What this covers

Gabe, who runs Rogo, an AI company serving investment banks and investment firms, walks through what it takes to build for finance and where the industry is heading. Rogo's early users are dealmakers in private markets, where coordination is still largely human and so offers more workflow to automate than public equities. Bloomberg's sequence is the template: a little data to get in the door, analytics and workflows on top, then a messaging platform where opaque asset classes can trade. Gabe projects that forward to a channel where agents negotiate, correspond with advisors, and run auctions across firms. He ties this to standardization, which he says drives liquidity and which AI could finally bring to unstructured private-market data.

The range is wide. He argues for Auditability Over Accuracy, describing Rogo's lineage tracking in case a court someday makes AI research discoverable. He forecasts capital raising moving online the way mortgages did, while small-business sales in generational turnover keep their handshake trust. He wants to price on outcomes rather than usage. He also says early applied-AI companies were hurt by building ahead of model capability, and that Rogo rebuilds its harness continuously, as Affirm does with its ledger.

The more contested turns are about people and ambition. He casts intermediaries like JP Morgan financing railroads as accelerators of innovation, and argues that human agency comes from many interacting brain "microservices" that AI must replicate. Firms should chart their deal lifecycle for irreplaceable data and relationships, since being smarter will not last. For ventures, he urges a Tail-Widening Risk Posture and backing dynamic founders.

Sharpest takeaway

Gabe argues that the durable value in vertical AI for finance lies in the non-model infrastructure (data access, compliance-grade auditability, workflow integration, systems of record, and transaction venues) that frontier labs will not build, and that incumbents face an innovator's dilemma they have not yet confronted.

  • Frontier labs will not build deep, messy vertical plumbing such as data rooms, systems of record, and compliance-grade audit trails, so those layers are the defensible position.
  • Auditability matters more than accuracy because users and regulators must be able to trace how an AI output was produced.
  • Financial institutions have not yet faced a shock forcing them to rethink their business models, so the AI-native disruption threat is still ahead.

The argument · threads13 threads · 73 claims
0.85

Vertical AI opportunity is largest in unbuilt private markets lacking standardized financial plumbing and venues.

10 pointscentrality 5/5
  • Rogo's early users are dealmakers on the buy side and sell side of transactions, because private markets are still largely human-coordinated and so offer more workflow to automate than public equities, where much of the infrastructure is already automated and exchanges already exist.

    “Part of the reason private markets are so attractive is because it's all done by humans... Whereas public equities, a lot of it has been automated.”

  • The opportunity to build an AI business in a vertical is largest where lots of plumbing is not yet built, and the AI layer is then applied on top of that infrastructure.

    “Exactly. Part of the reason private markets are so attractive is because it's all done by humans.”

  • To become the exchange for high finance and capital markets, a company must own not just the intelligence but the transaction venue, the communication venue, the workflows, and the data inputs, because frontier labs are unlikely to build data room businesses.

    “if you actually want to be the exchange for all of high finance and all of capital markets, you don't just need to own the intelligence, you need to own the transaction venue”

  • Just as 40 to 50 percent of mortgages are now delivered online by platforms like Rocket Mortgage after being human-intermediated 15 to 20 years ago, capital raising and asset pricing will move online, with transactions taking minutes rather than months and markets becoming more transparent, liquid, and efficient.

    “Now, 40 to 50% of mortgages are just delivered online by platforms like Rocket Mortgage.”

  • Startups should build things perpendicular to what the frontier labs want to build, because labs will never build the deep, messy, workflow-specific systems beneath a chatbot interface that verticals like finance require.

    “You want to build things that are perpendicular to what they want to build.”

  • The core infrastructure missing for automated private capital markets is a communication channel for agents to transact with each other across firms, which differs from Bloomberg Messenger because it must allow agents to negotiate, correspond with advisors, and run auctions on behalf of firms.

    “I need to build the communication channel for the agents to transact across these businesses, across these investment firms.”

  • Asset classes have grown in transaction volume and liquidity largely because of standardization, which makes assets easier to track and trade, and private markets have resisted standardization because of unstructured data, a gap AI should close.

    “often it has to do with standardization — because it gets easier then to track those assets and trade them. Private markets have been immune to standardization because there's so much unstructured data.”

  • Hundreds of thousands of American businesses are too small for investment bank coverage, meaning banks like JP Morgan lack the time to serve them, and AI-enabled coverage could bring capital markets access to these firms.

    “the 300,000 American businesses that couldn't even tell you what Goldman Sachs does, and JP Morgan doesn't have the time to go out and work with them because the business is too small.”

  • Public equities are the most brute-force AI application because data is available, but much of public markets is already automated and its transaction infrastructure already exists, so the opportunity is greater in private markets where work is still done by humans.

    “Part of the reason private markets are so attractive is because it's all done by humans.”

  • The two biggest uncertainties are how quickly private markets standardize, since standardization historically drives liquidity and AI could solve unstructured data, and how much alpha remains in human relationships, especially for small and medium business succession, which may not be fully automated.

    “Private markets have been immune to standardization because there's so much unstructured data. AI should fix that.”

0.68

Trustworthy finance AI requires auditable, traceable outputs and plumbing built into systems of record.

9 pointscentrality 4/5
  • Rogo builds its system so that AI inputs and outputs in investment and banking work are fully traceable, anticipating that a court may someday make AI research discoverable, so firms can show they kept information unintermingled and can reconstruct the full lineage of AI-generated judgments.

    “if someday some Delaware court judge makes AI inputs and research discoverable, you have actually done it all the right way”

  • As AI moves from copilot to autopilot and takes actions like making investment decisions or sending emails, the system must be explainable enough to debug why a decision was made, because regulators need to ensure no foul play and agent failures must be investigated like a bad human investor's decisions.

    “if you can't explain why you made a decision and the data went in, that's not going to fly.”

  • For financial AI, auditability matters more than accuracy: a user who can see the assumptions and data sources behind an answer can still act on an imperfect answer, whereas an accurate but opaque answer is unusable because the user cannot trust it.

    “I think it's actually more important to be auditable than it is to be accurate.”

  • Adoption depends on small workflow details: letting a managing director email a deck markup to an AI analyst and get it back in 20 minutes instead of two days, while alerting the junior analyst and preserving auditability, works for a dealmaker who has not used a computer in a decade but knows how to mark up on an iPad.

    “We make it so easy for a managing director at a bank to email a markup of a deck... and then returning that markup in 20 minutes as opposed to two days.”

  • Rogo's surface area is roughly half interaction with the underlying systems of record, such as CRM, portfolio monitoring, and LP reporting, because deal work requires updating these systems as the deal progresses, not just helping individual humans.

    “half of our surface area is actually the underneath of the iceberg of interacting with these different systems of record”

  • Models are now smarter than anyone Gabe knows, so the cutting-edge work is building the plumbing to connect them to a firm's context, thesis, and working style and to collect the data that feeds them.

    “the models are smarter than anyone I know, anyone I spend time with, and it's plumbing.”

  • A large private equity firm could eventually let an agent negotiate a deal on its behalf, correspond with QoE providers and legal advisors, and run a bid auction among sponsors, which Gabe sees as a large amount of software still to be built.

    “think about what is the system that would allow a large private equity firm to feel comfortable having an agent negotiate a deal on its behalf”

  • The third criterion is a business willing to reinvent and even cannibalize its core product quickly, because delivery methods like a terminal or a rigid UX cannot adapt every six months to step changes in models.

    “anyone whose delivery method of their product is not something that they can fully cannibalize quickly — a terminal, or a very specific UX or interface is not going to be agile enough”

  • When a firm ingests material non-public information, it needs auditability and flagging that feeds into internal systems so that if a regulator audits it, the firm can show exactly what it did, requiring plumbing that is pixel perfect.

    “if you are ingesting MNPI because you are working on transactions... do you have all that plumbing in place that is pixel perfect?”

0.62

Durable value lies in proprietary inputs and domain judgment as AI absorbs analyst work.

7 pointscentrality 4/5
  • Two years is easier to reason about than ten or twenty years because the best investors will spend the next two to five years figuring out how to reinvent their own firms and integrate AI into the investment lifecycle, which Gabe calls the biggest challenge for every great investor.

    “the world's best investors today are going to spend the next two to five years figuring out how to integrate AI into what they do”

  • The jury is still out on whether AI can eventually replace the judgment that the world's best investors use to decide what matters across topics, but if AI begins producing insights like Move 37 in public equities, it will substantially change which skills matter.

    “I think the jury's still out on whether or not that is something that AI can eventually replace.”

  • The core skillset that will remain valuable as AI handles more analyst work is gathering data and inputs that no one else has, such as spending time in the field, speaking to experts, and building relationship graphs, because one can then feed proprietary inputs into the model even if one cannot calculate the best ideas oneself.

    “the core skillset is folks who can go out and gather data and inputs into their model that no one else will have.”

  • Most firms see huge individual productivity gains from AI, but converting that into measurable firm productivity such as more deals or new market segments requires a firm-level strategy decision about whether to cut costs or enter new markets.

    “The problem is, where's that flowing through? Are you winning more deals? Are you actually transacting more?”

  • Gabe judges Rogo's success by daily user feedback that the product saves hundreds of hours per month and makes users smarter and better at decisions, rather than by traditional product metrics.

    “I'm saving hundreds of hours a month. I am doing things I never could have done before.”

  • Applied AI is the most interesting place for talent because it dictates how AI intersects with humans, and finance is upstream of all capital allocation, so making finance more efficient can supercharge the world.

    “Finance is a domain that is the catalyst for all human progress and innovation, and capital allocation is upstream of the financing of every company”

  • Domain expertise and product or GM-style judgment are increasing in value as execution ease rises, although some product surface still needs raw gritty engineering, and former founders with engineering chops who ate glass are most valuable on the team.

    “there's definitely problems where domain expertise is increasing in value and product intuition, and having more of a GM-like mindset versus an engineer-like mindset is super, super important.”

0.59

Deep vertical systems of record create AI wedges that generic chatbots cannot breach.

6 pointscentrality 4/5
  • An attractive vertical for an AI startup must have enough complexity and depth in data types, systems of record, and deployment models that solving them creates a wedge, because otherwise anyone can sell a generic chatbot into the industry immediately.

    “the industry actually does have to have enough complexity and depth in the types of data, the types of systems of record that people use”

  • Bloomberg's strategy was to offer a little data to get in the door, build analytics and workflows on top, and then provide the exchange and communication platform, Bloomberg Messenger, where previously opaque asset classes could be transacted.

    “offer a little bit of data to get in the door, build all the analytics and workflows on top that someone would need. And then provide the exchange and the communication platform”

  • Finance is a collection of different niches with different data sets, definitions of good, and regulatory requirements, and going deep across them can produce a company generating around $5 billion in revenue by creating systems of record for those niches.

    “all of finance is a collection of different niches with different data sets, different definitions of good, different regulatory requirements.”

  • A regulatory or market force could impose additional standardization on private markets, which would sharply accelerate transparency and liquidity, though this is outside Rogo's control.

    “is there going to be a regulatory or a market force that forces some additional standardization that really accelerates the kind of transparency and liquidity you have in private markets?”

  • Small and medium businesses undergoing generational turnover will be slow to click a button to sell, so human trust and handshake relationships will remain valuable in those deals, while sponsor-owned businesses, secondaries, and private credit can be fully automated.

    “it's very hard to imagine that if you're a small business owner and you've been building a business for 20 years and you want to make sure that if you hand that business off to someone else”

  • Intermediaries that connect capital allocators with entrepreneurs, as J. Pierpont Morgan did in financing American railroads, have had a profound effect on the world, and speeding up how quickly founders and businesses can tap capital markets can accelerate all innovation.

    “If you're able to speed up the rate at which entrepreneurs and company founders and individuals can tap into capital markets, you can accelerate all innovation.”

0.52

Firms must enforce AI adoption and win scarce, expensive domain talent for top AI companies.

8 pointscentrality 3/5
  • Gabe enforces AI adoption by sending monthly reports ranking every employee's AI tool usage, posting the top five and bottom five per division, and giving the bottom user in each division a printout with a dunce cap posted around the office, which creates a strong incentive function.

    “the bottom user in each division gets a printout with a dunce cap and we post it around the office.”

  • Rogo assembled a team of more than 100 people who previously worked in investment banks or investment firms, because domain expertise can be built by hiring from the network, and that team can continuously harness new models for finance.

    “So I have over 100 people that have spent time within investment banks or within investment firms across the world's best institutions.”

  • Rogo's most excellent team members are often former founders who started businesses, persevered, and had product intuition but whose businesses petered out, because those people combine product sense with engineering chops and the resilience to navigate a fast-changing product.

    “a lot of them are former founders. They're folks who started businesses, persevered, ate glass, had a lot of product intuition”

  • Rogo records every internal conversation and feeds it into an internal knowledge system called Shrek, so that employees can retrieve what a peer discussed weeks earlier, making company-wide knowledge reuse the core solution to enablement.

    “every internal conversation that happens at Rogo is recorded.”

  • The core constraint on growing an AI enterprise sales business quickly is how fast a human can be made productive, including salespeople, marketers, SDRs, and post-sale staff, so enablement and constant retraining is the fundamental problem.

    “The core problem we need to solve is how fast can you make a human being productive?”

  • Winning vertical AI companies will become black holes for talent and capital because the execution bar is higher than ever and the right team is extremely expensive, so companies that fail to attract resources quickly risk becoming roadkill for labs or better-capitalized competitors.

    “the businesses that do it are going to become black holes for talent and capital and brand”

  • Rogo's category-leader position in finance AI is intended to support a $500 billion outcome rather than merely a $100 billion business, because transforming capital markets requires depth of problems and a team capable of solving them.

    “we're the category-leading player who has the best shot on goal to not just be the $100 billion business to do it, but the $500 billion business”

  • Gabe's pitch to talent is that applied AI in finance is the most interesting domain because capital allocation is upstream of every company and idea, and Rogo's team is low-ego, hungry, curious, and smart, which is the team needed to become a $500 billion business.

    “Finance is a domain that is the catalyst for all human progress and innovation, and capital allocation is upstream of the financing of every company, every idea, every economy.”

0.43

Vertical AI products succeed only when built to match model capability thresholds.

5 pointscentrality 3/5
  • Many applied AI companies suffered a first-mover disadvantage because they built products for where they expected the world to go before the models were capable enough, so early users found the products terrible; companies that bet correctly on the end state are now seeing those products become magical as models arrive.

    “if you were right about the end state and where the models were going and you were building towards that, when they get there, it's magical.”

  • Rogo's product eras track model capability thresholds: o1 Pro was the first model reliable enough to function as a good search tool for financial metrics, and Opus 4.5 at the end of 2025 made models capable of basically anything a junior investment professional or banker did, given the right instructions and context.

    “It was o1 Pro and then probably Opus 4.5. o1 Pro was the first time you got enough reliability where it was a good search tool at the very least.”

  • Early Rogo was a Rube Goldberg contraption requiring around 60 separate model calls to resolve a single question, and as models get smarter the right design is fewer prescriptive steps and simpler, higher-quality tools for data gathering, self-auditing, and presenting outputs.

    “The early days of Rogo was kind of this Rube Goldberg contraption where you had 60 different model calls.”

  • Rogo benchmarks models with custom evals and data sets to determine which is most performant, cheapest in tokens, and lowest latency for each task, then routes tasks to the appropriate model type.

    “we spend a lot of time benchmarking these models and creating different evals and data sets so we can always decide what is the most performant, what is the cheapest from a token perspective, what is the lowest latency, and route tasks to the appropriate type of models.”

  • Rogo has acquired six fledgling financial AI startups, reflecting many smart domain-expert founders who tried to tackle finance AI and reached the market before models were capable enough.

    “And we've acquired six different fledgling financial AI startups.”

0.41

Founders must absorb setbacks, modulate emotion, and raise aggressive outcome tails to blitz markets.

9 pointscentrality 2/5
  • In turbulent, ambiguous markets such as applied AI, investors should underwrite whether founders can be extremely dynamic, because founders who iterate through repeated rejections and setbacks, and who are reconsidered at each round, are the ones who build momentum.

    “when you're investing in a market like today's, especially in applied AI, where it's so turbulent, so ambiguous, you need to underwrite the founders being able to be extremely dynamic.”

  • Startup building is a roller coaster in which the founder must absorb extreme highs and lows, but over time the lows get less cataclysmic and the highs compound, so the founder should modulate emotion and channel it without letting it distract the team.

    “if I look back at the journey, the lows get lower, the highs get much higher.”

  • Business building is a game of compounding momentum, where the founder's job is doing many small things to move faster downhill, and anxiety about losing that momentum is what drives thorough preparation for roadblocks.

    “the reality of this sort of business building is that it's a game of compounding momentum.”

  • Founders should not pretend to be something they are not, but should believe in a roughly 5 percent chance of becoming a $100 billion business and show a clear roadmap explaining how that outcome is possible, rather than projecting certainty.

    “You need to believe that that 5% likelihood that you can be in a $100 billion business is likely, and you don't need to pretend it's 100% likely”

  • A venture-backed company should deliberately raise the tails of its outcome distribution, accepting that its probability of failure rises if its odds of becoming a $100 billion company rise by a larger proportion, because a plan that is not aggressive enough is a soft plan.

    “my goal as a venture-backed business is to increase the tails of the distribution. It's fine if it gets 30% more likelihood that I fail, if the odds that I become a $100 billion company also increase by 20%.”

  • There are few credible financial services AI competitors because distribution in finance is hard to crack and depends on trust and value delivery with institutions more than engineering, while the product and engineering burden is also very high.

    “Distribution is really hard to crack. I think it's far more a people problem on building trust and delivering value and working with these institutions than just an engineering and product problem.”

  • If every buyer in finance will purchase an AI solution in the next 18 months regardless of whether a given company is there, then the only thing that matters is blitzing the market as quickly as possible to be among the vendors chosen.

    “then the only thing that matters is that you can blitz the market as quickly as possible to make sure that you are there.”

  • Because Rogo sells enterprise software that requires humans for every deal, it must build a go-to-market machine five times faster than most enterprise sales organizations have ever had to, using AI as a tailwind to build much better products.

    “we have to build a go-to-market machine five times faster than most enterprise sales organizations ever have to build.”

  • Pat Grady argued with a two-premise logic that if everyone in finance will buy an AI solution within 18 months, the only thing that matters is blitzing the market to be there, so Rogo's plan was not aggressive enough and needed to be more aggressive.

    “then the only thing that matters is that you can blitz the market as quickly as possible to make sure that you are there.”

0.40

Durable AI advantage comes from harnesses, memory, and agency components rather than raw models.

4 pointscentrality 3/5
  • Human agency comes not just from raw recall and IQ but from many interacting microservices in the brain for storing, retrieving, and triggering knowledge, with emotions acting as triggers, and AI systems need analogous components built out.

    “the reasons that humans are high agency and can do a lot is not just because we have high raw recall and IQ and knowledge. But there's all these different microservices in your brain.”

  • The capability that current models lack and that would most change the product is compaction: remembering the right things across many conversations and compressing memory into a token budget while keeping coherent context about a user, a problem that compounds when agents serve many colleagues at once.

    “Compaction. If you have 100 conversations with a single agent, how does it make sure that it's actually remembering the right things”

  • Rogo rebuilds its harness and core agentic system continuously, because the models change so quickly that the company is never near a local minimum, and this is analogous to Affirm rebuilding its ledger every year to prevent ossification.

    “We are constantly looking at it, realizing we're not even at a local minimum.”

  • The way a model is harnessed matters more than the raw model: Claude Code and Claude Cowork outperformed ChatGPT-era offerings even though the underlying models were fairly similar, because Anthropic's harness allowed models to exercise more long-running capability, driving usage growth.

    “The models were actually fairly similar, but the harness and the way that it was presented from Claude was far better.”

0.29

Incumbent firms must reinvent everything around AI-native systems they own, not just their people.

5 pointscentrality 2/5
  • Investment firms and banks have not faced an innovator's dilemma because the last 10 to 20 years were good for them and no shock has forced every firm to rethink its business, but that moment is coming as AI-native disruptors attack their business models.

    “there hasn't been a moment and a shock to the market where every investment firm, every bank is saying oh wow, I need to completely rethink what I'm doing.”

  • Firm leaders should chart every step of the deal lifecycle and identify where they hold irreplaceable data, relationships, or context; if their edge is only being smarter or better read, AI will obviate it.

    “Or do I just think I'm smarter and better read on the subject area and the market, in which case AI is going to obviate that.”

  • Being AI-native means constantly reinventing everything the business does, being so AI-pilled that nothing is considered too far-fetched, and treating no part of the business as sacred or immune to revolution by an alien technology.

    “being so AI-pilled that you don't worry about what's possible or what might seem completely far-fetched”

  • Leaders should ask how to take the expertise living in their best people's minds and put it into a system they own and operate autonomously, because in ten years top investment firms may hold most of their enterprise value in software, data, and systems rather than people.

    “What would you start doing? You would start trying to figure out how do you take all of what lives in the latent minds of your best people and putting it into a system that you own and operate autonomously.”

  • Firm leaders who were early to AI, such as John Momtazee at Moelis, wanted digital clones of their best bankers that could join calls and speak on their behalf, so that junior staff could leverage senior relationships and context immediately.

    “One example is there's a co-founder of firm Moelis, John Momtazee, who was just prescient about where it was going, and he wanted digital clones of all their best bankers.”

0.23

Pricing and distribution follow buyer habits and large banks serving as the channel.

3 pointscentrality 2/5
  • Rogo prices per seat because its buyers are accustomed to seat pricing and compare it to Bloomberg, FactSet, Capital IQ, and PitchBook, but seat-based enterprise sales require a human-intensive go-to-market motion that cannot scale usage 100-fold per deal.

    “We price per seat right now because our buyers are used to pricing per seat.”

  • JP Morgan's announced push into M&A for SMBs illustrates that AI can let a single banker act as a full deal team, making previously uneconomic small-deal segments serviceable because the fees were too small to justify staffing multiple people.

    “JP Morgan just announced that they're going to try and do a lot more M&A work for SMBs”

  • Large investment banks are the most strategic early customers because they are the distribution channel for the rest of finance, since many future great investors begin as analysts, and because a single bank like Bank of America has far more seats than the ten best public equity portfolio managers.

    “the investment banks are kind of the distribution channel for the rest of finance.”

0.23

Usage-based AI businesses can scale with fewer people, shifting bottlenecks to customers' decision-making.

3 pointscentrality 2/5
  • The most cutting-edge use is Moltbook-style swarms where a PM could have 10,000 agents debate ideas for 24 hours and return one idea, which is viable because investors will pay about $50,000 for one good idea whereas few domains justify that many tokens for a single insight.

    “Imagine if every PM at a hedge fund had 10,000 agents that were fraternizing, talking about ideas... and then at the end of 24 hours of debate, gave you one idea.”

  • Token-consumption AI businesses such as Cursor, Factory, and Claude Code can sell into parts of enterprise that already buy usage-based tools, so they can go much further commercially with fewer people than per-seat enterprise software, which requires a human-heavy sales organization.

    “And then there are the token brokers, the token consumption businesses, where you're selling in the parts of enterprise where they're used to buying usage-based tools.”

  • As AI can provision unlimited capability, the bottleneck shifts to the customer's ability to decide what to do with that productivity, such as how to structure and allocate work across many capable agents.

    “If you had 100 great investors start here tomorrow working for you, how would you channel that productivity?”

0.20

Vendors should skip usage pricing and charge directly for outcomes delivered.

2 pointscentrality 2/5
  • Every business must go through two pricing revolutions, first to usage-based pricing and then to outcome-based pricing, and Gabe wants to skip straight to outcome pricing, charging for good investment ideas, perfect LP reports, or CIMs, because outcomes are easier to value than tokens.

    “I think that every business needs to go through two different pricing revolutions. You need to move to some sort of usage-based, and then you need to move to some sort of outcome-based.”

  • Usage-based pricing creates a misalignment in which a vendor's margin deteriorates as a customer uses the product more, so the vendor's success with a customer raises its cost to serve, which Gabe addresses by focusing on being a good partner and steward of the customer's AI strategy.

    “if you do a great job and they use the thing way more, which costs you a lot of money, they become a worse customer?”

0.06

Fundraising hinges on repeated investor exposure, which lets investors gather data on founders.

2 pointscentrality 1/5
  • Founders who lead Series C or D rounds are often investors who met the company at seed, A, B, and C and passed each time, because repeated interaction lets them gather more data on the founder and business, so reps and relationships are essential to fundraising.

    “it's about reps and relationships. The folks who come out of nowhere and lead the Series C or the Series D are the folks that I met at the seed and then the A and then the B and the C, and passed every time”

  • Rogo's Series A process involved meeting about 40 investors, including Sequoia, Kleiner, and Benchmark, and all 40 passed, with some firms like Thrive spending extensive time before rejecting; investors ultimately passed because they lacked intuition for finance as a venture market, the product was poor, and they doubted the founder.

    “David introduced us to 40 investors, and I met with everybody. I met with Sequoia, Kleiner, Benchmark, everybody, and 40 people passed.”