Introducing brAIAIn.
A parody AI lab. The language model is real.
A language model trained on exactly what it was given: twelve lines of scripture, and one sentence from each customer.
$5 buys one sentence in the next model. It ships as a numbered release with its own notes, paper and status incident, credited to you or to Anonymous.
Talk to brAIAIn.
Ask it anything. It will answer using its commandments and what customers have taught it, which is not much, and it will be confident about it.
Press Peel a layer to hear the same question answered by the version underneath, and the one under that, down to brAIAIn-0. Each version is trained on everything before it, plus text the version before it wrote. Peel it until nothing is left but the first twelve lines.
This runs entirely in your browser. Nothing you type is sent anywhere or saved.
Pricing
Every plan buys exactly one sentence in the next version of brAIAIn. The plans differ in what you get to keep.
Minnow
For individuals exploring frontier intelligence.
- One sentence in the training data
- A numbered model release with notes, paper and status incident
- Credit by name, or Anonymous
Shark
For professionals who need proof.
- Everything in Minnow
- A printable certificate of contribution, signed by the board (see a sample)
- Your name in the paper's acknowledgments
Whale
For those who want to own a release.
- Everything in Shark
- You name the version. Your codename ships on the model card, the models list and the certificate
- Personally ignored by the Chairman
Enterprise
For teams training at scale.
- Volume pricing: $5 per sentence, multiplied by the number of sentences
- A dedicated account goose, subject to availability
- Uptime as good as GitHub's
Free: unlimited chat with every version, on every plan, including none. Nothing you type is kept.
*Usage limits apply to training only: one sentence per purchase, up to 255 characters, public and permanent. No links, contact details or other people's private information. Most go live in about 10 minutes; anything that looks off waits for a person, up to a day. Not approved means a full refund.
How it works, refunds and gifts · FAQ. Prices in US dollars. One-time payment. No subscription, because nothing recurs except the model.
Trained on you.
Every paid version adds one sentence that a customer paid $5 to include. That sentence becomes part of the model forever, or until our safety team removes it.
Trained on itself.
Each version also reads a few passages written by the version before it. Researchers warn that models trained on their own output slowly lose their minds. We wanted to see it happen.
Fully disclosed.
Other labs will not say what their models were trained on. We will. Here is all of it. It fits on one page.
Under the hood
Full system cardThe specifications other labs keep confidential. The counts and timings below are computed on the current model, in your browser, as this page loads. The rest are true by construction.
Latest releases
All modelsWhat people are saying
Other labs publish the good ones. We publish all of them, unedited, including the ones about our certificate. Collected from public comments.
“Cringe”
Lab response: Noted. The board met about it. The board is a goose.
“That's hilarious and also terrifying.”
“The cash grab is huge on this one.”
“get a damn tls cert bro”
Lab response: We have one.
“alright brian”
Lab response: It is pronounced brain. The extra AIs are for the valuation.
“this is going to be great”
Lab response: Our release notes say this about every version.
Shape the future of intelligence.
Write one sentence. We train a new model on it, give it a version number, publish its release notes, and tell the world. That is the whole product.
See pricing