a16z ecosystem partner and Cosign co-founder David Booth announced a product revamp: a new mutual interest network page, separating "cosign endorsements" and "list" into two sets of original terms, changing from an invitation system to a nomination system, and launching a new onboarding process. Priority will be given to those who have filled out the questionnaire.

Cosign was initiated by Booth and a16z general partner Erik Torenberg, positioning itself as a curated professional network for entrepreneurs, combining elements of a wiki-style directory, LinkedIn, and early AngelList. Its goals include directories for investors, companies, and operators, as well as an endorsement map of "who has shaped your career, who is willing to collaborate again, and who will be seen as future stars." Joining requires either an endorsement or a direct application. The two had conducted a similar experiment about seven years ago, which did not succeed.

The new rules categorize endorsements into three types: those who will become stars, those who have shaped your career, and those willing to collaborate anywhere. The list is akin to a playlist or collection board, which can include people, companies, and products, such as "designers willing to collaborate," with a weaker signal than formal endorsements. The team clearly does not want endorsements to be casually given to strangers. Nominators and nominees will be displayed on personal homepages.

The network page is used to see which mutual friends are endorsing which people and companies. The data side aggregates public financing, job changes, and public posts, claiming that even unregistered users can browse some profiles to alleviate cold start issues. The product also plans private intentions: willingness to invest, hire, or collaborate, with connections made after mutual matching. Officials state this is a community rather than a company looking to monetize.

Partners from a16z and former LinkedIn employees discussed the logic in a podcast: after AI turns applications and pitches into massive noise, the signal of real people willing to publicly endorse will become rarer. Job postings on LinkedIn can receive thousands of replies, and the cost of filtering increases after agents amplify them, making the endorsement map a next-level filter.

The buy side consists of entrepreneurs and investors looking for people, projects, and credible endorsements; the sell side includes incumbents willing to label private reputations into three categories. Funding is not currently reflected as transaction commissions but as relationship graphs and discovery rights concentrated in a16z ecosystem tools. Beneficiaries are those already in the map, visible to mutual friends; those under pressure are generic professional social networks relying on self-filled resumes for traffic and outsiders who cannot enter the nomination queue. The event-driven aspect comes from changing the rules in the first week after launch, using tightened original terms to prevent endorsement devaluation.

Source: Public Information