17 Questions About Me

10 min read
PersonalAIStartupsPhilosophyTechnology

I had a great conversation with a VC friend recently who was trying to understand my career motivations, and I thought these questions and answers were worth sharing.

(1) What has been your favorite job / internship you have held and why?
My first job — an internship at a small start-up that made me realize my heroes were within reach and that my ideas could bring value to the real world at an early age. I also wore a lot of hats and automated a lot of my job.

(2) What are your 3 core values?
Courage, truth, service.

(3) Why are you doing this? What motivates you?
At the heart of it all, I want to help bring about an increasingly beautiful and joyful world where people are better off tomorrow than they were yesterday, and I grew up seeing the technology I loved enable that at scale. I want to help people bring their dreams to life, and I want to bring about a world that wouldn't exist if I didn't do the best I can.

(4) Have you ever been #1 at something? What?
Debate! I was a successful high school and college debater, and that foundation has built within me an almost infinite sense of curiosity and ability to consider different perspectives. It also allows me to know very quickly what answers I need to know to reduce uncertainty about something because that’s how I would convince people that my side was the correct one within a few minutes.

(5) How do you learn new things?
Through real, compounding commitment to reading, dialogue, and practice.

Reading: I go into any text with an initial opinion or perspective that I can benchmark new material against. I also picked up a habit from one of my Harvard professors of becoming technically competent in any field by opening up a college textbook, which allows you to engage with the state of the art in research because you understand the lay of the land.
Dialogue: the benefit of working with people who are experts at different things — you can sit down with them and debate them so that you can adjust your priors and get to the cutting edge fairly quickly.
Practice: use the new technology or try to map out what would be required to reconstruct it. I like getting my hands dirty.

(6) How do you set goals? When you don’t achieve a goal, what do you do?
I set long-term goals as an outline on the basis of the world I want to see (think -- what do I want my life's story to look like if someone were to write a book about me?) and short-term experiments or wins that I can use to keep myself on track, and I do this across a wide, well-rounded “portfolio” of interest areas and personal aims.

If I don’t achieve my short-term goals, (1) I make sure to do very easy things that can give me at least some materially real progress; and (2) I recalibrate and set new, more achievable short-term goals based on my new understanding of what happened, and I treat these things like personal experiments.

But this is easier said than done — dealing with true failure is something I struggle with as I hadn’t really had to set goals with real stakes and considerations that go beyond what others had done before me until I graduated from school and started living without the safety net I'd had as a Harvard student.

(7) What are your 3 favorite hobbies/pastimes? (having nothing to do with work)
making perfume, reading high fantasy with beautiful world-building, long early-morning walks

(8) What thing about technology do you think is true that most people think is false?
I think the bulk of the next decade’s decacorns will come from technology with:
(1) clear, novel consumer use cases + lots of up-front R&D and domain expertise or serious regulatory or coordination challenges; or
(2) unparalleled cultural or community significance and loyalty
as opposed to the generalist software stack across different niches that defined the golden 2012-2018 era.

I think most people see AI as a force that will allow anyone to build a company and usher in an age of innovation as a result. While that’s likely true in some ways and people will continue to make some money building generalist software in different niches, those areas will become saturated and leave the really outsized outcomes to the groups I mentioned. The real benefit of AI comes from allowing you to focus on areas of differentiation.

Another way to describe it is that the only venture-scale technologies to exist after we reach a post-AI equilibrium will be technologies with significant market or technical risk and not just execution risk… this means a return to VC’s roots.

(9) Describe a startup you think should exist but doesn’t yet
AI for interactive group education in a classroom.
This wasn’t possible a year ago but advances in speaker diarization and agentic voice frameworks will enable the early technological foundation for collaborative interaction under the instruction of an AI within the next year.
The structural problems that make AI-augmented education inferior to classroom education at scale right now (requirement of deep behavioral changes; misunderstanding; lack of accountability; lack of true critical thinking; and most importantly, lack of socialization) are solved by embedding the AI in an educational setting that has other real students, many networks of which already exist.
This is the way forward for any serious educational or human capital use case, and a company that can build the technical infrastructure for AI-guided multi-party collaboration at scale (hardware, software, and operations) will win and unlock many new markets in the process, with education being a very sensible starting point saturated with latent demand.
There are almost 4 million teachers in the United States alone teaching upwards of 50 million K-12 students, and Americans spend approximately $17,000 annually per student for public education through their taxes. Even with that enormous number, there is a gap in education outcomes and demand for more teachers. Some of this value will be captured by advances in individual education, but there will always be serious demand for the other, complementary roles schools fill — socialization and a place for students to be while their parents are working. Selling to schools is not easy, so the company that can implement this technology at scale will have a strong moat.
This is a huge market — capturing $200 annually per student (1.18% of the annual budget paid for by taxes, or the equivalent of hiring additional teachers who can teach 200-400 students each within the existing paradigm) is $10 billion in revenue in the US alone. $2,000 per student, or 12%, is $100 billion.
There are also other large, diverse markets for AI in education across different kinds of schools, tutoring agencies, niche communities, corporate trainings, creator courses, and so on — with consumer, enterprise, and marketplace use cases that all have different sales cycles and price points — that could provide alternative revenue streams...

(10) How do you spot talent?
I've met some extremely talented people over the course of my life so far. These are some of their shared traits:
- optimistic frustration with the status quo and a need to fix it
- usually a general philosophy on how the world works that's rooted in history and the ability to spot gaps in what they perceive
- people whose interest in and vision for the world goes beyond themselves
- people who achieve a flow state or personal joy from the work they’re doing
- people who are able to carry a conversation with lots of different kinds of people with different areas of expertise and ask them good questions
- people who don’t just accept things as true without examining them themselves and who have a tenacious sense of curiosity
- people who could plausibly do any kind of job in the world and act instead on the basis of what they want to do
- humility + ambition

(11) If you could invest in any private tech company right now at the valuation of the last round, which company would it be?
Osmo AI. Not a standard seed company as it spun out of Google Brain as a joint venture with early-stage investors, but I predict that the company will become a decacorn within half a decade.

(12) If we asked you to pick an industry or theme within AI and bet the whole fund on that one industry or theme, what would it be? What about an industry or theme outside of AI?
AI:
Screenless productivity. If you can move human beings away from screens, not only do you make people’s lives better, but you’ve also figured out how to run an economy with agents. requires creating actionability from words or intentions, and covers new data, new interfaces, AR/VR, multimodal agents, etc.

non-AI:
If deep tech is too broad... maybe the creator economy across different kinds of content and in the broader micro-industrial sense. A world with more automated digital labor will push people to leisure, and the creator economy will facilitate new forms of entertainment, new forms of income, new forms of education, new forms of politics, and new forms of community, the five of which will form the cultural bedrock of day-to-day life and consumption for most people. Lots of different problems open for solving — new forms of social media, integration with real life (digital-physical bridge) and local communities, post-smartphone content, facilitating many-to-many participation, community fundraising and shared ownership/investment, governance, and so on. But they’re hard to get right. In simpler terms, what does the new economic engine look like when you make standard labor unnecessary? And my bet is on the rise of culture.

(13) What’s a question you’ve been pondering lately that doesn’t have an easy answer?
How do we revert the course of technology that'smade people sadder, more socially isolated, and mentally numb, and instead make them happier, healthier, and smarter?

(14) What online community are you fascinated by and why?
Urbanists on Reddit and Twitter. As an urbanist, I’m a huge fan of the ideas, but even just generally, these communities are full of thousands of people who have huge dreams for what the world could look like and no real outlet for bringing those dreams to life. I love talking to people who build mental worlds and work toward making them a reality, even if there's not always a present market for it.

(15) What’s the most interesting way you’ve incorporated AI into your day to day?
I've been recording my thoughts as voice memos for the last five years, and I run them through an LLM that gives me summaries, tasks, and first drafts for my action items across all my different thoughts. I'm working on integrating this even more into my life in real-time as an all-encompassing second brain with agentic capabilities.

(16) What application of AI do you think should be illegal?
Anything that facilitates surveillance or killing, including wide-net predictive law enforcement and legal verdicts without human decision-making. Even if there are potential safety benefits from some angles, there are some doors that don’t close once you open them, and they create lots of room for tyranny.

There are other applications that might make the world worse off and which I personally do not want to see take over, but I'm not sure whether they should be made illegal.

(17) Will VC be disrupted by AI? Why or why not?
I actually think it’s one of the few job functions that requires human beings as partners to the innovative process and which will always be necessary, but many of the less glamorous parts of the job can and should be automated so that good investors can focus on building their dream engine. Automated research will put us at an equilibrium that will allow truly novel thinking, reputation with founders, and speed/consistency to stand out.