a16z launches Cosign to trade reputation, but debases the currency
By Klaus Berger · Reporting from Frankfurt ·
By creating a positivity-only endorsement network with Cosign, a16z risks inflating social capital and stripping away the personal liability that gives professional signals real value.
Erik Torenberg poses a question on the latest a16z podcast that would make any traditional registrar shudder. "What's more impressive to you? Someone who went to Stanford or someone who is followed by Patrick Collison, Elon Musk, Mark Andre?" This thought experiment serves as the prospectus for Cosign, a new product designed to formalize the "peer-to-peer credential." Alongside Josh Elman, Olivia Moore, and David Booth, Torenberg argues that the most valuable signals in Silicon Valley are not institutional degrees. Instead, they are the "exhaust" of professional reputation—the mentors who shaped you and the colleagues who would work with you anywhere.
The securitization of the professional handshake
On the podcast, David Booth argues that the ability to "identify things before the market" is the industry’s most critical skill. Cosign attempts to turn this identification into a liquid asset. By allowing users to "cosign" individuals or companies, the platform aims to create a "social CV" that records conviction before it becomes consensus. Erik Torenberg frames this as a way to fix the "information problem" of talent discovery. This makes reputation more "durable" than the ephemeral chatter of social media. The ambition is to create a "reputation market" that is both efficient and liquid. This allows a young engineer’s talent to be recognized globally because an influential figure put their name behind them.
The structural failure of the one-way signal
However, I find the product's rulebook concerning. I have watched many "efficient" markets collapse under the weight of poor incentives. Torenberg explicitly states that the product "is just going to be positive," a place for people to "brag" and "celebrate." In any functional financial system, price discovery requires both buyers and sellers—or in this case, both endorsements and critiques. If a network only permits positive signals, it inevitably leads to reputation inflation. Josh Elman, drawing on his experience at LinkedIn, admits that the "strength of connection" in professional networks tends to fade over time. This happens as people "collect" connections. Without a mechanism to penalize a bad "cosign," the signal becomes as debased as a currency without a central bank.
Why liability must follow the endorsement
The strongest case for Cosign, as Olivia Moore suggests, relies on an era of AI-generated noise. In this environment, "human conviction" is the only remaining scarce resource. She argues that Silicon Valley’s "pay it forward" culture relies on these high-bar endorsements. But for a signal to have value, the giver must bear the risk. If I recommend a subprime borrower, my balance sheet suffers. If I "cosign" a mediocre founder, my professional standing should take a hit. Moore notes that a "reputation sin mechanism" exists naturally. However, Cosign strips away the possibility of negative feedback and focuses only on "celebration." This risks creating a moral hazard where everyone endorses everyone else to maintain social harmony. Stability in a professional network is not an accident; it is the result of rigorous, often unsentimental vetting. By removing the friction and the risk from the endorsement process, a16z is not making reputation more liquid. They are making it cheaper. A "cosign" that costs nothing to give will eventually be worth nothing to receive. We do not need more ways to "brag" for our friends. We need institutions that can actually enforce the value of a promise. Without the threat of social bankruptcy, the "social capital" being traded here is little more than paper.