Daily Page: 2025-11-17
• 2 min read 2 min • 342 words 342 words
Solving a Million-Step LLM Task with Zero Errors
Welcome to the Third Millennium, Post-Automation Era
`I just read a paper that completely broke my brain.
— Carlos E. Perez (@IntuitMachine) November 17, 2025
It describes a system that solved an AI task with over 1,000,000 sequential steps... with ZERO errors.
Using AI models that are known to be flaky and make mistakes.
How is that even possible? 🤯
We all know LLMs have an… pic.twitter.com/LYqIVQtJzp
What If?
The “Impossible” Math of BOTH Why AND How
A Third Millennium, Post-Automation Era Tool for Hacking
Closer To Truth
Global Cognition Group Cohorts
Models of Productive Intercognition
Three Examples
Theories of Everything, Ekkolápto, The Stoa
Intercognition ⊨ Intercognotion ⊨ Intercogmotion
What if thinking with intention is DOING (James 5:16)?
LEARNING is the Only True Wealth
It’s About Becoming a More Skillful Problem Solver
Especially When Supervising Dopey Vibe Coders
Unsupervised Vibe Code == GARBAGE AI SLOP CODE
What’s Planck Pixels Got To Do With It?
Because:
This:
It is what it is, and what it is, is.