Founders in Arms

Rahul Vohra is the founder of Superhuman, the email client that was acquired by Grammarly. Grammarly has since taken on the Superhuman name, and Rahul now leads Superhuman Mail. Before Superhuman, he built Rapportive, the first Gmail extension to scale to millions of users. LinkedIn acquired it 20 months after its first line of code. Rahul studied computer science at Cambridge, ran Cambridge University Entrepreneurs, and went through Y Combinator in Summer 2010.

In this episode, Immad Akhund and Raj Suri talk with Rahul about the origin of the Superhuman name, how a leaked demo link took Rapportive from 10 to 10,000 users overnight, and why being early to a new platform still pays off. Most of the conversation covers M&A: how to read a buyer, how to protect your leverage, and what a banker is actually for.

What you'll learn:

1. The four things an acquirer needs to be buying for a company to sell at a top valuation
2. How Rahul sold Rapportive at a premium with two weeks of runway left, including letting a no-shop period expire
3. Why building multiplayer features moved Superhuman's team net dollar retention from about 70% to 122%
4. Why founders should be early to every new platform, from browser extensions to ChatGPT apps and MCP
5. How the founder "bell curve" changes how much the idea matters
6. What M&A advisors actually contribute, and how they pace hot and cold buyers

Chapters:

(00:00) The four things acquirers are really buying

(01:07) How Grammarly became Superhuman

(04:15) Superhuman vs. superintelligence

(07:04) From the BBC Micro to a Cambridge PhD

(09:48) Fundraising on good vibes at Cambridge University Entrepreneurs

(12:22) Building Rapportive

(14:14) The leak that brought 10,000 users overnight

(18:19) Be early to every platform

(20:11) Email's trillion-hour problem

(22:33) Raising before Demo Day and meeting LinkedIn

(23:56) The LinkedIn API deal

(29:21) Does the idea matter? The founder bell curve

(31:33) Going multiplayer: net dollar retention from 70% to 122%

(33:45) Selling Rapportive with two weeks of runway

(44:20) Selling Superhuman: 16 buyers and the banker's real job

Stephen Balaban is co-founder and, as of a recent leadership change, CTO of Lambda, the AI cloud infrastructure company he and his twin brother started in 2012. It took five pivots — augmented reality, a facial-recognition contact book, a camera embedded in a baseball cap, the AI image app Dreamscope, then finally workstations and servers — before Lambda found a business that made money in 2017. Along the way, Stephen kept the company alive by consulting on the side, including projects with Airbus and the creators of South Park.

What you'll learn:

1. Why it took five pivots and 14 years for Lambda to find product-market fit
2. How side consulting work — for Airbus and even South Park's creators — funded the company through its leanest years
3. Why mainline Silicon Valley VCs kept passing on Lambda even while it was profitable, and why that didn't change after ChatGPT
4. How a Series D round pulled largely from Taiwanese manufacturers and family offices instead of traditional venture capital
5. What pushed Stephen to step down as CEO and bring in Michel Combes, a former CEO of Sprint and SoftBank International
6. Why Stephen thinks the disinformation around data centers — noise, water use — doesn't hold up

Chapters:

(01:10) Meet Stephen Balaban, co-founder and CTO of Lambda

(02:40) Palo Alto in 2012 and Lambda's earliest days

(03:42) Building Heads Up, a facial-recognition contact book for iOS

(05:06) The ImageNet moment and training neural nets on NVIDIA GPUs

(06:14) Five pivots: from augmented reality to Dreamscope

(08:55) Funding the company through consulting, including work with Airbus and South Park's creators

(12:13) Writing down the goal to IPO back in a 2012 notebook

(14:28) Raising a first $600K, including Austin Russell's $20K check at a $400K valuation

(24:29) Lambda's climb from $3M in revenue to a $1B run rate

(26:57) Why Silicon Valley VCs kept saying no — even after ChatGPT

(34:04) A Series D built largely on Taiwanese manufacturers and family offices

(37:13) Immad on Mercury Books, Mercury's new bookkeeping product

(44:12) Stepping down from CEO to CTO and bringing in Michel Combes

(52:34) Debunking data center disinformation, from water use to noise

What if the company students use to find their first internship became one of the most important players in training AI?

Garrett Lord, co-founder and CEO of Handshake, joins Immad Akhund and Raj Suri to break down Handshake's unlikely pivot. Handshake started as a way to help college students — regardless of where they went to school — find internships and jobs, and grew into a $200M+ ARR business used by most students in America. But over the last 18 months, Garrett has built a second business inside Handshake: using the company's network of 30 million students and alumni to help AI labs train their models on real, high-quality, professional-domain data — from oil and gas to finance to scientific research. That business alone has gone from zero to nearly $2 billion in revenue in about a year.

The conversation goes deep on how this actually works: recruiting domain experts, building task environments that function like video games, scoring model performance against expert-validated tasks, and why 70% of the money spent training a model today goes toward reinforcement learning rather than pre-training. Garrett, Immad, and Raj also cover the open-weight vs. frontier model debate, why China may already be ahead on robotics, and what jobs might look like in a world where AI models can eventually learn continuously, on the job.

The episode closes with a genuinely open-ended debate between Garrett and Immad about what humans will actually do for a living, and for meaning, if knowledge work is mostly automated — and how disruptive that transition might be along the way.

What you'll learn:

1. How Handshake used its network of 30 million students and alumni to build a second, multi-billion-dollar business training AI models
2. Why 70% of AI training spend now goes toward reinforcement learning, not pre-training
3. How AI labs identify gaps in their models and commission the specific data needed to close them
4. Why data, not algorithms, may be the real long-term moat for AI companies
5. Why computer use — AI navigating real software and websites — has recently gotten dramatically better
6. Why China may already be ahead of the U.S. in deploying real-world robotics
7. How enterprises like Mercury are likely to use a mix of frontier and open-source models going forward
8. What Handshake learned scaling a data business from zero to nearly $2B in under two years
9. Garret and Immad's differing views on what human work and meaning look like if knowledge work becomes automated

Timestamps:

(00:19) Introduction and Handshake's origin story

(01:39) Handshake's new business: training AI models on real-world data

(02:30) How Handshake's 30M-person network became a moat

(04:24) Inside the "video game" environments used to train agents

(05:00) Why 70% of AI training spend now goes to reinforcement learning

(08:18) How AI labs commission specific data from Handshake

(12:10) Why computer use has finally gotten good

(14:11) What models are still bad at, and why

(17:00) The "8 people can agree" test for what AI can be trained to do

(18:23) China's 2 million working robots, and why the US is behind

(22:12) Open-weight vs. frontier models, and how enterprises will use both

(31:15) Scaling from zero to $2B: what broke along the way

(34:02) Handshake's "Olympic pace" culture value

(39:41) Why continuous learning is the next frontier for AI

(42:07) Bill Gates' essay on AI, job loss, and taxing tokens

(45:23) Immad and Garret debate what jobs and meaning look like in an AI-driven future

This week, Immad and Raj sat down for a wide-ranging catch-up on the biggest stories in tech right now — from record-breaking acquisitions to what they're each giving AI access to in their own lives.

The conversation kicks off with the OpenRouter-Stripe acquisition and Cursor's $60B deal, and what both say about investing in "obvious" ideas when the underlying trend is right. From there, Immad and Raj get into the economics of secondary markets (including Immad's own purchases of SpaceX and Anthropic shares pre-IPO), why staying private longer might be bad for retail investors, and the case for making it easier for smaller companies to go public.

They also dig into consumer AI hardware — why simple, single-purpose devices like Pocket are breaking through where more complicated products haven't — and trade notes on what they've each connected their own AI assistants to, from email and calendars to health results and scheduled tasks.

What you'll learn:

1. Why "obvious" ideas can still be some of the best investments, if the trend is right
2. What's driving the OpenRouter-Stripe and Cursor acquisitions, and why they matter for developer tools
3. How Immad and Raj think about the risks and opportunities in secondary markets
4. Why Immad believes deep secondary liquidity could be bad for retail investors and the broader economy
5. What's made simple, single-purpose AI hardware devices succeed where more ambitious ones have struggled
6. How Immad and Raj are using AI assistants in their own lives, from productivity to personal health
7. Why "PMF doesn't exist anymore" in consumer products, according to a recent conversation Raj had with Character AI's CEO
8. What it will take for AI to handle more complex, multi-step tasks like buying insurance
9. How Anthropic and OpenAI's revenue numbers compare going into the back half of the year

Timestamps:

(00:47) Introduction

(01:21) OpenRouter's acquisition by Stripe

(02:18) Cursor's $60B deal and the case for "obvious" ideas

(06:03) Why big exits justify high seed valuations

(08:47) AI adoption is still low — why Immad is bullish on the next 5-10 years

(12:56) Buying into SpaceX and Anthropic pre-IPO

(15:17) The case against deep secondary markets

(18:21) Why Pocket is winning in consumer AI hardware

(20:02) Talking to Matic's robot vacuum

(22:07) An idea for family video, and why photo frames haven't solved it

(24:14) What Character AI's CEO said about PMF at a recent Tribe event

(28:28) What Immad and Raj have given their AI assistants access to

(33:08) Scheduled AI tasks, and why AI still can't do the last mile

(35:59) Anthropic and OpenAI's latest revenue numbers

(38:06) The debate over housing density and California's building laws

Mo Al Adham is the founder and CEO of Frec, a brokerage platform he describes as "brokerage 2.0" — building on core trading primitives to offer more sophisticated strategies like direct indexing, long-short direct indexing, and options overlays. Before Frec, Mo co-founded Twitvid, an early video-for-Twitter startup, and later spent five years at Twitter. He founded Frec in 2021 and launched the product in October 2023.

What you'll learn:

1. Why the $1-30M wealth segment — about 10 million US households — controls 40% of all investable wealth in the country, and why it's the fastest-growing segment
2. How direct indexing creates "tax alpha" by harvesting capital losses, and why that's a deferral of taxes rather than an elimination of them
3. How a step-up in cost basis at death effectively forgives the deferred tax bill
4. The concrete numbers: how much a $100k investment can harvest in losses via a classic direct index versus a long-short direct index
5. Why long-term, sophisticated investors have proven far less fee-sensitive than the market assumes
6. Mo's path from Twitvid — an early video app built on top of Twitter — to five years working inside Twitter itself
7. How a frustrating experience with a wealth manager who charged 1% fees for little added value planted the idea for Frec
8. Why Mo's six months of "top-down" market research largely failed, and why a "bottoms-up" approach — starting from what he actually cared about — led him to Frec
9. Why Frec had to resequence its roadmap when rising interest rates undercut its original plan to lead with a cheap line-of-credit product
10. Immad's framework for company OKRs (which he calls "COR") and why he insists on including non-measurable results
11. Why Frec has deliberately stayed out of banking, unlike some robo-advisor competitors
12. Mo and Immad's picks for financial products that should already be obsolete

Chapters:

(00:00) The $1-30M wealth segment and why it holds 40% of US investable wealth

(01:03) Introducing Mo Al Adham and Frec, "brokerage 2.0"

(02:07) Targeting sophisticated investors vs. democratizing access

(03:12) Why long-term investors are stickier and less fee-sensitive than assumed

(07:08) Tax alpha explained: deferral vs. elimination

(09:20) How direct indexing lowers cost basis through loss harvesting

(12:45) Long-short direct index and portfolio tilts

(14:16) Mo's first startup, Twitvid, and getting outpaced by Twitter

(16:29) The wealth manager experience that inspired Frec

(19:49) Vetting the idea: six months of top-down research that failed

(22:18) Switching to a bottoms-up approach and finding conviction

(30:39) Immad's approach to OKRs, called "COR"

(36:24) Frec's pivot from lending to investing as rates rose

(52:16) Rapid fire: AI in fintech, obsolete products, and more

Guillermo Rauch is the co-founder and CEO of Vercel, the company behind Next.js, and previously created the widely-used Socket.io library. In this special episode, recorded live in front of an audience, Guillermo joins Immad Akhund and Raj Suri for an open Q&A covering pivots, ethics, investors, and the future of work in the age of AI.

What you'll learn:

1. The difference between a "lowercase p" pivot (refining focus) and an "uppercase P" pivot (starting over) — and how to know which one you need
2. How to build an ethical framework for operating in an industry full of shortcuts and noise
3. How to extract real signal from investors without letting them drive your roadmap
4. Real pivot stories from Presto (restaurant tablets to voice AI), Lyft (carpooling to peer-to-peer ride-hailing), and Mercury's early product-market-fit signal
5. Why blaming distribution is often easier than blaming the product — and why that's a trap
6. How founders can get their teams to think about prioritization the way they do
7. How Mercury created early demand by deliberately recruiting a broad, vocal set of seed investors
8. What "the future of work" looks like when your team's job shifts from producing outcomes directly to building the systems that produce them
9. How growing up outside Silicon Valley shaped each panelist's belief that they could build something from scratch

Chapters:

(0:00) Lowercase p vs. uppercase P pivots

(1:05) Q&A begins

(1:23) Building an ethical framework in Silicon Valley

(4:38) Balancing customer signal vs. investor advice

(9:53) Pivot stories: Presto, Lyft, and Mercury's obvious PMF moment

(15:34) Why founders blame distribution instead of the product

(16:08) Getting your team to think about prioritization like you do

(18:21) How Mercury created early demand with 60 seed investors

(19:48) The future of work: agents, harnesses, and factories of output

(24:25) Growing up outside the Valley: mentors and self-belief

(28:04) Closing

Dan Teran is the co-founder and managing partner of Gutter Capital, an early-stage venture firm investing in vertical AI and marketplace businesses. He previously founded Managed by Q — an operating system for commercial spaces that grew to employ nearly 1,000 people, expanded nationally, and was acquired by WeWork in 2019. Dan joined WeWork as head of corporate development before leaving after a turbulent six months. He now runs Gutter Capital's third fund ($75M) and the Elbow Grease accelerator, sponsored by Mercury, which invests in early-stage founders in New York City.

What you'll learn:

1. How Managed by Q found extreme product-market fit in lower Manhattan — and why that made expansion harder, not easier
2. Why winning a market can be a trap when the TAM is smaller than you thought
3. The real story behind the WeWork acquisition: a three-year relationship, a theatric walkout, and why great exits are always principal-to-principal
4. Why over-capitalization was more ruinous to Managed by Q than any external factor
5. How to think about Series A benchmarks for non-AI companies today (2–3M ARR, renewals, one productive AE, 3x growth)
6. Why AI-enabled services businesses can be great companies even if they're not venture-scale outcomes
7. The mismatch between what early-stage founders need to raise and what top VC funds are mandated to deploy
8. Why founders should play the hype game — but stay ruthlessly honest with themselves about what game they're playing
9. Dan's take on Adam Neumann: what made him exceptional, where he fell short, and why Dan wouldn't bet against him
10. The "leaders eat last" philosophy — and why holding people to high standards and having their backs aren't in conflict

Chapters:

[00:00] The hype trap founders fall into

[01:31] Managed by Q: founding story and early growth

[02:39] Scaling nationally and selling to WeWork

[04:17] The state of co-working and commercial real estate post-WeWork

[07:18] In-person vs. remote — what actually matters pre-PMF

[11:16] How the WeWork acquisition really happened

[15:06] Realizing the TAM was smaller than expected

[17:09] Raj's parallel experience at Presto

[20:04] FOMO-driven investing and the AI diligence problem

[22:04] Series A benchmarks for applied AI companies today

[25:27] Why founders should aim for break-even before raising

[28:56] The mismatch between venture fund mandates and founder needs

[34:32] What Dan learned about fundraising after becoming an investor

[37:30] Adam Neumann, WeWork, and Flow

[39:30] Leadership, high standards, and the "leaders eat last" philosophy

[42:12] Why founders learn the wrong lessons from Steve Jobs

[47:31] FarmEvo: the drone ag company Dan flew to Karachi to diligence

Colin Angle spent 33 years building iRobot — bootstrapping for eight years without venture capital, surviving 15 failed business models, and ultimately launching Roomba in year 12. What followed was a decade of overcoming consumer skepticism, 70%+ global market share, a public offering on Nasdaq, and eventually a blocked acquisition by Amazon. Now he's back with a new company, Familiar Machines and Magic, building robots designed for human connection — priced to compete with the cost of owning a pet.

What you'll learn:

1. Why Colin believes iRobot would have failed with early VC access
2. How iRobot funded itself for eight years through customer contracts instead of investors
3. The sales tactic Colin used to get Fortune 500 CTOs to fund iRobot's R&D
4. How DoD mine-hunting algorithms and a Hasbro partnership became the technology inside Roomba
5. The wallet share framework for evaluating whether a consumer robot idea can actually work
6. Why adding features to a consumer robot often reduces perceived value
7. How iRobot priced Roomba at $199 with a $42 BOM — and what that discipline required
8. What it felt like to go public, and how everything changes when what you say can be monetized
9. The full story behind the Amazon acquisition attempt and why the EU and FTC blocked it
10. What Familiar Machines and Magic is building and why the pet economy is the target comp

Chapters:

00:00 – Regulators celebrate blocked deals — what Colin saw on FTC examiners' doors

00:53 – Introducing Colin Angle, co-founder of iRobot and Familiar Machines and Magic

02:00 – The "if not us, who?" moment that started iRobot

03:54 – First business model: privately fund a moon mission, sell the movie rights

07:03 – Eight years without VC: "completely unfundable"

08:09 – The CTO sales tactic: present a problem half a step from their real one

09:00 – "Work for no profit, cancel anytime" — the deal structure they used five times

12:05 – Built for 10,000 units, sold 70,000 Roombas in three months

15:03 – "If I had VC early, iRobot would have failed"

18:40 – $199 retail, $42 BOM — the Roomba economics

20:31 – The wallet share framework: which consumer spend are you actually replacing?

32:39 – First interview as a public CEO: "My wife says Roomba doesn't work"

34:42 – The Amazon acquisition gets blocked — 15% market share and falling

42:09 – Familiar Machines and Magic: the new company and the original vision

46:12 – Building robots for human connection, not task automation

Guillermo Rauch, CEO of Vercel, joins Immad Akhund and Raj Suri at a live Founders in Arms event to break down the full arc of building one of the most widely used developer platforms in the world—from a contrarian bet that VCs said was already solved, to a multi-product company powering the future of the web.

Guillermo walks through the three chapters of Vercel's growth: finding focus (trimming a portfolio of open source projects down to the one that had undeniable traction), building repeatability (anchoring go-to-market around customer-led ROI stories), and scaling the company itself as the product. Along the way, he shares how he thinks about feedback, why consensus is a red flag for startup ideas, how customer-led innovation beats internal roadmaps, and what "brand permission" has to do with why Google keeps failing at social.

The conversation also gets into the current moment in SF—the AI supercycle, the anxiety around who gets left behind, and why Guillermo's answer to all of it is the same: product market fit solves most problems. Just stay focused on building.

What you'll learn:

1. Why Guillermo treats everything—including silence—as feedback
2. The "pain discovery" method he uses to extract what's actually broken
3. How Next.js started as a personal solution and became a wedge into the entire cloud
4. Why he deliberately ignores competitors when building
5. The three chapters of Vercel's growth and what drove each inflection point
6. How customer-led innovation produced some of Vercel's biggest revenue lines
7. Why your second product has a higher bar than your first
8. The iPhone and AirPods framework for thinking about adjacencies
9. What "brand permission" means and why it explains Google's failures
10. Why consensus around an idea is a signal to walk away

Chapters:

00:00 – Managing your own psychology as a founder

00:51 – Welcome + live event intro

02:55 – Vercel's web stack vs. agent stack

04:04 – Guillermo's background and first exit to WordPress

05:15 – Spotting the waves: cloud and front end in 2013

08:49 – Everything is feedback; the pain discovery method

10:40 – Short-term pessimism, long-term optimism

13:14 – Opinions vs. ideas: the Jony Ive mental model

16:40 – Chapter 1: Finding focus — how Next.js became the wedge

21:03 – Why consensus is a red flag for startup ideas

21:40 – The MacBook moment: simplicity wins

25:37 – Chapter 2: Repeatability — e-commerce as the GTM unlock

29:30 – Chapter 3: Scaling the company as the product

34:41 – iPhone and AirPods: smart adjacencies to a strong core

38:41 – Brand permission: why Google keeps failing at social

40:18 – The SF culture divide: AI optimists vs. AI anxious

43:09 – The AI gentrification of San Francisco

49:05 – Being your own coach; founder loneliness and burnout

50:46 – What fundraising actually feels like

Karri Saarinen is the co-founder and CEO of Linear, the product and issue tracking platform built for high-performing software teams. A designer by training — with stints at Airbnb and Coinbase — Karri took a different path to founding than most Silicon Valley CEOs. Linear has become one of the most beloved tools in the startup ecosystem, known for its speed, design quality, and now its deep integration with AI agents.

What you'll learn:

1. How Linear evolved from issue tracking to a full product-building system with AI agents
2. Why speed and quality — not features — were Linear's winning strategy in a crowded market
3. How Karri thinks about AI's role in design and why average startup design is getting worse
4. Why designers rarely become founders and whether AI will change that
5. The "Quality Wednesday" ritual Linear uses to keep polish standards high at 120 people
6. How Linear's feature roast process catches blind spots before anything ships
7. What Linear borrowed from Coinbase's hiring playbook — and how work trials outperform interviews
8. How Linear built an open agent platform and why it now hosts more agents than any tool in its category
9. Karri's take on whether designers should write code — and where design thinking matters most
10. Why Linear intentionally pushed PM thinking to engineers and designers instead of hiring traditional PMs

In this episode, we cover:

(00:00) Why designers rarely become founders

(00:53) Introducing Karri Saarinen and Linear

(01:27) How Immad and Karri met 15 years ago

(02:00) What Linear actually is — and where it's going

(03:13) Mercury running compliance workflows on Linear

(05:12) Immad's regret: not investing in Linear early

(06:17) How Linear broke through a crowded market

(08:08) Speed and quality as a product moat

(09:26) Why Mercury and Linear win the same way

(14:23) Linear's AI agent strategy and open platform

(17:40) Coinbase and Ramp building custom agents on Linear

(19:27) Linear's upcoming coding agent and PR review interface

(21:31) Karri's background as a designer-CEO

(23:33) Why designers don't start more companies

(27:15) How AI is blurring the lines between design and engineering

(31:03) What AI can't replace in design thinking

(34:05) Bleeding roles without losing specialization

(36:47) The AI slop problem in product features

(37:02) Maintaining quality culture at 120 people

(39:31) Quality Wednesdays explained

(41:16) The feature roast process

(44:18) How Linear collects user feedback

(46:33) What Linear borrowed from Coinbase's culture

(47:21) Work trials: how they work and why they're better

(53:32) Why work trials benefit candidates too