the anti-chatbot learning workspace
The Learn workspace is a 7-step coaching loop that teaches like a tired-but-fair Stack Overflow senior who actually wants you to learn. State your problem properly, show your work, and get taught, not spoon-fed.
You posted code, a real error, and what you expected. We can work with this.
Your coroutine never runs because you call it without awaiting it: do_work()
returns a coroutine object, it doesn't execute.
import asyncio
async def main():
await do_work() # await, don't just call
asyncio.run(main())
Run it. Then come back and tell me why it works now, not that it works. I'll grade you.
example output. Your mileage, and your effort, may vary.
One question becomes a full session. The rail walks you through every step, in order, and remembers where you stopped.
The effort gate. Three required fields: your problem, what you think is happening, and what you've tried. Yes, all three. 'It doesn't work' is a mood, not a bug report.
The concept behind your bug, explained at three depths so you can go as deep as you need. The why, not just the fix.
Drills generated from your actual question, so you rehearse the exact thing you just got wrong, not generic flashcards someone else authored.
It answers wrong on purpose. One deliberate, real flaw, calibrated to your level. Catch it before the reveal, because correcting beats reading.
Re-explain the fix in your own words and get graded. Copy-paste is not a learning outcome. You only understand what you can re-explain.
Your real misses get filed and scheduled back at you, so the thing you learned today is still there next week.
“The best way to get the right answer on the internet is to post the wrong answer.”
People correct faster than they answer. Cunningham mode hands you a deliberately flawed answer and asks you to find the bug. Catch it and you have understood it.
“If you can’t explain it simply, you don’t understand it.”
Every answer ends with a challenge to re-explain the concept in your own words. The coach grades your explanation, not your vibes.
The whole pedagogy is free: the effort gate, theory, practice, Cunningham, Feynman, and your scheduled review. You only ever pay for capacity and continuity, never for the learning itself.
Want unlimited questions? Bring your own Hugging Face token. It goes from your browser straight to the model on each request. We never see, store, or log your key, and frankly we don’t even want to see your code.
No dark patterns. The meter is visibly honest, the limits are stated in the app instead of buried, and cancellation is one line, honored instantly. We would rather you be annoyed at the math than oversold.
It’s faster than you think, and you’ll actually remember it.