The Reputation Graph of Silicon Valley | Introducing Cosign

Key Themes

Reputation as an Underpriced, Undercapitalized Asset

The core thesis of the episode is that professional reputation — who believes in whom — is the most valuable and least structured data set in Silicon Valley. David Booth frames it as the real credentialing system of the industry, more powerful than formal education. "What's more impressive to you, someone who went to Stanford or someone who is followed by Patrick Collison, Elon Musk, Marc Andreessen, etc. I think most people would say the latter." [00:00:07] Erik Torenberg reinforces this from an investing standpoint: "It's like who has conviction and who, and who can give you kind of a real read on a person for so many decisions... And there's such information asymmetry right now." [00:06:23]

The Mentor Economy Is the Real Operating System of Silicon Valley

Cosign is explicitly built around capturing "who shaped your career," treating mentorship endorsements as a durable, transferable form of career capital. David Booth explains: "If Josh Elman is my mentor, that means... he is putting... a signal to other people that I'm worth other people spending time with too. Because if you respect Josh, you respect me by association." [00:04:36]

Social Media Reputation Is Powerful but Ephemeral — and That's the Gap

The founders argue that valuable endorsements already happen constantly on X/Twitter and in DMs, but they vanish. Olivia Moore notes: "There's so much that just happens and it's wonderful in the moment and then it gets lost. And I think this idea of building a profile... that lets other people... shine on your strengths, I think is really, really powerful." [00:19:53] David Booth adds the goal is to make reputation "durable" rather than a fleeting quote-tweet moment. [00:18:40]

LinkedIn's Network Is a Flat Graph — Connection Strength Is the Missing Layer

Multiple speakers agree LinkedIn succeeded at building the world's professional graph but never captured relationship depth. Olivia Moore, who worked at LinkedIn in its early days, explains the original design intent versus what it became: "It was even better if you imagine... this idea of this trusted graph... was the important condition. But over time it started to fade from that strength... everybody kind of got into a collecting [mode]." [00:30:21] Erik Torenberg gives a concrete example of the failure mode: "90% of the time when I say to someone, hey, I saw you're connected to this person on LinkedIn. Can you intro me? I will get the, hmm, I don't think I know them." [00:33:13]

AI Is Making Human Endorsement the Scarcest Resource

As AI-driven outbound and application volume explodes, the speakers argue verified human conviction becomes more valuable, not less. Erik Torenberg describes posting a hiring form and getting overwhelmed: "I get 2,000 results... it's going to be 10,000... a month from now as browser use agenting gets better... it's going to be 100,000... in an age of endless kind of outreach and interest with AI, the scarce thing is the human endorsement or the human conviction or the human knowledge." [00:34:55]

Category-Defining Products Don't Cannibalize Each Other — Everything Great Just Works

David Booth cites a framework from Benchmark's Eric (Vishria) via the Patrick O'Shaughnessy podcast to argue new products don't need to "kill" incumbents to succeed. "It's all going to work. Like, everything is going to be huge... when people say... does the world need another podcast?... another great podcast comes in and it doesn't kill the other. It just all works." [00:41:55]

The Product Is Deliberately, Structurally Positive-Only

A notable design decision: Cosign will not host critique or negative signal — only endorsement. David Booth states plainly: "The reason why it's all positive is because products have to win on some dimension and this product is going to win on people using it to show off... like, this represents the full manifestation of what I'm capable of." [00:21:50]

Recognition Should Extend Beyond Founders to the "Village" That Builds Companies

Olivia Moore and others push the idea that engineers, designers, and operators — not just founders — deserve durable, portable reputational credit. "It takes a village to build every one of these companies. It takes so many great designers and engineers and marketers and product managers..." [00:24:02] This echoes a16z's earlier "Rise Awards," explicitly designed, per Josh Elman, "around this thesis that... founders get a lot of publicity, a lot of credit, but it's actually the people who are the operators behind the scenes." [00:24:34]

Contrarian Perspectives

Sharing Your "Alpha" on People Actually Compounds Rather Than Dilutes Your Advantage

Conventional wisdom in venture says you protect knowledge of great undiscovered people as proprietary alpha. Josh Elman argues the opposite is true over a long time horizon: "Someone might rightfully say, well, why would I share this co-sign? This is my alpha... But if you take a very long-term view... the person you want to co-sign is the one that's raising a pre-seed round two years from now... they remember you... And of course, I'm going to call her first [for] my pre-seed round." [00:47:52]

College's Credentialing Function May Already Be Worth Less Than a Twitter Account

David Booth opens with a claim that would be heretical at most institutions (notably delivered by a founder of a company that just launched a school): "My Twitter account is more valuable to me than my college education... my Twitter has given me more value to my career." [00:01:52] He extends this into a broader claim that peer-endorsement ("followed by Patrick Collison, Elon Musk, Marc Andreessen") now outweighs institutional pedigree as a credibility signal. [00:02:30]

LinkedIn's High Usage Masks Deep Dissatisfaction — and That's Not a Contradiction, It's Structural

Rather than treating LinkedIn's dominance and its low NPS as a paradox, Olivia Moore argues LinkedIn's model is inherently double-audienced and that dissatisfaction and habitual use can coexist indefinitely because of network lock-in: "I actually think the NPS is really high for the people who aren't actively in the flow and in the searching... But for those mildly happily employed people, they aren't constantly... checking LinkedIn." [00:37:56] David Booth adds that despite 20 years of well-funded, well-designed attempts (including crypto-incentivized ones) to dislodge it, "it's just fascinating... they have like low NPS on it or something. And they know that a better version can exist. And yet... it's been impossible to disrupt." [00:37:36]

The Most Valuable Career Signal Isn't What You Say About Yourself — It's the Absence of a Feed for Critique

Most social/professional platforms (X, Glassdoor, even parts of LinkedIn) derive value partly from critique and friction. Cosign is explicitly designed to exclude that dimension entirely, betting that a purely celebratory graph is more valuable than a "balanced" one that includes negative signal — a direct rejection of the "authenticity requires criticism" norm of modern social platforms. Olivia Moore: "There's plenty of places for critique and we see all that all the time too. But I think that doesn't need to be recorded for posterity in the same way." [00:22:33]

Companies Identified

Cosign — New a16z-built product creating a comprehensive, durable directory of people and companies in the startup ecosystem, centered on professional endorsements ("who shaped your career," "would work with," "person to watch"). Mentioned as the subject of the episode and framed as filling gaps left by LinkedIn, X, and AngelList. "We're super excited to announce Cosign, a product that is around professional reputation for the startup community." — David Booth [00:01:24]

LinkedIn — The dominant professional network, described as an iconic, undisrupted 20-year product despite widespread user frustration. "One of the most iconic products of all time... it's been impossible to disrupt." — David Booth [00:29:13, 00:37:56]

AngelList — Cited as an inspiration and cautionary tale; once "owned the Investor Graph" for the startup community but drifted from its free public social network as it made other product bets. "They made a series of other bets that then sort of made them focus on other parts... or deprecate the free public social network that used to exist." — David Booth [00:42:51]

X (Twitter) — Praised as a free, high-signal discovery and credentialing platform, but criticized for not being built for professional matching use cases (following ≠ hiring intent) and for making valuable social exhaust ephemeral and hard to export. "I wish I could just download my followers... it's not built for this use case." — David Booth [00:28:13]

Crunchbase / PitchBook — Referenced as valuable but paid products that Cosign intends to differentiate from by being free and community-oriented. "Crunchbase, PitchBook cost money, et cetera. They're great products. They have to be businesses. But we have the luxury of being able to just really offer this for the startup community." — David Booth [00:09:42]

GitHub — Cited as an example of a superior niche "LinkedIn" for a specific vertical (engineers), supporting the idea that hyper-niche reputation networks can outperform generalist ones. "GitHub is arguably a better LinkedIn for engineers than LinkedIn itself." — Josh Elman [00:40:12]

Product Hunt — Mentioned via founder Ryan Hoover's early co-sign of Erik Torenberg, illustrating the mentor/endorsement dynamic the whole episode is built around. "Ryan Hoover... co-signed Eric and brought Eric on to become one of the leaders of building the community." — Olivia Moore [00:46:43]

Cognition — Referenced through Russell Kaplan, now president there, as a proof point of early talent-spotting paying off. "One of them was Russell Kaplan, who's now the president at Cognition. And he was like a 22-year-old just engineer at Tesla." — David Booth [00:05:35]

Databricks — Referenced via Ali Ghodsi as an example of a founder who rose after being co-signed by senior industry figures. [00:46:32]

Airbnb — Used as an example of how founders publicly credit their earliest believers years later, reinforcing the long-term value of early endorsement. "Brian Chesky, when he... goes public with Airbnb, who does he write in the blog post? He writes... YC, you know, Paul, his first investor." — David Booth [00:48:43]

People Identified

Josh Elman — a16z general partner, former LinkedIn employee, co-builder of Cosign; known for identifying trends and people before the broader market. "The most important skill in this town or in this industry or perhaps in the world is your ability to identify things before the market." [00:00:00]

David Booth — Builder/founder driving Cosign at a16z; originator of the episode's central thesis on peer-credentialing over institutional credentialing. Also referenced as having run a talent-scouting experiment seven years earlier that surfaced people like Russell Kaplan and Olivia Moore before they were well known.

Olivia Moore — a16z partner, former early LinkedIn employee who worked directly on LinkedIn Jobs and the "social resume" concept; brings historical LinkedIn product context to the conversation. "What is a social resume?... a resume annotated by who else can help you reach that person." [00:31:18]

Erik Torenberg — a16z operator, co-founder of a long-running Sunday startup jobs newsletter with his twin sister Justine Torenberg, and a person whose own career was "co-signed" by Ryan Hoover of Product Hunt. "My twin sister, Justine, also works here at Andreessen on our infrastructure team... we started a newsletter that we ran every Sunday for I think like six and a half or seven years." [00:07:34]

Justine Torenberg — Erik's twin sister, works on a16z's infrastructure team; co-ran the startup jobs newsletter with Erik for ~7 years, credited as part of the earliest version of this "curation" thesis.

David Perel — Friend of David Booth whose quote inspired the episode's opening thesis. "My Twitter account is more valuable to me than my college education." (as relayed by David Booth) [00:01:24]

Zach Frankel — Cited as someone who has built an outstanding personal reputation specifically around discovering talent early. "Zach Frankel has done an amazing job, right? So many people have done amazing, building a reputation for discovering people." — David Booth [00:03:58]

Russell Kaplan — President at Cognition; cited as an example of someone identified as a "person to watch" at age 22 while an engineer at Tesla, validating early talent-spotting. — David Booth [00:05:35]

Paki McCormick — Investor noted for strong signal in deep tech and "vertical integrators," credited with having identified 37 companies worth watching in that space. "Paki's got 37 companies that he's chosen. He says, these are the companies if you want to get into one into this space." — Josh Elman [00:09:10]

Nikita Bier — Referenced twice as a benchmark for consumer growth tactics (leveraging social proof) and jokingly as a comparison point for David Booth's own reputation-building work. "It's the Nikita Bier. Like, your friends just answered this question about your strategy, which worked so well three times." — Erik Torenberg [00:44:23]

Charles Hudson — Investor referenced for having publicly called out, over a decade ago, the opportunity to disrupt LinkedIn. "Charles Hudson, you know, great, great investor... 12 years ago wrote a blog post, like where are the LinkedIn competitors?" — David Booth [00:37:07]

Paul Graham — Cited for his famous early co-sign of Sam Altman, listing him among history's best entrepreneurs before Altman had a major exit. "Paul Graham co-signed Sam Altman famously... he said, like, the five best entrepreneurs... and Sam Altman. And at the time, Sam Altman hadn't had a huge exit yet." — David Booth [00:46:02]

Sam Altman — Referenced as the beneficiary of Paul Graham's early endorsement, which propelled his career including running YC. [00:46:02]

Ben Horowitz — Mentioned in passing as another example of someone whose career was launched via a senior co-sign. [00:46:32]

Ali Ghodsi — Databricks leader mentioned as another example of a co-sign beneficiary. [00:46:32]

Dwarkesh Patel ("Dorkesh" in transcript) — Highlighted as proof of Silicon Valley meritocracy: a 25-year-old with no prior job history who became one of the most important intellectuals in tech via podcasting, after early co-signs from prominent figures. "One of the most important intellectuals right now is Dorkesh. He's 25 years old. He's never had a job, as far as I'm aware. He just... came to San Francisco, started podcasting." — David Booth [00:47:23]

Patrick Collison — Referenced as one of the highest-value people to be followed/endorsed by, and possibly an early guest/co-signer of Dwarkesh's podcast. [00:02:30, 00:47:23]

Elon Musk — Referenced as a top-tier endorsement signal (being followed by him carries reputational weight). [00:02:30]

Marc Andreessen — Referenced both as a top-tier endorsement signal and as someone whose public blog post about a hire ("Mark Andreessen wrote a blog post about me joining the firm") is cited as an aspirational durable reputation artifact. [00:02:30, 00:20:50]

Michael Truel — Cited as an example of a young (mid-to-late twenties) founder who became one of the most important people in the industry, reinforcing the meritocracy theme. — David Booth [00:47:23]

Ashton Brenner (referenced as Aravind or possibly garbled name) — Cited as another young (24-25) major figure in the industry as further evidence for the meritocracy thesis. — David Booth [00:47:52]

Reed Hoffman — LinkedIn founder, quoted via Olivia Moore's recollection describing LinkedIn's three-stage strategy (growth, usage, revenue) and pushing the "social resume" concept internally. "Reed would push us a lot and he'd say, what is a social resume?" — Olivia Moore [00:31:43]

Eric (Benchmark) — Referenced via the Patrick O'Shaughnessy podcast for the "everything is going to work" category thesis. — David Booth [00:41:55]

Ryan Hoover — Product Hunt founder credited with co-signing Erik Torenberg early in his career. — Olivia Moore [00:46:43]

Operating Insights

Build Public and Private Endorsement Lists as Standing Infrastructure, Not One-Off Favors

Josh Elman describes maintaining literal running lists of people he'd privately recommend versus publicly endorse, so that when a founder asks for a referral, the matching work is already done rather than reconstructed from memory each time. "Every time a founder comes to me and they say, David, I need a new designer. I need a growth guy. I need an engineer. I can say, well, do I consult my private signals or do I send them my public list?" [00:16:51]

Design Events Around Structured Intent-Matching, Not Just Attendance

David Booth's anecdote about the SV Angel conference (attendees pre-swiped on a list of 150 people before the event, triggering automatic intros on mutual match) is a concrete, replicable event-design tactic for maximizing connection density without relying on organic mingling. "Basically, if they swiped right on me, then we would get an automatic intro... That event provided the context for us to... become friends and collaborators." [00:13:05]

Treat "Timing Mismatch" in Your Network as a Solvable Data Problem, Not an Inevitable Loss

Erik Torenberg identifies a specific, recurring operating failure — knowing two parties who should connect but at the wrong moment, and having no durable way to resurface that match later — as something worth systematically capturing rather than relying on memory. "I meet someone else three months later who's looking for that skill set. And I have to figure out like, remember this person, check if they still want the intro, make the intro." [00:14:35]

Give Credit Publicly and Early — It Compounds Asymmetrically in Your Favor Later

Beyond the general "networking is good" advice, the specific tactical claim is that public, early credit-giving to unproven people generates outsized future reciprocity precisely because it's rare and remembered disproportionately. "The benefits just accrue to you in the sense of people always remember like by order of magnitude more, the people who believed in them first." — David Booth [00:48:43]

Overlooked Insights

The "Mildly Happily Employed" Segment Is the Real Network-Effect Moat — and Almost No One Optimizes For It

Buried in the LinkedIn discussion is a genuinely underappreciated strategic insight about network effects: LinkedIn's durability doesn't come from active job-seekers but from the vast, passive population who never open the app but remain reachable — and that passive reachability, at scale, is what makes a 20-year moat. Olivia Moore names this specifically: "We used to talk about the... mildly happily employed because there's a lot of people who like what they're doing... But if the right thing popped... you might be like, hmm, should I actually think about that?... That's the kind of network effect that really sticks." [00:37:56] This is a generalizable insight for any founder building a marketplace or network product: the users who barely engage may be the actual source of defensibility, not the power users — a point neither host explicitly draws out or challenges after Olivia states it.

Verification Infrastructure (Bot Detection) Is What Actually Protects LinkedIn — Not Content or Network Density Alone

Erik Torenberg mentions, almost in passing, that LinkedIn's real defensibility against competitors isn't just network lock-in but its anti-scraping/bot-detection capability, which prevents rivals from bootstrapping via data theft — a distinct and separate moat from the social-graph argument the rest of the conversation focuses on. "They are famously good at bot detection. So you can't really scrape it in real time. There's whole businesses that are built off of scraping LinkedIn and selling that data." [00:39:31] This is a non-obvious operating insight for anyone building a network-effects business: the defensibility isn't only "who's on the network" but "how hard is it to extract the graph programmatically" — a technical/legal moat layered on top of the social one, mentioned once and never revisited.