There's a version of knowledge that gets published — in articles, books, courses, documentation. It's the knowledge that someone decided was worth writing down, that was legible enough to structure, that could survive the translation from lived experience into text. It's useful. It's not the good stuff.
The good stuff is the knowledge that never gets written down because it's too specific, too contextual, too hard to explain without ten minutes of back-and-forth. It's the contractor who knows exactly which building inspector in the county is strict about a particular code and how to prepare for that inspection. It's the immigration lawyer who has handled 200 applications of a specific type and knows which arguments consistently fail even though they look reasonable on paper. It's the VC who can tell you which metrics actually matter at Series A for your category versus which ones look right but mislead founders.
None of that is on Google. It never was. It's always existed exclusively as expertise — knowledge that lives in people's heads and is only accessible through a relationship with those people.
The Knowledge Iceberg
What's on the internet vs. what experts actually know
The reason this matters now more than it used to is that AI has significantly closed the gap for the type of knowledge that is on the internet. If you want a well-structured explanation of how Series A valuations work in general, AI does that well. If you want to understand standard contract language in a commercial lease, AI does that well. The easily-available tier of knowledge has gotten dramatically better and more accessible.
This has a counterintuitive effect: it makes the hard-to-access tier relatively more valuable. When basic information is free and instant, the premium moves entirely to specific, contextual, judgment-based knowledge. The gap between "what you can find online" and "what you actually need to make a good decision" hasn't narrowed — it's become more visible because the first tier is now fully saturated.
People are already intuiting this. The popularity of expert newsletters, paid communities, and high-end advisory services has grown alongside AI, not shrunk. People who can afford it are paying more for direct access to genuine expertise, not less. The market is trying to solve the problem — the mechanism just isn't efficient yet.
The ideal solution would be something that makes the tacit knowledge tier as accessible as the public internet tier — not free (that would just recreate the same problem), but accessible on demand and at reasonable cost. If you could ask an expert directly about your specific situation — not a generic answer but a response grounded in real expertise applied to your context — the quality of decisions people make would improve substantially.
That's the structural opportunity. The knowledge that matters most has always been locked away from the people who need it, not because experts are hoarding it, but because there was no efficient mechanism to deliver it. You can learn how expert vaults work — but the broader point is that the problem has always been structural, not a matter of willingness to share.
The best advice has never been on Google. Making it accessible — really accessible, not just "technically available for $500/hour" accessible — is one of the more important problems worth solving right now.