Questions & answers

What this research is. And what it is not. How Unbroken Take works, how to read its public-disclosure research, and the legal and practical boundaries that matter.

Last updated
August 14, 2026
Scope
Purpose / research / boundaries
01

About the project

Purpose, process, and selection.

What is Unbroken Take?

Unbroken Take began with a simple idea: individual investors have access to more information than ever, but information is not the same as research.

Investor-tracking platforms are good at showing what someone bought, sold, or held. They are less useful when the more important questions begin: What kind of business is behind the filing? Why might a long-term investor find it interesting? How is its competitive position changing?

Recent advances in AI make it possible to approach these questions with greater depth and consistency. Unbroken Take applies that capability to long-term, value-oriented company research, focusing on business models, competition, capital allocation, and meaningful changes in the underlying business.

Institutional research can be excellent, but it is written for particular audiences and purposes. Unbroken Take offers a different perspective: independent research based on public information, intended to help individual investors understand a business and identify questions worth investigating before reaching their own conclusions.

How is AI used?

Unbroken Take is influenced by Harness Engineering, an approach developed around AI coding agents in which reliability depends not only on the model, but on the environment built around it: persistent context, explicit instructions, tools, validation, and feedback loops.

The same principle can be applied to company research. Rather than asking an AI to “analyze a company” in a single prompt, Unbroken Take treats the research methodology much like a codebase: documented, constrained, reviewed, and improved when it fails. Research principles define what should be examined, source rules determine what evidence should carry weight, and persistent project documentation preserves context across different companies and research cycles. Separate review steps then challenge factual claims, inconsistencies, and reasoning. When recurring problems are identified, the underlying methodology is revised so that the correction improves future research rather than only one article.

AI agents are therefore used within a research system, not as a substitute for one. They assist with gathering, comparing, structuring, challenging, and drafting information. Published research remains subject to substantive human review and editorial control, and AI output is never treated as a source.

How are superinvestors selected?

The investors followed by Unbroken Take are selected for the research value of their disclosed portfolios, not simply for fame or assets under management.

The strongest preference is for investors with concentrated portfolios, long and credible track records, and a clearly identifiable investment philosophy. Their holdings should reflect deliberate fundamental judgment rather than broad diversification, short-term trading, or strategies that cannot be meaningfully understood from public disclosures.

02

Research & disclosures

How to read the evidence.

What can a Form 13F filing tell me—and what can it not tell me?

A Form 13F is a delayed quarterly snapshot of certain reportable U.S.-listed long positions. It can show that a manager reported a position at a quarter end and, when comparable filings are available, how the reported position changed. It does not reveal the exact trade date or price, the manager’s full portfolio, cash, short positions, or the investment thesis. Amendments and confidential-treatment requests can also change what is visible and when. The SEC’s Form 13F frequently asked questions explain the filing requirements and exceptions.

A reported holding or change is therefore a research lead, not proof that a security is attractive and not an instruction to copy a trade.

How current and accurate is the information?

The site preserves reporting periods, filing dates, source context, and known gaps where practical. Data is derived from public regulatory disclosures and structured processing, but completeness, timeliness, and accuracy are not guaranteed. Filings may be delayed, amended, incomplete, or superseded, and processing errors are possible. Verify material facts in the original SEC filing before relying on them.

How is published company research reviewed?

AI agents may gather, compare, structure, challenge, and draft public-source research. Published work is subject to substantive human review and editorial control. The site aims to distinguish sourced facts from interpretation, keep uncertainty visible, and identify evidence that could disconfirm a thesis. Publication does not make a conclusion final; new filings or better evidence may require revision.

04

Questions & contact

Questions and feedback are welcome.