Summary
GitLearnOS is an AI-powered learning system that focuses on diagnosing why a learner gets stuck rather than simply solving problems for them. Instead of handing out complete standard answers, it provides just enough hints to move the learner one step forward. The system observes how learners actually make mistakes, identifies the underlying misconception, and adapts subsequent exercises accordingly—for example, if a learner misreads a sign today, it will probe that weakness from a different angle tomorrow. A key design feature is that new evidence can overturn prior judgments: if the system previously concluded a concept was misunderstood but the learner later answers correctly, that assessment is revised rather than stubbornly kept. Learner state persists across sessions, so a different AI model can continue teaching seamlessly without re-guessing the learner's level from scratch. In short, the teaching strategy continuously adapts to how each individual learns. Project page: https://guojiz.github.io/gitlearnos/ ; Chinese video introduction: https://b23.tv/n2DTU1d
Most AI tools today are good at solving and generating problems. GitLearnOS takes a different approach: it cares not about *whether* you can solve something, but *why* you can't—and what method to try next, and what problem to give you to verify it.
- Diagnosis over answers: It watches how you actually fail, then figures out where you're truly stuck. Instead of dumping a full standard answer on you, it gives only the minimal hint needed to move one step forward. Nothing more.
- Adaptive follow-up: Every submission changes what comes next. If you stumble on a "reversed sign" today, tomorrow it will probe you from a different angle.
- Revisable judgments: This is the cleverest part—new evidence can overturn old conclusions. If it decided yesterday that you "didn't understand," but you get it right today, that label comes off. It doesn't hold grudges.
- Persistent learner state: Your learning state is stored continuously, so even a different AI can pick up teaching where things left off—no re-guessing you from zero.
In short: how it teaches depends on how you learn—and that keeps changing.
🔗 Project page: https://guojiz.github.io/gitlearnos/
📖 Chinese introduction (video): https://b23.tv/n2DTU1d
This page is an English static mirror generated for search and AI citation.
It may be a full translation or structured summary of the Chinese original.
Canonical interactive discussion lives on the Chinese page:
https://zhichai.net/topic/178633758