Dialog's Secret Ranking: How the Thiel-Linked Club Grades Members

Dialog, the private network cofounded by Peter Thiel and Auren Hoffman, grades event attendees on a hidden A/B/C scale and compiles dossiers containing home addresses, private phone numbers, dates of birth, photos, emergency contacts, food allergies and political labels.

Of 192 dossiers in the leak, 130 are tagged as members; the rest are prospects. The records show the group pairs an A/B/C grade with a "value-add" score of 1 to 4, a separate moderation tier, and notes staff use to decide seating, fees, introductions and whether someone should be disinvited.

How does Dialog grade members?

Dialog uses a three-tier letter grade plus a 1-to-4 "value-add" score, and it reviews and revises grades after every retreat. Of the 192 dossiers examined, 141 were marked B, one in seven received a C, and A appears reserved for older, established members whom graders consider less notable; staff perform a "post-retreat code review" to revisit scores.

Grades are justified by wealth and fame in staff notes: one investor is summarized by the "$30 billion in assets under management" he oversees, while others are downgraded for "Small AUM." Fame is measured against whether the "average person" would recognize someone, comparing prominence to a "Fortune 500 company or a top celebrity" in some dossiers. The group pairs grades with a moderation tier that identifies who is trusted to run workshops or hold "Soapbox" sessions.

What data and tools feed those scores?

The dossier leak includes nearly 200 people and spans contact details, personal photos, allergies and volunteered political leanings, and Dialog supplements those disclosures with internal assessments. The records show an AI tool was used to assemble dossiers on at least 26 people, but staff overruled the tool in some cases, as when graders promoted Tyler Cowen after the AI described him as not the leader of a household-name organization.

The database also tracks matchmaking signals: roughly 10 percent of respondents opted into a singles pool, more than three-quarters already have algorithm-suggested matches, and staff manually refine introductions. It also maintains "do-not-pair" flags for spouses, professional associates, and other barred combinations, including members flagged against Dialog staff and organizers.

Dialog's internal notes record political labels too. In the data for August's event, 165 people disclosed their politics and more than half identified with the left; despite that, those labeled on the right were more than twice as likely to carry a C. The files show staff sometimes overrode members' self-descriptions: eleven members were assigned labels despite disclosing nothing, and 15 self-descriptions were changed.

Why it matters

The system turns private social access into a quantifiable product: grades affect seating, introductions, and pricing. Bottom-grade attendees are placed on the full-price tier roughly 70 percent of the time, compared with about a quarter of those considered VIPs, and Dialog's events can cost attendees into the tens of thousands of dollars. That links personal data, perceived public prominence and financial outcomes in a closed network that convenes politicians, investors, military leaders, academics and journalists.

The leak also reveals demographic and cultural imbalances: women account for roughly a third of those graded but hold only 18 percent of top marks, and staff judgments about who is "widely recognized" or a "VIP" influence who gets access to high-value connections. The presence of an AI tool alongside manual overrides shows human moderators still steer the system's outcomes.

What to watch

Dialog's annual retreat scheduled in August outside Dublin, a two-day program on artificial intelligence, geopolitics and modern warfare, is the immediate test of how the grades and matchmaking play out in practice. Watch whether post-retreat code reviews change members' grades or pricing assignments, and whether any attendees publicly challenge the use of political labels or the club's matchmaking lists.

"We have no ideological agenda," reads a document shared by a past participant, a line Dialog uses to defend its process even as the leaked dossiers show staff assigning political labels and prioritizing fame, wealth and perceived recognizability in deciding who sits with whom, who pays what, and who stays.