Evaluate sources and write a personal operating policy

Use a source ladder

Match the source to the kind of claim. For how a protocol behaves, start with its specification and official documentation. For what the law requires, start with the regulator or the legislation itself. For what a product does, start with its terms, its published security reviews, its status page, and behaviour you can verify on-chain.

Journalism and creators occupy a real but secondary position. Good reporting points at primary evidence, and its value is in finding and contextualising that evidence—not in replacing it. When a claim matters, follow it back to whoever is actually responsible for the fact.

Note also that the loudest sources are selected for engagement rather than accuracy. Anything that reaches you without your having sought it out has passed through a filter optimising for attention, which is uncorrelated with being right.

Follow incentives

For any consequential claim, ask six questions. Who wrote this? How would they know? What do they sell, hold, or get paid for? Is any payment or position disclosed? What date and jurisdiction does it describe? And what evidence would change the conclusion?

The last one is the most discriminating. Someone who can describe what would change their mind is reasoning. Someone who cannot is advocating, however technical the vocabulary. Confidence and popularity are not methods, and in this field they correlate with promotion more often than with accuracy.

Write rules before pressure

The purpose of a written policy is to move decisions from moments when you are rushed, excited, or frightened into a moment when you are neither. This is the same reason a test transfer is written down in advance: the version of you that encounters an urgent message is not the version that should be setting limits.

Cover the recurring situations: unfamiliar links, installing new wallets, test amounts before a new route, approval limits, whether you use leverage at all, maximum position size, how backups are handled and where, what records you keep, and exactly who you contact after an incident. Keep it short enough that you actually remember it.

Review the policy after any mistake, not only after a loss. Near misses are the cheapest information you will ever get, and they are the only feedback available before something expensive happens. Review it again after a significant change in what you hold or in your circumstances.

• No seed phrase in a website, chat, photo, or cloud note—ever.
• Every new route starts with a test that is actually credited.
• No decision made under artificial urgency.
• Approvals limited in amount and reviewed periodically.
• Legal and tax claims require current sources for my own jurisdiction.
• One named person or channel to contact after an incident.