building a recursive self-improvement loop:
- an unnamed pre-release model more capable than GPT-5.6 Sol
- GPT-5.6’s RSI Index and the reported 100× growth in internal coding-inference compute and 22× growth in agent-token usage
- GPT-Red self-play and OpenAI’s comments about next-generation models
- statements that GPT-5.3 Codex helped create itself and GPT-5.5 improved its own serving stack
- Sam Altman’s “whole cluster” tease and comments about GPT-6 memory/personalization
- Bloomberg reporting that Altman will brief officials on OpenAI’s upcoming generation
The post also points to roadmap items such as recursive self-improvement safety, synthetic RL, self-play, personalized superassistant work, long-term memory, evolving user models and closed-loop automation for Stargate.
It is essentially a source map for the GPT-6 / RSI narrative rather than a separate announcement.
More from AGI Musings
- Adam Marblestone's Podcast Reading List: Evolution of Intelligence to Digital Minds — KordingLab · 2026-09-11
- Superintelligence will be maximum good, not stupid or evil, argues Patterson — davidpattersonx · 2026-09-11
- Mathematician Daniel Litt Launches Problem Repo to Track Human vs AI Progress: 15 Problems, 1 Solved — littmath · 2026-09-11
- Should AI models be taught morality? Breakout incidents expose missing ethical training — Pfungus_ · 2026-09-11
- SoftBank's Masayoshi Son predicts 100 trillion self-replicating AIs: "humans' era as top life form is ending" — Puzzleheaded-King584 · 2026-09-11
- We are witnessing the unreasonable effectiveness of inference-time scaling — sqcai · 2026-09-11