What LinkedIn Will Never Do: Open a Searchable Web of Professional Trust
Cosign, a16z's bet, argues that the next chapter in hiring isn't mass-applying for jobs—it's having someone you trust, or eventually your own AI agent, run the first round of screening for you.
The video won't play here. Listen to the audio instead:
The argument · tap a timestamp to hear it
Talent data gets trapped inside corporate silos, invisible across companies
Sophia observes that knowledge of who's a growth expert or top engineer usually stays confined to each company or small circle, never crossing organizational boundaries. Dhruv responds that Cosign Awards aims to break down these silos—transforming what companies normally keep quiet through internal gossip (like ‘which company has the highest talent density’ or ‘which product had the biggest impact this year’) into searchable public information, not just awards for famous founders and CEOs.
— SophiaNiche award categories surface far more talented people than expected
David cites specific numbers: the public intellectual category has 65 nominations, podcaster of the year has 76, and Rising Stars—his personal favorite—already has 135 nominations, currently the largest category. None of these align with traditional job titles, revealing a massive population of people with strong internal company reputations but no public visibility, who need a window to be seen.
— David BoothThe real product isn't the network—it's your surface area for serendipity
David quotes Jobs: you can't plan to meet the person who changes your life, but you can place yourself on the path where luck travels and grab the opportunity when it appears. He applies this to Cosign—building profiles, populating the graph, browsing relationships—is fundamentally increasing the odds that future partners or the right person will spot you. He calls this ‘coincidence of intent’: when two people with complementary skills wanting to build the same thing can actually be matched by the system.
— David BoothThree layers build the complete graph: profiles, lists, then co-signs
David breaks Cosign into three layers. The bottom is a Wikipedia-like profile system for 40,000 to 50,000 people, clickable through companies and investors. The middle layer is ‘lists’—about 400 people have published lists so far, like Packy McCormick's 36 companies he thinks are worth joining. The top layer is ‘co-sign’ itself, where someone formally tells the platform ‘I'd work with this person anywhere.’ All three together form the complete graph.
— David BoothLinkedIn's search limits are a choice, not a technical constraint
Dhruv believes LinkedIn will never open an MCP that enables full-text search across its entire relationship graph—even with hundreds of thousands of accounts, ordinary users still can't find the people they need. He recounts his own experience: hunting for startups worth joining in New York, he asked Gemini, then ChatGPT, and both gave poor answers. His conclusion: factual ‘who's doing what’ questions fail because the underlying data simply doesn't exist. Cosign's graph, fed by real professional relationships, fills that gap.
— Dhruv GuptaThe next frontier: agents that negotiate on your behalf, not just write emails
Dhruv notes the personal-agent wave lasted only a month or two; Muse hit 10 million users, but even the most tailored agent today can only draft an email in your inbox, send it to theirs, and wait for their agent to pick it up and reply—essentially the old ‘send message, wait for response’ loop. He points to early examples like Instagon where agent-to-agent protocols are already emerging (such as car-rental customer-service integrations), imagining the future: when job hunting, your agent could negotiate a round with the company's agent, assess whether your intentions align, and only then decide whether to surface you to the actual hiring manager.
— Dhruv GuptaEveryone's inbox treats all signals equally—and that's the core problem to solve
Dhruv points out that in today's social networks, Sam Altman's inbox and a regular person's inbox—even his 14-year-old sister's—work exactly the same way, with no signal-strength differentiation. This is exactly what Cosign wants to fix: ensuring you only spend time with people you should actually meet, rather than sifting through mountains of connection requests yourself.
— Dhruv GuptaIn their own words · checked verbatim
The biggest opportunities in your career often come down to one person willing to vouch for you.
you can't meet, you can't plan to meet the person who changes your life, but you can certainly put yourself in the path of luck and you can know that, you know, when that person appears, you can grab it with both hands and run with it.
David Booth8:15
it's kind of, it's an extension of the idea that you are the average of the five people you spend the time with.
David Booth12:23
I think that one thing that LinkedIn will never do is introduce an MCP that allows you to have vast search access over the graph.
Dhruv Gupta17:39
But it still doesn't have a way to actually connect with the people that is finding you. The best it can do is draft an email in your inbox and send it to someone else's inbox where maybe an agent picks it up and sends back it.
Dhruv Gupta19:41
And in fact, like our agents should connect with, my agent should be able to connect with the company's agent and do a first pass of whether it's worth elevating me to the hiring manager.
Dhruv Gupta20:42
I mean, on networks right now, the inbox works the same way for Sam Altman, as it does for me, as for my 14-year-old little sister.
Dhruv Gupta21:42
Figures
| Rising Stars nominations | 135 | 5:15 |
| Podcaster of the Year nominations | 76 | 5:15 |
| Public intellectual category nominations | 65 | 5:15 |
| Cosign lists published | approximately 400 | 14:27 |
| Companies on Packy McCormick's list | 36 | 14:27 |
| Muse users | 10 million | 19:41 |
Glossary
- MCP / Model Context Protocol
- An open protocol enabling AI agents to access external data and tools.
- coincidence of intent
- The moment when two people's goals and needs align perfectly enough to warrant an introduction.
- golden handcuffs
- Salary and benefits structured to keep talented people from leaving their current employer.
- agent-to-agent protocol
- Direct negotiation between two AI agents acting on behalf of their respective users.
How to listen
Founders and investors tracking AI agents in hiring and social networks, plus anyone curious how trust-based networks turn into products.
After 22:42, the host wraps up with pleasantries and a preview of the next episode—low on new information.