Special Announcement: I joined Anthropic working directly with the co-founder, and why successful founders are joining frontier AI labs
From co-founding a $200M+ rev/yr company to working directly with the co-founder of Anthropic, why I am now working on humanity’s biggest transformation
Exactly one year ago, I stepped back onto the board of my company Super.com, generating $200M+ in annual revenue.
Today, I’m sitting in Anthropic’s office working directly with co-founder Ben Mann (one of the first author on GPT-3 paper) and CPO Mike Krieger (Instagram co-founder), helping shape the technology that will define the next decade of human civilization.
In the last eight years, I grew Super.com from zero to $200M+ in annual revenue, over $1 billion in annual GMV, and 50M+ users, while raising $150M+ in VC funding and achieving profitability across multiple revenue streams.
Since stepping back to the board , I’ve spent the past year actively building in the AI space - creating content that’s reached 20M impressions across nearly 100,000+ followers (including the viral AI Crash Course with 5,000+ GitHub stars), founding the Lean AI Leaderboard, and angel investing in 30+ AI startups.
I am now thrilled to announce that I’ve joined Anthropic in a unique Special Projects/Technical Chief of Staff role, working across the product, engineering, and research organizations.
This is a calculated bet on the biggest tech transformation, made at what could be the most critical inflection point in human history, that will determine our relationship with AGI.
Here’s the full story of how this happened and what I’m seeing inside one of the most important companies in the world.
The Super.com Foundation: Building for What’s Next
The hardest conversation I’ve ever had with my co-founder happened in December 2023.
“I think I need to step back,” I told him on Christmas. “Not because anything’s wrong. Because everything’s right, and I can see what’s coming next.”
The company was firing on all cylinders: hundreds of millions in revenue, profitable across multiple streams, operating a legitimate super app that millions of people used daily. We had the growth engines humming and a leadership team that I knew could continue to execute and grow the company.
But after being a founder for 8 years, you develop an instinct for inflection points. And every signal I was seeing told me we were standing at the edge of the biggest technological shift in human history.
I had already done everything at Super.com, covering the entire journey from 0-to-1 through scaling to hundreds of millions in revenue and back to new 0-to-1 initiatives. I led engineering, product design, growth, strategy, supply chain, M&A, organizational design, and people management - the full founder spectrum.
The transition was not some sort of dramatic founder exit you read about in TechCrunch. I spent nine months methodically preparing everyone for life after Henry: Hiring the right people, preparing leaders who were ready to step up, and making sure every handoff was seamless.
By the time I actually stepped down, the company didn’t skip a beat (exactly how it should be).
There was also a strategic element to my timing.
We’d been one of Anthropic’s earliest enterprise customers back when they were small enough that we had a shared Slack channel with their entire workspace. I watched their progress firsthand. I saw the models evolving month by month.
When ChatGPT launched in November 2023, I had a hunch it was the starting point for the most significant tech transformation since the internet.
I remember sitting in my office that December, looking at our quarterly numbers, and thinking: “We could keep growing this for another decade. But if I’m right about what’s coming, I’m going to want a front-row seat to making a large impact on AI.”
The pieces were aligning for my next move.
The Break Year That Changed Everything (Spoiler: I Wasn’t Actually Retired)
My girlfriend still laughs about my semi-retirement. It turns out that I may have worked harder as a retired founder than I did running a company.
But I had a mission: I wanted to expand my impact beyond Super.com. I wanted to give back to the founder community that had supported me, build deeper relationships across Silicon Valley, and explore AI in a meaningful way.
So I started writing about AI and entrepreneurship, documenting the emergence of lean AI companies: businesses generating $20+ million annually with just 5 employees using AI tools. That’s when I created the Lean AI Leaderboard to track and gamify this trend.
The timing was perfect. This movement was exploding, and I became one of the primary chroniclers of it.
During this year, I also built a comprehensive investing and content creation experience:
Angel invested in 30+ companies
Worked as a venture partner, sitting in 10+ partner meetings
Became LP in multiple top funds across stages and tiers
Amassed nearly 100k followers across LinkedIn, X, and Substack
Created content that reached over 30M+ impressions
This gave me direct visibility into investment decisions from every angle: startup, VC, and LP perspectives.
But more importantly, it gave me insights, credibility, and relationships that proved crucial for my next decision.
Why I Rejected the Traditional Options
Everyone kept asking me the same question: So what’s next? VC or another startup?
As if those were the only two paths for a successful founder. I get it. That’s what most people do. Build something, sell it, or step back, then either start writing checks or start building again.
But I wanted to understand what I was actually signing up for. I evaluated both options.
Option 1: The VC Route
I threw myself into the venture world for 6 months. Attended dozens of Monday morning partner meetings. Had coffee with over 100 GPs across every tier-one fund you’ve heard of. Sat in on Monday morning partner meetings and deal reviews.
This made me realize that VC is a pure sales job. Period.
90% of deals are outbound. VCs identify trends, analysts research targets, and partners chase down the winners.
This is the paradox of the VC world: Give money to people who don’t need your money, and don’t give money to people who actually need your money.
I carefully evaluated my potential impact in this role and decided not to pursue it because I’m young enough and still have the energy and drive to build something meaningful, rather than allocating and selling capital.
Plus, I could always return to the venture world in the future.
Option 2: The AI Founder Route
Starting another company should have been the obvious move. That’s what founders do, right?
But when you are a repeat founder, the bar gets higher.
If you are 22 and going to Y Combinator, you can pivot three times, find some product-market fit, build a decent B2B AI-Wrapper SaaS business, and call it a win. That’s actually a great outcome for a first-time founder.
But when you’ve just stepped down from a $200+ million revenue + profitable company going towards IPO, the opportunity cost of doing anything mediocre is massive.
I spent months exploring ideas: Lean AI solopreneurship opportunities, various AI application areas, and traditional startup concepts. I even sketched out plans for some interesting B2B tools.
But every idea felt like I was building an AI wrapper: taking someone else’s model and slapping a nice interface on top. Sure, some of these could probably become $10-50M ARR businesses. But why spend the next decade building something that OpenAI or Anthropic could make obsolete with their next API release?
That’s when I realized I was looking at this all wrong.
Instead of building on top of the AI revolution, what if I helped create it?
The Anthropic Decision: Choosing the Future
With the VC or founder options crossed out for now, I decided to explore frontier AI labs. There were 4 major players: OpenAI, Google, Anthropic, and xAI.
xAI eliminated itself as I didn’t want to spend my career making inappropriate content.
Google was complicated. I had worked there a decade ago and remembered the politics of the time. Even current employees described the same dysfunction between Search and AI initiatives. For a company that should lead AI, they seemed to be their own biggest bottleneck.
The real decision was between OpenAI and Anthropic.
Multiple factors pushed me toward Anthropic:
Mission alignment: Anthropic’s philosophy resonated deeply. Building powerful AI while prioritizing safety felt like the best and most responsible approach for technology that could reshape humanity.
Network effects: I knew many people at Anthropic personally who shared similar values:
My Waterloo CS lab partner Cam McKinnon
My former college dormmate Alex Rodriguez (who built Embark Trucks and was youngest person to take a company public)
Co-founder Ben Mann (first author on GPT-3), and
Mike Krieger (Instagram co-founder and now CPO).
Our early customer relationship from the Super.com days had already created trust and valuable connections with the team.
Talent concentration: As a smaller company, the density of exceptional people was the highest of the four labs. Every interaction reinforced this. Empirically, they also have the highest retention of all labs.
Technical superiority: Claude has a superior performance in coding and economically valuable real-world tasks, as well as in practical capabilities. They don’t chase hype or superficial benchmarks, which aligns with my mission and values.
Contrary to what many people assume, my decision to join Anthropic was not a part of some grand strategic plan. In hindsight, it was a series of conversations that kept getting interesting.
I’d stayed in touch with Ben over the years since our Super.com days. One conversation in particular stuck with me.
He was discussing his predictions for the arrival of economic AGI around 2027-2028 and how everything was scaling exponentially. The way he laid out the timeline and the technical fundamentals really made me think deeply about what was coming.
That’s when the weight of the decision hit me. If there’s even a 10% chance he’s right, then sitting on the sidelines writing Twitter threads about AI would be the biggest missed opportunity of my lifetime.
My calculation was simple: relative to a career span, 1.5 years is nothing. If AGI arrives in 2027/2028, I will be inside a frontier lab, watching from the front row seat. If it doesn’t happen, I’ll understand why and what’s actually possible.
The risk-reward ratio was immensely positive.
Life Inside Anthropic: What It’s Actually Like Inside
Walking into Anthropic for my first day felt surreal. Here I was, the guy who’d built Super.com from zero to $200M, feeling like a kid on the first day of school.
The office energy is unlike anything I’ve experienced. You’ve got the Instagram co-founder setting product vision and strategy. The first author of GPT-3 sits beside me, and the creator of Claude Code is across from me.
My old Waterloo lab partner Cam (the same guy who could solve impossible computer science problems the night before they were due) is now building the future of AI. I’ve also worked with a researcher who took his previous company public and countless accomplished founders who have raised from every top fund in SV.
The talent density is extraordinary. But what struck me most was the culture. Despite having some of the world’s smartest AI researchers, everyone is super friendly and approachable. There are zero ego games or office politics.
The people are genuinely kind, empathetic, and helpful. The mission is truly real, and people truly believe in the greater impact of safe, responsible AI.
I especially appreciate Dario’s leadership style. He sets an incredible tone and has a very direct and raw style of communication.
Dario and Jared casually chime in on my Slack threads and are super engaged. They speak with full transparency, answer every tough question, and there’s zero corporate-speak.
There is also complete financial transparency and this radical openness builds trust and mission alignment throughout the organization.
The Chaos Behind the Success
Anthropic’s growth has been insane:
2024: $100 million revenue run rate
January 2025: $1 billion revenue run rate
July 2025: $5 billion revenue run rate
From the outside, this appears to be perfect execution.
But behind all the “fastest growing company in history” headlines, it’s absolute controlled (and sometimes uncontrolled) chaos.
There’s an overwhelming amount of information, and Slack is chaotic. There are hundreds of channels I should be monitoring. Someone’s always pinging about some breakthrough that happened overnight.
And everyone (literally everyone) is a Member of Technical Staff, which sounds egalitarian until you realize nobody knows who actually owns what.
But it still works because everyone’s brilliant and mission-driven. As someone who spent years building operational excellence through OKRs and Mission Aligned Teams at Super.com, watching this beautiful chaos is both terrifying and fascinating.
My Unique Role: At the Frontier
My position at Anthropic is unique - I serve in a Special Projects role as the “Technical Chief of Staff”, working directly with co-founder Ben across product, engineering and research.
This provides significant scope without people management overhead (best of both worlds). My time breakdown atm is:
50% frontier prototyping: Working with research teams on emerging model capabilities and building 0-to-1 product surfaces. This combines my entrepreneurial experience with an AI/ML background.
25% organizational excellence: Process improvement, OKRs, decision frameworks, technical reviews, and helping the company scale efficiently.
20% percent developer productivity: Making engineers more effective.
Remainder: Team building, hiring, and hands-on coding.
The frontier work is especially intellectually stimulating. I regularly encounter research projects that could be standalone startups raising $50 million. The caliber across the organization is extraordinary.
I get insights into the six-month roadmap and help shape where capabilities develop next. I get to see what the next Claude 3.5 moment for agent coding will be, which enables companies like Lovable and Cursor.
Being able to influence these developments as they are emerging is an incredibly humbling experience.
That being said, my role is extremely dynamic. I’m constantly setting up teams for success and hiring backfills to replace me. It’s exciting, and nobody knows what tour of duty and adventure I’ll take next.
Mission Over Money: The Proof
I know many outsiders who are skeptical about Anthropic’s strict focus on AI safety.
Because many AI companies say they care about safety. It’s become the equivalent of saying you are customer-focused or data-driven. Everyone says it. Few actually mean it.
But then I saw how Anthropic actually makes decisions:
The billion-dollar mistake they made on purpose: This decision shaped my understanding of the company’s priorities. Anthropic had Claude ready to ship before OpenAI launched ChatGPT. They could have been first to market. They could have owned the entire conversation around AI.
Instead, they held back because they believed that releasing that level of capability to millions of people overnight could kick off a race to build AGI.
This could have made them a household name in AI, like Google became synonymous with search. But they chose to wait and let OpenAI take that moment.
The history books could have been completely different. We might all be saying “I’ll Claude that” instead of “I’ll ChatGPT that.”
The culture fit check: The culture interview process carries heavy weightage in hiring decisions. I’ve personally been on multiple hiring panels where candidates with exceptional technical credentials were rejected due to a misalignment of culture and values.
Internal decision-making: Every so often, we have discussions where revenue growth directly conflicts with safety considerations. And every time, I watch leadership choose the harder path.
Because the people making these decisions genuinely believe that how we build AI today will determine whether it becomes humanity’s greatest tool or its biggest mistake.
Sitting in those rooms, watching people turn down easy money (billions of dollars) to do the right thing, that’s when it clicked for me: This is exactly the kind of company I want to help build.
The Broader Trend: Founders Join Labs
I’m not alone in this transition. Many founders are making similar moves to frontier labs.
The reasons are compelling: Instead of building AI wrappers externally, you work with cutting-edge internal models and influence their development. Plus, lab equity is often more valuable and liquid than startup equity.
No wonder many founders choose to work with frontier labs:
One founder spent months prototyping a complex problem, making incremental progress. Then Claude Opus launched and solved his entire use case in a single prompt. That was the moment he realized he needed to be building from inside the lab rather than chasing capabilities from the outside. He joined Anthropic shortly after.
A founder who joined Anthropic after raising $100M+ in VC funding and scaling his own company to a unicorn and exiting to a public company.
Another founder who joined Anthropic after scaling his company to hundreds of employees and taking it public.
A founder just sold his 50 person company and decided to join Anthropic instead of staying with the acquirer.
Two founders, each running their own 8-figure healthcare companies, have now joined Anthropic to lead the health and life sciences team.
A 2x founder joined last week
Another 2x founder (also worked at Google Brain) joined Anthropic 10 months ago.
Even from a lab’s perspective, these founders are valuable employees as they understand the whole stack of bringing new innovative products to market.
Having spent years navigating product-market fit, scaling teams, managing complex stakeholder dynamics, and shipping under pressure, they bring first-hand operational experience that pure researchers may lack.
This creates a powerful win-win dynamic: founders get access to create cutting-edge technology products and meaningful equity upside, while labs gain operators who can bridge the gap between breakthrough research and scalable products.
As we continue to move rapidly towards AGI, I expect this trend to accelerate. (In fact, a lot of founders reach out to me frequently, expressing their (secret) desire to join the labs).
Even at Anthropic, over 20% of the candidates I interview are founders or former founders, and this trend is expected to continue. If you are a founder and are thinking about joining a lab, it’s never too late.
Trust me when I say this, because when I joined, I thought I was late to the party myself. But Anthropic’s valuation tripled the day I joined, proving it’s never too late.
What This Means
My move represents a shift in how careers evolve in the AI era.
The traditional founder → angel → investor path is expanding to include founder → frontier lab → unknown. We don’t know what comes next because the landscape changes so rapidly.
Being inside during this transformation provides value regardless of outcomes. Whether AI achieves predicted breakthroughs by 2027 or takes longer, the impact, learning, and relationships built at the frontier will be invaluable.
And now what I’m seeing from the inside has only reinforced my conviction about AGI, though with more nuance.
I’m constantly amazed by new capabilities and breakthrough projects. But I also see the massive amount of engineering work and technical challenges that need to be solved to get to AGI.
We could get there faster than expected if everything clicks, or slower if we hit unexpected bottlenecks. My personal median expectation is still around 2027, but the range of outcomes is broader than I initially anticipated.
Nonetheless, I’m more bullish than ever on the AI transformation now that I have a more complete picture of what it will take to get there.
Thankfully, Anthropic explicitly approved my previous community involvement, so I will continue angel investing, creating content, and supporting founders. This gives me the best of both worlds: frontier access plus ecosystem engagement.
Final Thoughts…
I am grateful for the opportunity to work directly with Anthropic’s co-founders and a highly talented team on a once-in-a-lifetime mission
The research projects underway could each become billion-dollar companies if they were spun out tomorrow.
What keeps me up at night (in a good way) is that we are building the technology that will determine whether AI becomes humanity’s greatest achievement or its final invention.
I am also coming to terms with the fact that I was building a company with tens of millions of users, while now I am potentially shaping technology that billions of people will depend on. The scale of impact is different by multiple orders of magnitude.
The response to this announcement has been overwhelming. Since posting on LinkedIn, many people have asked when I’ll go back to being a founder.
Honestly, I don’t have an answer right now, but what I can tell you is it’s the beginning of something I couldn’t have imagined when I first started building companies.
For now, I’m helping us win the race to the top
And I’ve never been more excited about what comes next.
Onwards and upwards!
Thanks for the great post Henry! Follow up question for you: How do you think you all think about hiring (you mentioned culture as a big driver) versus how you did it at Super or seen it done at other places like Google?