Tesla Must Buy Super Bowl Ads

FSD 14 needs a traditional marketing approach

While most of our content comes in the form of our Tuesday newsletter, we also find it important to occasionally share our thoughts on where we believe the industry needs to head. In the spirit of the 2026 Ride AI Summit theme “It’s Time to Market,” we thought it was worth publishing this piece on how the industry must do more to educate everyone on the importance of what is coming with autonomous vehicles and how much of the future is already here. We will also be publishing one of the largest surveys on consumers’ perceptions of autonomy at our Summit. We look forward to the discussion and hopefully seeing many of you in 2026.

I. The Invisible Hand Behind Modern Success

In 2006, sociologist Duncan Watts built an artificial music market called MusicLab, as described in a 2007 New York Times Magazine piece. He divided 14,000 listeners into eight parallel “worlds,” each of which was asked to rate and download unknown indie tracks. Some worlds saw only song titles. Others saw download counts from previous listeners.

The results have stuck with me for nearly twenty years. Whenever listeners could see those counts, inequality and unpredictability exploded. Hits diverged from world to world, and middling songs occasionally rocketed to #1. Quality still mattered at the extremes (terrible tracks seldom triumphed), but for most songs, chance plus social proof decided everything. As the New York Times put it, “the best songs rarely did badly, and the worst rarely did well, but any other result was possible.”

Watts had quantified what Hollywood script doctors have said forever: nobody knows anything. The crowd’s early whim snowballs, and the rest of us mistake momentum for merit. A song in the Top 5 for quality had only a 50% chance of finishing in the Top 5 for success. If you’re an artist or a marketer, that insight is either incredibly scary or liberating. Brilliance alone won’t save you, but you can engineer the early inputs: seed the right tastemakers, craft share‑worthy liner notes, and design launch day for visible uptake.

II. Memes and the Architecture of Spread

Richard Dawkins coined the word “meme” in his 1976 book The Selfish Gene, arguing that ideas spread and evolve by natural selection just as genes do. More than three decades later, psychologist Susan Blackmore picked up the thread in, appropriately, one of the earliest viral TED talks, “Memes and Temes” (2008), showing how digital networks supercharge this replication.

Memes don’t ask for belief; they ask for repetition. Their superpower is that they recruit each participant as a carrier. The first person to say “6-7“ and move their hands around gave it a slight edge. The hundredth gave it momentum. By the thousandth, it had become inescapable—not because it was profound, but because visibility had compounded into virality. Now, when the crowd cheers for a basketball team because the score hits 67, it only amplifies the meme, beaming it into even more brains.

Fifteen years later, memes have become the purest expression of cumulative advantage online. They’re subject to the same network effects Watts documented: early shares compound into inevitability, regardless of quality. 67 clearly didn’t proliferate because it’s profound, but because visibility snowballed into more visibility. The mechanics are identical to MusicLab: what catches on early gets amplified until random becomes inevitable.

III. Markets: The 80/20 Reality and Double Jeopardy

In 1896, Italian economist Vilfredo Pareto found that 20% of the population owned 80% of the land. That skewed pattern—the Pareto distribution—has appeared everywhere, from software adoption curves to Spotify streaming data. In many categories, 1% of products now drive 90% of sales.

The pattern has a name in marketing: Double Jeopardy. The term traces back to 1963, when sociologist William McPhee discovered something strange: the fewer people who knew about a movie star, the less those people liked them. McPhee called this “Double Jeopardy”—unpopular options suffer twice. His explanation was elegant: “The lesser-known alternative is known to people who know too many competitive alternatives.” Those with more choices have higher standards.

Marketing scientist Andrew Ehrenberg later documented this pattern across brands: small brands suffer twice—they have fewer customers, and those customers are less loyal. But Double Jeopardy burst onto the marketing scene with Byron Sharp’s How Brands Grow, becoming a kind of meme in consumer packaged goods circles. The biggest brands benefit from both reach and repeat. Why? Because availability drives both buying and liking.

Harvard Business School professor Anita Elberse confirmed the pattern held in movies in 2008: hits weren’t just more watched—they were more liked. The long tail is full of products people only consume when the big hits aren’t available.

The implications are stark: cumulative advantage isn’t just strong—it’s nearly unbreakable. Despite breathless media coverage of “disruption,” the data tells a different story. Dollar Shave Club’s viral launch video garnered millions of views, but after a decade, it captured only an 8% market share, while Gillette maintained over 50%. Most of the time, when we use the word “disruptor” in marketing, it’s a euphemism for small.

The uncomfortable truth is that big brands tend to remain big, while small brands often remain small.

IV. Technology: Visibility as the Product

Cumulative advantage doesn’t just govern what succeeds—it determines what we believe is possible. The gap between reality and perception has never been wider than in autonomous vehicles.

Back in June, I bought a Model Y. Since then, I have put on about 9,000 miles, of which I’d estimate 90-95% were in FSD (since they introduced tracking in FSD 14, I’ve got over 2,000 miles with 96% Tesla at the wheel). When I ask friends what percentage they think it drives, they guess 30%, imagining it only works on highways. This perception gap is the root of the entire problem. The technology exists. The capability is here. But without visibility, there’s no snowball. Check out the most recent data from Tesla on the penetration in the United States, for example. Despite this technology being nothing short of mind-blowing, only around 12% of the Tesla user base uses FSD regularly. What this means is that only about 1 in 1,000 Americans has ever experienced Tesla FSD. We have built this interactive funnel to give you a sense of FSD’s penetration in the US, which you can play with.

In a recent BRXND Dispatch, I called this the “Evenly Distributed Blindfold.” The tools aren’t hiding—they’re being ignored. ChatGPT can already draft research, analyze data, and write code. Yet, intelligent and accomplished people still ask when AI might be able to do these things. The gap isn’t access; it’s acknowledgment.

The autonomous vehicle industry keeps asking when the tipping point will arrive. But, as we outlined in the announcement for this year’s Ride AI, that’s the wrong question. The tipping point requires visibility. It won’t come until enough people see self-driving cars that perception finally matches reality. Waymo’s new city rollout, Zoox’s SF launch, Tesla’s robotaxi fleet—each adds visibility. But the real opportunity is the millions of cars, specifically Teslas, that are already on the road, silently demonstrating a future their passengers don’t realize they’re living in.

Which brings us to an obvious question: Why doesn’t Tesla advertise? Elon Musk has been clear: “I hate advertising,” explaining, “Tesla does not advertise or pay for endorsements. Instead, we use that money to make the product great.”

But this isn’t about manipulating opinion—it’s about revealing reality. Show FSD navigating a school pickup line. Show it handling construction zones. Show it’s almost sentient ability to park. Make the extraordinary ordinary. Every advertisement would be a testimonial at scale, each impression adding weight to the snowball.

The irony is that Tesla has solved the most challenging part—the technology—but hasn’t closed the easiest part—the perception gap. The snowball is sitting there, waiting. Someone just needs to give it a push.