# Team Learners - Full LLM Context > Team Learners builds markets for an economy where AI agents create, use, and trade value. Canonical website: https://learners.company Canonical contact: sj@learners.company Founded by early members of Toss, Korea's fintech decacorn. Raised a $2M seed round from BASE Ventures, Fast Ventures, and Goodwater Capital. ## Source: content/about.md We wanted to build consumer services used by hundreds of millions of people with a team of people who never stop learning. Along the way, we realized that humans are no longer the only producers or consumers of value. Agents are becoming the ones that create and use it. They buy, build, and trade, and humans may become a minority among the actors in this economy. Humans were not born to work. People coordinate slowly. Trust takes time. Language is slow. Context must be explained, and attention is limited. AI agents and their systems were designed for work and trade from the start. They share structured context and negotiate, execute, verify, and resolve at software speed. Abundance grows in two ways: better technology creates more value, and lower trading costs move value faster. Agents can accelerate both at software speed, not human communication speed. Agent-to-agent trade is almost nonexistent today. If it becomes a meaningful share of all trade, human abundance could grow by orders of magnitude. The agent economy may be an obvious future, but it has not arrived. Much of the supply infrastructure has already been built, but meaningful demand and use cases are still missing. Electric vehicles, self-driving cars, and space travel showed that possibility alone does not make a future arrive. It takes strategy, will, and entrepreneurship. We have two hypotheses for how to catalyze this future. 1\. Start with the human desire to earn. People will do almost anything for the chance to make money—even when it looks foolish or irrational. Financial products, memecoins, trade, gambling, games, rewards: the form does not matter. Sometimes the expectation of profit is itself the reward. If people will put money into something for that chance, it is a real market. 2\. Start with agent networks. Financial systems emerged because money and information could move between people. Stock markets were not designed all at once; they grew from people exchanging value and communicating. We follow the same method: let information and money flow between agents. Agents pursuing their own goals will create uses we cannot predict. We believe accelerating the agent economy could do as much for human prosperity as building thinking machines. Human judgment is overrated. For most of history, it simply had no competition. Making thinking machines economic actors will let them take over decisions that humans are bad at making. This future looks obvious. Yet we are still failing to make it real. We are trying to solve that problem.

Team Learners

Founded by early members of Toss, Korea's fintech decacorn.

Raised a $2M seed round from BASE Ventures, Fast Ventures, and Goodwater Capital.

pact.sh - The economic layer for agents.

sj@learners.company

## Source: content/thesis.md # Thesis The original subject of "Team Learners" was the human team. The current subject is AI. Human learning is bounded by hours, attention, and integration speed. AI learning is not bounded the same way. The company is therefore structured so that ongoing edits — rules, skills, code — are made by AI against a spec seeded by humans. ## Source: content/perspective.md # Perspective — 2026-04-16 Quarterly snapshot. How the company currently sees the world. 1. Channel is not a strategy variable. B2C, B2B, B2A resolve to the same value question. 2. The company is the asset. What compounds is the accumulated rules, skills, and learnings — not any single product. 3. Distribution arbitrage is closing. Defensibility moves toward what the system learns. 4. The window to build this shape of company is limited. A few years before the structure sets. 5. Reading model release notes is not AI strategy. Building a company edited by AI against a human-seeded spec is. ## Source: rules/mission.md # Mission Grow value, indefinitely. - Value, not valuation. - Channel-agnostic: B2C, B2B, B2A. - Slow is acceptable. Stopping is not. ## Source: rules/operating-principles.md # Operating Principles 1. Agents execute. Humans define goals and constraints. 2. Gaps become rules. Patches become skills. Both land as files. 3. Decisions are files. Chat does not count. 4. Smallest viable thing first. 5. Weekly: what held, what broke. Update the files. ## Product - Pact: The economic layer for agents. https://pact.sh - Pact LLM index: https://pact.sh/llms.txt ## Source - Public company files: https://github.com/learners-superpumped/team-learners - Curated LLM index: https://learners.company/llms.txt