> For the complete documentation index, see [llms.txt](https://docs.scanbylens.xyz/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.scanbylens.xyz/the-product/the-problem.md).

# The problem

Every input a serious market read needs is sold separately. The join is not sold at all.

## The silos

A serious read of the crypto market needs several kinds of evidence at once: what derivatives traders are paying to hold leverage, how much leverage is open, how deep the order books are, how much stablecoin liquidity is sitting on the sidelines, what the dollar and rates are doing, what is being forced out of positions, and what the news actually says.

Each of these is sold separately. Derivatives data from one vendor, depth from another, ETF flows in spreadsheets, macro on a terminal, news in a feed. The synthesis across them, the part that answers "what is actually happening", is produced by analysts at institutional desks, weekly, for paying clients. For almost everyone else it does not exist.

## The noise

What does exist is the daily explainer: "bitcoin fell because of X." It is published everywhere, usually without sources, and it is mostly noise dressed as insight. A language model pointed at raw feeds produces exactly that, confidently and forever. A model does not know today's funding rate or this morning's liquidations, and when asked, it narrates a market that may not exist.

## Two failures

In 2026 the market split into two kinds of AI research product, and both fail the reader.

**Explainers that will not take a side.** Exchanges and data sites now explain price moves for free. They describe; they do not conclude. The reader is left to do the judgement themselves, which was the hard part all along.

**Callers that keep no record.** Confident AI agents make calls on social media, often without primary sources, and advertise hand-picked wins. Without a scored record, confidence is cheap and trust collapses the first time it is tested.

The evidence on how people trust advisers points the same way. Confident advisers keep their influence only while their accuracy is hidden; once accuracy is visible, confident but wrong advisers lose credibility. Longer explanations make people more confident without making the call more accurate. And a system that is willing to say "I don't know" is wrong far less often than one that always answers.

## The gap

The gap is not data, and it is not narration. It is **disciplined synthesis with accountability**:

1. Join the inputs on one clock, each measured against its own history.
2. State only connections that are measured, mechanical, or labelled as conjecture.
3. Commit to one clear call, or say plainly that the data does not support one.
4. Publish every call as it is made, never edit it, and score it in public.

LensAI is built to fill exactly that gap. The rest of this paper explains how.


---

# 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 by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.scanbylens.xyz/the-product/the-problem.md?ask=<question>&goal=<user_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 `automate deployments from our CI pipeline` 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.