why I started Temet
Why I started Temet
A founding manifesto about AI, self-improvement, feedback loops, personal agents, and why Temet exists.
When AI stopped feeling like a tool
I have been using AI since ChatGPT launched in November 2022, and very early on I understood that my work as a product creator would not simply become faster. It would become fundamentally different. At first, the shift looked familiar: writing faster, researching faster, prototyping faster, organizing ideas faster. But after enough hours spent inside these systems, I realized that speed was only the surface. The deeper transformation was that AI could become a continuous layer around my life and work: a layer of memory, feedback, judgment, and applied knowledge that stayed close to me and kept evolving with me.
Building a fleet of personal agents
That realization changed the way I started building for myself. I did not just want assistants. I wanted systems that could help me think better, communicate better, negotiate better, and learn faster over time. So I began creating a small fleet of personal agents, often rooted in books or methodologies I wanted to internalize more deeply. Chris Voss for negotiation. Esther Perel for positive communication and relational intelligence. Marcus Aurelius for self-observation, journaling, and private reflection. Each of these became less like a chatbot and more like a living micro-application: a framework with memory, history, context, and exercises that could meet me where I was and keep pushing me forward.
From builder to architect of augmented intelligence
What fascinated me most was not that these systems could answer questions. It was that they could begin to participate in a loop of self-learning. They could preserve what I had already struggled with, remind me of patterns I kept repeating, and turn abstract advice into small actionable feedback tied to my own behavior. My productivity accelerated, of course, but that was not the most important effect. The more profound shift was that I was no longer just building products. I was becoming an architect of augmented intelligence: someone designing systems that learn with me, improve with me, and quietly reshape how I work.
The realization that changed everything
The turning point came during a speaking practice session. The structure was ordinary enough: three people in a small group, one person speaking, the others listening, reacting, giving feedback, helping refine tone, structure, and clarity. But while I was inside that loop, I felt something click. This was not just a learning format. It was a pattern. Practice, immediate correction, another attempt. Feedback, adjustment, repetition. And suddenly it became obvious that AI could provide this kind of loop constantly. Not occasionally, not when a human coach was available, but every day, all the time, in the middle of real life.
Why traditional learning is too slow
Traditional learning is slow not because the material is bad, but because the loops are too loose. We read books, watch videos, take courses, highlight passages, feel briefly inspired, and then return to ordinary life where almost none of it gets reinforced. A few weeks later, most of it is gone. The problem is not knowledge. The problem is distance. Too much time between the lesson and the attempt. Too much time between the attempt and the correction. Not enough repetition, not enough implementation, not enough pressure from reality. The most effective form of learning is brutally simple: do the thing you want to improve every day, and receive immediate feedback on what you just did. That is exactly where AI becomes transformative.
Why Temet exists
That is why I started Temet. Not to build another assistant, but to explore what happens when AI begins to function as a continuous layer of feedback, memory, and capability around a human being. I believe we are entering a world where people will not just use AI to produce more. They will use AI to become more. To notice patterns in how they work. To turn repeated behavior into explicit skills. To preserve useful memory. To track what changes. To improve through tighter loops. To build software that helps them grow, not only software that helps them output. A world where learning is ambient. A world where self-improvement becomes operational. A world where anyone can have access to a form of coaching that is personal, persistent, and always within reach.
Thanks for reading. Arnaud
FAQ
Is Temet mainly about productivity?
No. Productivity is one effect, but the deeper idea is continuous learning through real feedback loops, memory, and repeated application.
Why start from personal agents and coaching systems?
Because the most compelling use of AI is not only output generation. It is building systems that can help a person improve over time in context.
How does this connect to Temet today?
Temet starts by reading real AI work, surfacing skills and repeated patterns, and building a clearer profile from what your work already proves.
Next step
Use this guide in practice with Temet's audit, tracking, and profile workflow.
Connect your agentPublished January 1, 2026