A snapshot of what I’m interested in working on next
This post will be continuously updated. Last updated June 19, 2019
“I GOT NEXT!” is what you say when you walk up to a pick-up basketball game that’s in session to declare that you’re next-up. You gotta do it loudly so people know. I’m at the phase of my startup career where I’m picking my next game to play and I’m doing it loudly.
How I am thinking about my next project
Strong leverage from technology: Leverage can come in many forms (e.g. finance, brand, patents, scale) but the best is still technology leverage. 100% of the decks I’ve seen positions tech as the secret sauce because that’s how you raise from VCs, but does tech really matter? There’s a spectrum with pure SaaS business on one end and WeWork on the other end that could still succeed without any tech. I lean to the pure SaaS side of the spectrum.
Product matters: If the product is actually transformative to the customers (the proverbial “10x better”), you can make a lot of mistakes and the customers will still carry you. There are many startup plays that are about efficiency gains where you might be 2–5x better but the experience is not transformative because it’s stuck in the old paradigm. 10x better frequently requires breaking the existing paradigm.
Disruptive / value re-alignment / bundle, un-bundle, re-bundle: I like playing with these configurations kind of like a puzzle. Pick an underserved segment of the user base, especially in industries that have giant monopolies or duopolies like social, and play with the value alignment to see if you can create a better experience for them.
Shortcut to get started: The shortcut for my first business was jumping on Twitter as a platform and taking advantage of the early user enthusiasm. The shortcut for my second business was selling a B2B service to startups through the YC, StartX, and my wide founder network. A solid shortcut gives us the rocket boosters we need for the initial ascent and from there we can up-resource and figure out how to take on the market.
Working with people I admire: Startups are tough and there’s a lot of luck involved, but I love it because I get to work with people I admire. To increase the enjoyment level, I am looking to work with even more people I love and respect for my next project.
Personal RFS (Request For Startups)
Operationalizing Y Combinator
Many venture capital firms sell value-add to differentiate their money, and while some individuals in these firms can add tremendous value, as an industry they are generally overselling the value-add. In response, Y Combinator and StartX have built software to operationalize something as vague as “helping” founders. While they’re not perfect, they have much higher hit rate at adding value than traditional VCs. I think we can take this further.
Pioneer is operationalizing Silicon Valley by creating a “a home for all the ambitious outsiders” and for the participants it feels like going through a program like Y Combinator, except it’s completely online. All the mentorship, education, community, and regular feedback are online. I think we can “unbundle” Silicon Valley this way. I wrote about this in “The next Silicon Valley is already online.”
The future of social media
Mark Zuckerberg has declared that Facebook’s focus will be on private messages with smaller groups of people. If they move quickly before the public completely loses trust in Facebook, it’s possible they can win. However, as long as Facebook’s business model continues to rely on advertising in its current form, they leave themselves vulnerable to a new paradigm where chasing user engagement is actually detrimental.
Which populations today already feel like “engagement” is detrimental? What metric do they care more about instead?
Thought experiment: Even though Citizen’s police scanner activity is addicting to consume and would be a perfect fit to juice NextDoor’s user engagement, should NextDoor acquire or integrate with Citizen’s content? Or would this cause NextDoor users to abandon the service? My guess is that NextDoor users wouldn’t care about the downside of consuming “crime porn,” meaning engagement is still an okay metric for NextDoor to optimize for. But are there populations of users who care a lot about feeling “safe” or “healthy” in their digital media diets? Enough to pay to control their digital environment?
Evan Spiegel describes Snap as the “camera company.” If they are, I think they’re mostly a front-facing camera company. Recently I tried Highlighter and thought the interaction of using the camera to capture texts on a page and using your finger to help guide the OCR is a very satisfying interaction that I wonder if there are opportunities to either extend this specific OCR + finger interaction to other use cases or if there are other “back-facing camera” opportunities that are not purely for fun and does a little bit of computer vision magic. For example, can you take a picture of what you are eating and automatically track the nutrition information without manual logging (hard, I know)? Can you actually make a bill-splitting app work well enough to replace Venmo if you take a picture of the bill (bill-splitting app is one of the most common startups ideas people tell me about…)? Most AR or object-recognition plays have felt underwhelming, but it feels like there’s opportunity for more product-thinking here.
New platform plays
What would be really interesting is if cars become platforms because of true autonomous driving and we’ll see countless entrepreneurial opportunities for when we are sitting in a car and they won’t be limited to tech entrepreneurs. But we are too far away from true autonomy.
Where are the new platforms, especially ones people aren’t talking about yet. Is it possible to build on top of Zoom given the superior video and audio experience it has enabled? Can we build on top of WeWork that’s not simply selling to their tenants, but actually leverage the commonalities in the physical spaces themselves?
The platforms would have to bring more than just a user base. For example, while I love my AirPods, Siri still sucks at understanding my instructions so we are limited by Siri there. Similarly, Slack has a giant user base but we are also limited by our ability to build truly capable chat-based assistants.
In Tony Sheng’s piece on the psychology of mass movements, he talks about how there are usually three stages to a movement.
- Forebearance (I don’t think this is the correct spelling or definition, but it comes from Sheng): “Men of words” sow the seeds of revolution with idealism that alludes to a better world and discredits the current powers
- Conception: “A fanatic leader,” inspired by the words of the intellectuals rallies a ripe population with promises of sudden and substantial change. The promises are vague and grandiose, set against “the devil” to fear and fight.
- Legitimization: Once the movement reaches critical mass, “men of action” do whatever possible to legitimize the mass movement and convince the population that, “the new order [is] the glorious consummation of the hopes and struggles of the early days.”
He uses the example of Bitcoin:
- Forebearance: Cypherpunks publish subversive pieces like The Cryptoanarchist Manifesto leading to Nakamoto’s Bitcoin Whitepaper.
- Conception: Fanatics such as Andreas Antonopolis, Fred Wilson, Roger Ver, and more evangelize Bitcoin, promising freedom from the tyranny of authoritarian regimes and legacy financial systems.
- Legitimization: We’re not here yet, but this will look something like a group of pragmatists doing everything in their power to maintain a perception that Bitcoin succeeded in fulfilling its promises.
Software companies are great for “legitimizing” a movement by baking the philosophy into a product and aggressively selling it to the market. Intercom did it for the religion of “lean startups / product metrics” and Hubspot did it for “inbound marketing / SEO” What are some of the “religions” of our day?
- Forebearance: Tim Ferriss’ 4-Hour Workweek created digital nomads and people interested in remote work.
- Conception: Matt Mullenweg of Automattic has been leading by example here for many years. Joel Resnicow at Buffer followed and now Pieter Levels is a new radical voice on Twitter touting remote work.
- Legitimization: Github is still largely remote and they were acquired for $7.5B. Rainforest and Zapier are two fast-growing venture-backed startups that are remote. Many companies have WFH days and some companies are permanently partially remote. The tools are starting to come together with the popularization of Slack, Zoom and WeWork, but there’s still a lot of room to grow to make remote possible especially for more dynamic kind of work like early-stage startups.
- Forebearance: DHH and Jason Fried have been banging the drumbeat against startup #hustleporn and published books like Rework and It Doesn’t Have to Be Crazy at Work.
- Conception: DHH and Jason Fried have also been the fanatical leaders of the movement.
- Legitimization: I don’t think startups can be built this way, but I’m sure we can expect more awareness of our energy level and mental health at work and become more suspect of bullshit work practices. Silicon Valley is already better at than most other industries at these things. Slack’s Do Not Disturb feature is a manifestation of a legitimate product feature that is part of this trend.
- Forebearance: Jef Raskin wrote The Humane Interface
- Conception: Tristan Harris made waves by calling out how popular tech companies are addicting their users with A Call to Minimize Distraction & Respect Users’ Attention at Google, then he left Google to create the Time Well Spent movement and Center for Humane Technology together with Aza Raskin.
- Legimization: We are not there yet, but Apple has built greyscale and Screen Time. Facebook launched Your Time on Facebook. Many people I know have deleted Facebook or regularly go on digital detoxes.
Other religions: FIRE (Financial Independence, Retire Early)
Full-stack startups / vertical integration
“The Rabois rule” is great.
Since my last full-stack/vertically-integrated startup failed, I’ve been learning my lessons to figure out how I can apply my experience to increase the odds of success next time around. I wrote about the lessons in “Should you build a full-stack startup? 7 factors for consider.”
I’ve been perusing the complaints on the Better Business Bureau website and evaluating industries with opportunities to vertically integrate. Additionally, I’d like to learn more about franchise models like what Wonder School and WeeCare are doing for pre-schools. My understanding is that franchise models are only as strong as your ability to control essential parts of the business (e.g. brand, technology, and secret ingredients). I suspect that on the spectrum of horizontal to vertically-integrated, franchising could be a good middle ground for some industries (some call this “managed marketplaces” but definitions are hazy).
New tools for new jobs
“Social media manager” was a new job title in 2010. Crowdbooster, my first startup, was built for them from the ground-up and we ended up in the “skills” section of their resumes and in numerous marketing textbooks.
Startups have targeted “gig economy workers,” but that didn’t feel attractive to me maybe because the common surface area isn’t large enough or those weren’t jobs that could move into the “professional” class. What Checkr did, however, was brilliant. There are now recruiters hiring gig workers at unprecedented rates.
There aren’t terrific tools out there yet for data scientists. Startups with widely distributed markets and physical assets need people on the ground. They call them “city GMs” — is that an opportunity? What about Instagram influencers?
Anti-virus for the new world / new ways to build trust
Anti-virus software always played catch-up to the latest vulnerabilities but still offered sufficient protection for the majority of users. In today’s world, the equivalent of viruses are deepfakes, crypto scams, 3rd-party trackers, filter bubbles, flame wars, and dark UX patterns. What should anti-virus software look like today in the era of deepfakes?
For most of the “new viruses,” we won’t be able to fight them head-on without a new paradigm of trust-building. Our current proxies of brand, ratings and reviews, real names, ID verification, mutual friends, background checks, licensing, etc. are becoming insufficient. Crypto’s solution is to make it “trustless” — which may work in some scenarios but doesn’t map to real-world behavior where we prefer proxies for ease and recourse. I would tackle this by trying to solve a problem in an existing low-trust environment or where trust has recently broken down.
Time-shifted content consumption
In a world dominated by algorithms feeding us on-demand content, regardless of when the content was produced, we are starting to see the gap between the time when a piece of content is published and when it’s consumed widen. That means that there will be fewer and fewer experiences where we watch the same TV show or read the same article with our friends at roughly around the same time, but we still crave that shared experience. For example, I frequently read articles I saved months after, and then I’d search Twitter and Hacker News to catch up on the discussions about that article especially since we don’t often see comments below articles anymore. It’s my personal hack and I wonder how much better it could be.
Incidentally, this is the idea we got into Y Combinator with and quickly pivoted away from almost ten years ago.
Best of the best
Over the past few years I’ve consumed some incredible content that consist of experts analyzing world-class experts. I loved the AlphaGo documentary, Kobe’s Detail, Coach Nick’s BBallBreakdown, Tristan Parades, Foolish Baseball, Every Frame a Painting, and the list goes on. Penn and Teller have been breaking down magic forever, ESPN has former players help analyze the game, and Washington Post’s wonkblog, NYTimes’ The Upshot, and Vox.com are examples of similar projects from major media companies.
The secret seems to be that an expert can see things that the rest of us can’t, and if you find the right expert who can articulate what he sees, that makes for great content. Good candidates are people who are #25–#100 in the world at what they do, or coaches. Watch Steph Curry teach you how to shoot a basketball on Masterclass and you’ll see how difficult it is for the world’s very best to empathize with the rest of us. Watching Coach Nick break down Steph Curry’s shot is much more entertaining to people like me.
Ideas that sound like jokes
Toreba — remote control a real-world crane machine. Screw AR, VR, Mixed Reality. How about CR? “Control Reality.”
Uber for wingmen — could be a way to sell companionship without calling it that. They do it in Japan.
Woke Co. — products to help make the world more “woke”
Asian people Yelp — this comes up all the time like how Yelp sucks for authentic Asian food because the reviews are not from Asian people. On a more serious note, black patients / students perform better when they are assigned to black physicians / teachers. It’s probably not PC to advocate for shaping critical services around identities, but maybe there are less sensitive areas (for example, many dating apps are identity based).
I’m doing this in public so I can have more people challenge my thinking as early as possible. I’ll be digging deeper and starting to talk to customers. If you think I can think bigger, tell me how. If you are an expert in one of the areas I mentioned, share your experience with me. If you think you have a strong RFS of your own, pitch me. I’m on Twitter and my DMs are open.
Thanks to Avichal Garg, Michael Gummelt, and Ming Yeow Ng, Dave Yen, David Tran, Jessica Tsoong for reviewing a draft of this post.