Socratic tutor for the LLM age
People increasingly get instant correct answers from AI and lose the ability to absorb knowledge. learnerbot deliberately reintroduces productive struggle — one concept at a time.
Ask ChatGPT, get the answer, feel productive, retain nothing. The slow, durable way of building knowledge — concrete examples, then names, then higher abstractions — has been quietly traded away for an autocomplete loop.
Problem-first
A module opens with the concrete problem it solves — never with a taxonomy you have to memorize. You feel why an idea is needed before you meet its name.
Choose, then reason
Questions answer with multiple-choice options and a "what do you think?" — not a verdict. The wrong options are the actual common misconceptions, not strawmen.
Guide, don't grade
When you answer wrong, learnerbot asks for your reasoning and works *with* your mental model toward the insight — instead of marking you down.
Three steps. No setup, no infrastructure, no app to install — just open the chat.
Anything you want to actually understand — system design, statistics, immunology, the Curie family. learnerbot researches it, including the common ways people get it wrong.
Modules open with a real problem. Concepts come as multiple-choice questions where every distractor is a misconception a real learner would pick — not a strawman.
The tutor asks why you chose what you chose, then works from your mental model toward the insight. If you circle the same concept three times, it eases off pure Socratic mode and helps more directly.
Socratic dialogue is powerful but exhausting. After a few rounds of going in circles on a single concept, learnerbot drops the strict Socratic stance and helps more directly — because productive struggle stops being productive when it becomes frustration.
Open the tutor and try a subject you've been meaning to actually understand.