Turn anything into an interactive course

Describe a topic or drop a PDF. You get a course of interactive lessons, quizzes, and instant feedback. You learn by solving, not skimming.

Try:

Free to start · no credit card

On the free plan, courses you create are public — your chats stay private

2Why sunsets turn orangeWhy is the sky blue? · lesson 2 of 4

Slide the sun down toward the horizon and watch the sky change.

Why does the sky go orange at sunset?

Pick an answer to check it.

From a sentence to a course you can solve

1

Describe it

A topic, a paper, a PDF, or an internal directive. Tell it who the course is for.

2

Watch it build

Lessons appear one at a time, each with a widget or quiz. Keep chatting to reshape any of them.

Solve your way through

Every lesson is graded instantly. Spaced review brings back what you got wrong.

See it in action

Why not just ask ChatGPT?

Ask a chatbot to explain something and you get a wall of text — read once, forgotten by Friday. Moldavite uses the same class of AI; the difference is everything that happens after the explanation.

Active recall, not passive reading

Every lesson makes you solve, answer, and check yourself as you go — retrieving beats re-reading, and a chat never asks you anything back.

Spaced review, so it sticks

Quizzes you answer come back right before you'd forget them — spaced repetition scheduled from how you actually did.

Progress you can see

XP, streaks, and per-lesson progress across a real course structure — not a conversation that scrolls away.

A course, not a chat log

Structured lessons you can revisit, share with anyone, and reshape by chatting — "make lesson 2 more visual" just works.

Built by other learners

Browse all courses

#Computer Science✦ Claude Sonnet 4.6

Longformer: The Long-Document Transformer

A hands-on course for ML practitioners and NLP researchers covering Longformer (arXiv:2004.05150) end-to-end: why standard Transformers fail on long documents, how Longformer's three-part attention mechanism scales linearly, the LED variant for summarization, pretraining from RoBERTa, and empirical results on character-LM, QA, and summarization benchmarks.

Created Jul 10, 2026

#Uncategorized✦ Claude Sonnet 4.6

Graph Representation Learning

From lookup tables to message-passing neural networks: how graphs became differentiable, and why that changes everything for fraud detection, chemistry, and beyond.

Created Jun 5, 2026

#Computer Science✦ Claude Sonnet 4.6

Flash Attention: Fast & Memory-Efficient Transformers

A presentation-ready deep dive into Flash Attention — the IO-aware algorithm that makes Transformers fast on long sequences. Covers the exact mechanism, the math, tiling & recomputation tricks, and real NLP applications.

Updated Jul 13, 2026

#Computer Science✦ Claude Sonnet 4.6

Byte-Pair Encoding (BPE)

Master Byte-Pair Encoding (BPE) — the subword tokenization algorithm powering GPT, LLaMA, and most modern LLMs. Learn how it works step by step, why it beats word-level and character-level approaches, and how it handles rare and unknown words.

Updated Jul 15, 2026

Start with the thing you've been meaning to understand.

Free to start · no credit card · built by one person who couldn't find this.