> For the complete documentation index, see [llms.txt](https://docs.brainnetwork.app/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.brainnetwork.app/the-idea/what-we-refuse-to-fake.md). # What we refuse to fake Most distributed-compute projects die of the same disease: the dashboard is more impressive than the network. BRAIN was built with a short list of things it will not do, and the list is enforced in code and tests, not in a values page. ## Never trusted from the client Contributors are assumed to be adversarial. The server does **not** trust any of the following when a browser or node reports it: * GPU model or name * VRAM * Compute units completed * Job completion * Benchmark score * Uptime * Token counts beyond what the coordinator streamed itself Reported values are stored as `reported`, labelled **REPORTED** wherever they are shown, and used only to *exclude* a node (a card that reports 8 GB is never sent a model that needs 20 GB). They never raise a score or a payout. ## Never mixed Every receipt, accounting event, order, treasury record, epoch and credit event carries `source: "REAL" | "SIMULATED"`. Stores index on it, snapshots filter on it, and the test suite asserts that simulated data cannot enter a real total, that a simulated earning can never offset credits, and that a client cannot self-report verified compute. By default the site shows **real only**. A footer control, "Show simulated data", blends a simulated network in for people who want to see how the product reads at scale, and every such figure gets a **SIM** badge. The choice lives in your browser; it changes nothing on the server. ## Never claimed * No TFLOPS or VRAM figures. Browsers do not expose VRAM, so BRAIN does not display it. * No fabricated benchmark results. Every score shown for your device was measured in your tab and verified by the server. * No "X % cheaper than" comparisons. BRAIN has not measured one and will not invent one. * No token emissions, no staking yield, no price expectations. * No cryptographic proof of execution. The receipt's `resultHash` is an integrity digest anyone can recompute; `attestation` is `{ kind: "none" }` for browser work and `{ kind: "signature" }` for coordinator-signed node receipts. Hardware attestation is **PLANNED** and labelled so. * No "verified" label on an upstream model's response. If BRAIN AUTO routes to an external provider, the receipt says `unverified-provider-response`. * No revenue, customers or savings that are not in the accounting records. Unpriced work is **UNKNOWN**, not estimated. ## Zero verified compute, zero reward This is the load-bearing rule. A wallet with no verified compute in an epoch receives nothing from that epoch, no matter what it holds, how long it was online, or what it reported. The reward formula has a test for it. Benchmark jobs and canaries issue no receipt and earn nothing. The accounting ledger accrues only from receipts. If you remember one thing from this book, make it this one. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters: ``` GET https://docs.brainnetwork.app/the-idea/what-we-refuse-to-fake.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `build a script that syncs our docs to a CMS` lets GitBook tailor the answer to that use case. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.