The Gingerbread Man is Our Final Defense
Speed is everything in the AI era
How does a tech startup stay in front of the competition? How does it gain a “moat” to protect itself and continue to grow?
One of the most straightforward and common initial approaches is to pursue a segment where others aren’t playing. Many start this way with a “new idea” and “new approach.” This is wonderful right up until the moment you succeed and others, whether startups or existing players, decide that they want a piece of that.
Pebble launched a successful and viral smartwatch and app. This was great right up until every major technology hardware company decided to build their own and Pebble went bankrupt.
“Nobody else is doing this yet” is not a moat. So, what can be a moat? Until recently, I’ve liked simplifying it down into three D’s: Distribution, Design, and Data. Before we go over why AI upended our apple cart of competitive moats, let’s walk through what these moats traditionally looked like.
Distribution consists of two abilities: a) acquire customers, b) and retain them.
Uber was legendary at simultaneously building “two-sided” markets of drivers and users in new cities. A virtuous cycle of network effects sustained that advantage, but city entrance skills were key.
Atlassian (Jira, Confluence, Trello, etc.) has minimal network effects across companies, but it reached a $20 billion market cap essentially without a sales team. Instead they pursued a word-of-mouth, marketing, and minimal friction sign up.
Data is one of the most commonly recognized moats. In some categories, it’s difficult to acquire customers without data, and it’s difficult to acquire data without customers. The data could come directly from your users, or they could come from a parent company or other relationship. The key to a data moat is that whoever possesses the most data will have the best product.
Bloomberg has gained a reputation as the “comprehensive repository of accurate, real-time data.” That’s a long hill to climb to provide parity with Bloomberg, and your best hope to compete is to carve small niches.
Because Google captures how customers respond to search results, they can serve the most relevant search results. The design underlying technology is important, but Google’s granularity for human behavior on search engines was unparalleled for decades.
Design moats create a customer and product experience that creates brand loyalty and equity.
Headspace is a mindfulness and meditation app that lives on a principle of Design.
A user’s 10‑minute session does not improve because millions of others subscribe.
The catalog is proprietary, but guided‑audio IP is easy for Calm, Insight Timer, or YouTubers to replicate
Headspace’s moat relies on the instantly recognizable design language and trusted tone
Squarespace’s library of high‑gloss, editorial‑grade templates sets the aesthetic bar for DIY sites.
One store or portfolio does not accrue extra value from another customer’s site, so little network effect exists.
Competing builders (Wix, Webflow) can match features, and page‑analytics.
What keeps many creatives paying is the perception that Squarespace = tasteful design out‑of‑the‑box.
AI is coming for your moats.
Data moat erosion – Because AI has picked up information across the entire internet (stored in its “latent space”), you often only need to tap into that data by providing customized user information and scenarios into the AI. Once you do that, you can generate high quality analyses, classifications, and decisions in return.
New product surface area – Google previously knew “everything” about you, yet in just a few months of ChatGPT use, users willingly divulged far more detailed and valuable personal information.
Tech Speed Unlock – Both incumbents and startups can develop technology faster than ever with AI. Incumbents can rapidly build and deploy AI features to existing customers. Startups can quickly launch competitive AI features within just a couple of months.
These examples are only the beginning. The landscape has shifted.
Only the Gingerbread Man Moat Remains
What is the last moat standing? Velocity. Go faster than the competition. Many used to poke fun of “velocity” as a competitive edge. Erik Torenberg from a16z reminded me of the strategy Ben Thompson once dubbed as “The Gingerbread Man” moat.
Run, run, run as fast as you can.
You’ll never catch me, I’m the gingerbread man.
Cursor is considered one of the most successful companies of the AI era due to the speed of its high-quality implementation. The design is sharp, the data is helpful, and the distribution strategy is solid. However, if Cursor paused development for six months, it probably wouldn’t survive.
Currently, it feels like velocity is the only moat that matters in AI. If you pause for six months, your company is effectively dead. That wouldn't have spelled doom for any of today's trillion-dollar market-cap companies. Speed has always been important, but Amazon, Facebook, and others likely would have endured a six-month pause. Once your company gains visibility, expect a pack of rabid wolves at your back for years. It’s exciting for bystanders… and exhilarating and terrifying for participants.
How extreme is the velocity shift? GitHub Copilot just flipped to open source. They open-sourced their product because they were falling behind Cursor. If GitHub alone can't outperform Cursor, perhaps the open-source community can at least put them on equal footing.
It seems like a crazy move, but GitHub can afford it because Copilot is a critical moat supporting GitHub’s core product. If GitHub Copilot were a standalone product, they would be fighting for their lives.
“Work smart not hard” only works against dumb people, and AI unleashed a lot of smart in the world. The competition is working both very smart and very hard.
Velocity is the moat that matters. AI’s pack of rabid wolves is nipping at our heels. Life would probably be a lot chiller if I ran an HVAC business right now.