Idea Compass #3: Faster Horses
Netflix, BlackBerry, industrial carbon capture, and my own failed startups
Today we're looking at the top-left quadrant: faster horses. If you consume the same kinds of podcasts and blogs that I do, you'll probably recognize the title as a reference to the apocryphal quote attributed to Henry Ford:
“If I had asked people what they wanted, they would have said faster horses.”
Faster-horse-like ideas tend to involve linear thinking, and a moderate improvement on something that exists. Faster horse ideas are probably obvious to a lot of people at a given time (e.g a better mousetrap). The really groundbreaking ideas – the metaphorical cars – are a platform shift, and involve some level of creativity, nonlinear thinking, and reframing the problem.
I do want to emphasize at the outset that incremental improvements are not a bad thing! I think most progress in the world comes from small and steady improvements – the “1% inspiration and 99% perspiration” view of progress – and we give too much credit to mythical lone inventors.
Being anywhere above the x-axis is fantastic. But I think some people set out to create a breakthrough idea without realizing they’re working on merely a faster horse.
Netflix, and going from faster horse to breakthrough
Wanting to emulate the internet-based success of Amazon, Netflix started an online DVD rental business. Back in the day, you would choose a queue of movies online and they would be mailed to you. As a kid, I have fond memories of getting those red envelopes in the mail. This was way more convenient than driving to Blockbuster, cost less, and you didn’t have to worry about late returns. It was a business model innovation, but still a sort of “faster horse” of DVD rentals.
Netflix could have stayed a DVD rental service and died like Blockbuster. Believe it or not, the mail-delivered Netflix service still had paying customers1 as of September 29, 2023 when the final discs were sent out. But they successfully anticipated and capitalized on another platform shift, from DVD players and the early internet to smart TVs and fast internet for everyone. In the chart below, you can see an inflection point around ~2010 when the average bandwidth in the US started going up year over year. It was becoming feasible for most people to stream content over the internet, and also run apps on their TVs – a great time to launch streaming services.
If you asked DVD buyers what they wanted in the 90s, they probably would have said “Blockbuster but with cheaper DVDs and a location closer to my house.” Obviating the DVD stores entirely was the superior path.
If you talked to Netflix users in 2010, they might have asked you to get rid of mail, and let them just download the movies as a file to their TV or computer. But streaming the data instantly and in real-time unlocked an even better user experience.
Email machines
There’s a great clip where Steve Ballmer (CEO of Microsoft at the time) is asked what he thinks of the new iPhones:
“Five-hundred dollars, fully subsidized, with a plan?? I said that is the most expensive phone in the world, and it doesn’t appeal to business customers because it doesn’t have a keyboard, which makes it not a very good email machine.”
Obviously, the iPhone was massively successful, a staple of modern life that people will stand in line to pay $500-1000+ for. Steve Ballmer’s parochial concept of a phone was missing two things:
First, he thought of phones as primarily “email machines”, missing the broader possibilities of being able to open up the internet wherever you go in a few seconds through a delightful and flexible UI/UX.
Second, he’s imagining a phone as just a miniaturized computer – so of course it must have a keyboard! Taking something that exists and adding an adjective – smaller, bigger, lighter, faster, internet-connected – is often a recipe for faster horses. It’s easy to see in hindsight, but touch screens were the better option if your goal is to let people do everything on a single device in their pocket. Touchscreens support keyboard layouts for any language and give you the maximum viewing area when the keyboard is hidden. Phones don’t need to play by the same rules as a PC.
Nevertheless, early keyboard-based phones like the BlackBerry were quite successful. Research in Motion (creators of BlackBerry) survived for nearly 20 years, sold over 50M devices per year at their peak, basically invented push notifications, and were ahead of their time on encrypted messaging. Pre-iPhone people apparently loved the keyboard on their BlackBerry, according to my parents. But simply shrinking a computer down to the size of a phone wasn’t enough of an innovation to stay relevant.
Industrial carbon capture, and band-aid solutions
One way to create a faster horse idea is to take something that’s broken and try to fix it by adding more steps or complexity2. We might call these the “bandaid” subcategory of ideas.
For example, the easiest and most obvious way to decarbonize a factory is to keep it the same but add carbon capture and storage (CCS). Many industrial processes have a small number of places where very concentrated streams of CO2 are coming out, so relatively it’s straightforward to capture the emissions. This is a strategy for decarbonizing steel production and cement production, and it’s the most boring one.
But CCS has two problems: (1) it’s a bandaid, and a leaky one at that, and (2) it adds cost. In Neil Hacker’s post on green steel, he points out that CCS only captures about ~50% of the emissions from steelmaking, and it also adds a hefty green premium of around $150-160/ton of steel. Because CCS is taking the business-as-usual process and adding a step to it, it will always have a green premium and never compete on cost.
In this comparison of green steel technologies above, we can see the projected costs of different pathways in 2021 and 2050. By 2050, every approach could become cheaper than BAU except CCS. This is why market shaping initiatives like the First Movers Coalition set their success criteria such that CCS is effectively ruled out as a solution. CCS should be a last resort for the processes that we truly can’t find other ways to decarbonize.
Some of my own ideas
It’s easy to criticize others’ ideas, especially with the benefit of hindsight. I’ve come up with many “faster-horse” ideas in the past, and probably have more of them ahead of me.
During the pandemic, I worked on a startup called Blue Meadow, which was developing sensors and automation equipment for ocean farming. The failure of that idea was overdetermined; the market was too small, farmers didn’t want the technology, I was a solo-founder, etc. But even if those issues didn’t exist, the idea was a faster horse. Better sensor data might improve farm yields a bit, and catch disease earlier. Automated farms could reduce labor costs, but wouldn’t solve the existential pain points that farmers had around permitting and finding buyers for their product. I was building a solution that wasn’t 10x better along any axis3.
A few years later, I worked on another startup called Climate Refarm. We were developing a new kind of carbon offset4 for sustainable food, which would incentivize institutions like schools, hospitals, and corporate cafeterias to adopt more low-emissions (i.e plant-based) options. Failure was also overdetermined: quantifying marginal emission reductions from dietary changes is hard if not impossible, buyers lost faith in the voluntary carbon market around 2022, and the amount of money that our projects could offer institutions was only significant in developing countries. Our mission and team was solid, but carbon offsetting ultimately turned out to be a band-aid solution; it could change incentives a bit on the margins, but wasn’t going to radically transform the food system. Our hardest and best decision was to shut down, return the remaining runway to investors, stay friends, and think hard about how we could each have the most impact.
Takeaways
The original point of this essay was to encourage people not to settle for “faster horse” ideas. In the course of writing, I’ve reversed my opinion. We do need faster horses!
Almost by definition, most progress in the world is incremental. Transistors get a little bit smaller each year, solar panels get a little bit cheaper, someone’s Ph.D thesis extends human knowledge a little bit further on some esoteric topic, and so on. Evolution can only produce amazing capabilities like winged flight through a series of incremental improvements that are each viable5. If you want to build a slightly better grocery delivery app, I say go for it – if you can pull it off, you might make life a tiny bit better for everyone.
I think Rohit Krishnan defends incremental ideas well in this podcast:
... I'm not entirely sure that the types of ideas that the marginal person should be pursuing are going to be more ‘Why don't you go off and try to solve cancer' kind of questions versus ‘Why don't you try and make process X within biotech that deals with clinical trials 3% more efficient?' Because I think the latter is what creates a ton of incremental efficiencies over a long enough period of time — just because there are so many more of them — that the platform that we're standing on imperceptibly slowly creeps up.
There’s nothing wrong with being in the top-left of the chart. But if you do want to avoid faster horses:
Reframe the problem in its most fundamental form. People didn’t want a better DVD store, they wanted convenient entertainment. People didn’t want an email machine, they wanted a portable way to do anything on the internet.
Avoid band-aid solutions. Don’t try to solve a problem by adding more complexity or more steps, especially if cost matters. It might be better to think from the ground up.
Avoid solutions that stack many small improvements on top of each other. The exception is when you’re a vertical integrator and you are effectively your own customer!
Look for imminent platform shifts, and make sure they won’t render your idea obsolete in 5-10 years! Looking at some relevant graphs on Our World in Data is well worth your time.
I guess DVDs still make sense for people with old TVs or very low-bandwidth internet? Shipping DVDs is a form of high bandwidth (but extremely high-latency internet).
Band-aid theories and explanations exist too. Before Copernicus, we had the geocentric (Ptolemaic) model of the solar system, where everything orbited the Earth. When observations disagreed with this model, astronomers had to keep adding new complexities to explain them, like epicycles, equants, and eccentrics.
A common heuristic you’ll hear in the startup world is that a production has to be 10x better for people to adopt it. Packy McCormick’s essays convinced me that there is an exception to this rule called the vertical integrator. Vertical integrators stack many small-to-moderate improvements across an entire supply chain to achieve a superior product or service.
I was tempted to put carbon offsets in the “gimmick” category, since they have low impact and tend to rely heavily on marketing to persuade buyers. However, I do believe that trying to value externalities like emissions or biodiversity through markets is a paradigm shift. Indeed, it’s a seductive concept that seems to appeal to technical people.
Evolution is not a perfect analogy for human invention. Humans are capable of creativity that leapfrogs past many incremental solutions at once.