Skip to content
Grasp

Pilot cohort · opening soon

Explain it well enough and we pay you.

Pick a topic. Learn it however you like. Sit a twelve-minute spoken exam with an examiner that keeps asking why. Pass and the money is yours — and that topic is closed to you forever.

Join the pilot waitlist
First cohort: 100 people. No fee to sit an exam.
Transcript · base rates 00:00
Passed · all four ≥ 3 $10.00
Accuracy4
Depth3
Transfer4
Epistemics3
Two of ten exchanges.

We give you nothing to read.

There are no lessons here, no videos and no course to click through. Where you learn a topic is your business — a library, a lecture, an argument with a friend. All we do is check, carefully, whether you actually understand it, and pay you when you do.

Step 01

Claim a topic

Choose from the open board. Each topic shows what it pays before you start.

Step 02

Learn it anywhere

Take a day or take a month. Use any source you trust. Nobody is watching this part.

Step 03

Sit the spoken exam

Twelve minutes, out loud, in real time. Explain it, defend it, then apply it to a case you have never seen.

Step 04

Get paid, then move on

Passing pays out and closes the topic for good. The only way to earn again is to understand something new.

Four things every exam scores.

Reciting a definition is not understanding, so the marking does not reward it. Every transcript is scored blind on four separate dimensions.

Dimension A

Accuracy

Is what you said true, and do you catch your own slips before the examiner does?

Dimension B

Depth

The examiner asks why, twice more, on your weakest sentence. Memorised answers stop here.

Dimension C

Transfer

A fresh scenario, generated for your exam alone. Arithmetic slips are forgiven. Structure is not.

Dimension D

Epistemics

Where does the idea break down, and what would change your mind? Knowing the edges counts.

Passing needs 3 or better on all four. Nothing is averaged — being right without being able to apply it is exactly the failure this exists to stop rewarding.

The opening board.

Everything currently open, with what it pays. New topics go up every week, so the board is meant to stay ahead of you — if you ever clear it, that is our problem to fix, not your ceiling.

$1,000 on the board today, if you passed every topic open right now
41 topics you can claim, from $10 foundations to $50 unscripted
Weekly new topics added, so there is no total to finish and no cap on what you can earn
Confirmation bias Foundations $10
Correlation and causation Foundations $10
Base rates Foundations $10
Survivorship bias Foundations $10
Sunk cost Foundations $10
Regression to the mean Foundations $10
Anecdote versus data Foundations $10
Valid versus sound arguments Foundations $10
The availability heuristic Foundations $10
Motivated reasoning Foundations $10
Falsifiability Foundations $10
Necessary and sufficient Foundations $10
Bayes' theorem, applied Machinery $20
Expected value Machinery $20
Natural selection Machinery $20
Compound growth Machinery $20
What a p-value does not tell you Machinery $20
Confidence intervals Machinery $20
Sampling and selection bias Machinery $20
Opportunity cost Machinery $20
The prisoner's dilemma Machinery $20
Signal and noise Machinery $20
Conditional probability Machinery $20
Marginal thinking Machinery $20
Placebo and control groups Machinery $20
Statistical power Machinery $20
Reading a study end to end Integrative $35
Goodhart's law Integrative $35
How a randomised trial works Integrative $35
The scientific method as error correction Integrative $35
Evaluating a news claim from scratch Integrative $35
Significance versus importance Integrative $35
Cost and benefit under uncertainty Integrative $35
Replication, and why it fails Integrative $35
What peer review does not do Integrative $35
Meta-analysis Integrative $35
Steelmanning a position you reject Unscripted $50
Naming what would change your mind Unscripted $50
Designing a test for a claim you hold Unscripted $50
Tracing a belief back to where it came from Unscripted $50
Auditing a decision you got wrong Unscripted $50

Fail and you get the full scored transcript telling you exactly where the understanding ran out, plus a retry after 72 hours. The feedback is free either way.

Why would anyone pay you for this?

It is a fair thing to be suspicious about. Money offered for not much is usually a trick, and anything promising to teach you how to think is usually selling you what to think. So here is the whole thing plainly, including the parts that would be easier to leave out.

Where the money comes from

Out of one person's pocket. There is no investor, no token, no advertising, no fee to sit an exam and nothing to buy at any point. The pilot is a hundred people rather than a hundred thousand precisely because it is funded personally. If that ever changes — a foundation funding a topic, a company paying to verify a record — it will be written here before it happens rather than after.

What happens to the recording

The exam is a spoken conversation, so there is a recording and a transcript. They exist to be graded and handed back to you. They are not sold, not published and not shown to anyone else. Delete both whenever you like — the pass stays on your record.

Whether this is an AI training scheme

Twelve minutes of someone reasoning carefully out loud is genuinely valuable material to an AI lab, and it would be reasonable to assume that is the real product and the payout is the bait. It is not. Nothing recorded here is sold or licensed for model training, and the providers handling the audio are contracted not to train on it. If that position ever changed, the honest thing would be to stop running exams rather than quietly edit a policy page.

Whose idea of correct is this?

The topics have settled answers — base rates, natural selection, what a p-value does not tell you. Nobody is scored for agreeing with anything. A quarter of every mark is for naming where the idea breaks down and what would change your mind about it, so the exam pays you for doubting the material rather than repeating it. Where reasonable people land in different places, the mark is for how you got there. If you find a doctrine in here, that is a fault worth reporting.

Why passing closes a topic for good

Because paying repeatedly for the same understanding turns this into a job, and jobs get optimised. Retiring a topic the moment you pass it means the only way to earn again is to understand something new. That is the entire mechanism, and it is also why nobody can grind it.

What it costs you

Nothing. No entry fee, no subscription, no card on file, and failing costs you nothing but the seventy-two hours before you can sit that topic again. The only thing asked of you is that you actually learn the thing.

Who is behind it

I grew up in a high-control religion. I was taught what to think and never how to think, and when it came apart the question that stayed was: if I could be that wrong about something I would have risked my life for, what else am I wrong about? Grasp is what I wanted to exist while I was working that out — something that checked whether I actually understood a thing rather than whether I could repeat it back.

— the person who built it

A certificate says you attended. This says you can be questioned.

Every pass is a recorded exam you could be asked to defend again. Over time that record is worth more than the payout that prompted it.

Waitlist

Sit one of the first exams.

The pilot takes 100 people. Tell us which topic you would go after first and we will use the answers to decide what goes on the board next.

One email when the pilot opens. No newsletter, no sharing your address with anyone.