How To Be Contrarian — And Right
Garry, Harj, Jared, and Diana discuss why founders should pursue contrarian ideas in a crowded AI market. They analyze how companies like Uber, Coinbase, and Flock Safety found massive success by tackling non-obvious, legally ambiguous, or seemingly impossible problems that others ignored.
The Crowding of Obvious AI Ideas
As Peter Thiel says, “competition is for losers.” In the current AI landscape, competition is intensifying. The initial “gold rush” period, similar to the dawn of the internet and mobile, where numerous greenfield opportunities existed in various verticals, is coming to an end. For about two years, it was relatively easy to find a workflow in a sector like insurance or banking and apply AI. However, most of these obvious ideas have been picked over, and the market is now crowded with multiple startups in each vertical.
Furthermore, the rapid, step-function improvements in AI models that previously expanded the idea space have slowed. This new environment demands that founders move beyond obvious applications and discover a unique, contrarian insight to build a defensible company.
The Nature of a Contrarian Bet
A non-obvious or contrarian idea doesn’t just feel intellectually novel; it often feels dangerous and scary. It’s the kind of idea that could lead to wasting years on a failed venture. These ideas often go against prevailing mental models and media narratives. Founders might hear that “no one has ever built a big company in this space,” or that an idea is a “tarpit” because of past failures.
However, a major technological shift like the current AI revolution is precisely the moment to question these assumptions. The fact that a space is littered with the “dead bodies” of pre-AI companies can be a signal of opportunity, as new capabilities (like agents that can think and act) can fundamentally change the odds of success. The key is to find signal in what others dismiss as noise.
Case Studies in Contrarian Success
1. The Mobile Revolution: Uber, Lyft, and DoorDash
When the iPhone launched, no one predicted that the biggest winners would be companies like Uber and DoorDash. These ideas were profoundly non-obvious.
- Legal Gray Area: A primary reason these companies weren’t started sooner was that their core operations seemed illegal. Founders of services like Lyft were genuinely worried about going to jail. The prevailing laws, written decades before the smartphone, were designed to prevent unaccountable, unsafe taxi services.
- First Principles Thinking: The founders bet that if a service provides immense value to the end-user, society and its laws will eventually adapt. The smartphone provided a new layer of accountability and tracking that made the old taxi medallion system obsolete. The lesson is not to do illegal things, but to identify laws that are misaligned with new technological realities and profound user needs.
- Challenging Playbooks: DoorDash entered a crowded food delivery market. At the time, the popular playbook was the “full-stack startup” (e.g., Spoon Rocket, Sprig), which involved not just delivery but also cooking the food in ghost kitchens. DoorDash’s contrarian bet was to reject this model and focus purely on being a marketplace and logistics app, which ultimately proved to be the right approach.
2. Coinbase: Embracing Regulation
In the early days of Bitcoin (2010-2012), the dominant culture was cypherpunk and anti-establishment. The community valued anonymity and was hostile to laws and regulators. Brian Armstrong’s contrarian bet was the exact opposite: he actively worked with banks and regulators, implementing KYC/AML laws. This increased friction and enraged the existing user base, but it was based on the non-obvious belief that, eventually, regular people would want to trade crypto and would require a trusted, regulated on-ramp.
3. Flock Safety: The “Unfundable” Startup
Flock Safety presented multiple red flags for a typical VC:
- Hardware: It sold a physical camera.
- Go-to-Market: It initially sold to small neighborhood associations, a market with a seemingly tiny Total Addressable Market (TAM).
- Location: The company was based in Atlanta, not Silicon Valley.
Despite these “rules” that made it look unfundable, the company was solving a desperate human need: safety. Founder Garrett Langley focused on the customer’s problem from first principles. The initial product—a solar-powered camera with on-device computer vision—was simple but effective. When the technology started solving major crimes, including kidnappings, its value became undeniable. This real-world impact created viral, word-of-mouth growth, as one police department would see a neighboring city solve a major crime and immediately want the technology. Flock Safety is now valued at over $7.5 billion and solves 10% of all reported crime in the US.
The “Sci-Fi Founder” and Impossible Ideas
Another category of contrarianism involves tackling ideas that seem technically impossible or are dismissed by the expert establishment.
- OpenAI: When OpenAI started, the academic AI research community was largely negative, criticizing the founders for their lack of peer-reviewed papers and the audacity to claim they could build AGI. The focus on scaling laws and spending millions on GPUs for practical outcomes, rather than publishing papers, was a contrarian approach that went against the established incentives of the field.
- SpaceX: Elon Musk was not the first billionaire to start a rocket company, and the press was initially dismissive. The idea of reusable rockets was considered blasphemous by established rocket scientists. It required immense conviction to persevere through years of failed launches and negative press.
In these cases, founders must be a magnet for the few people who share their belief, even when nine out of ten people—including experts—tell them they are crazy. The key is to derive your sense of reality from first principles, direct user feedback, and personal experience, not from the prevailing narrative on social media or in the press.