The Longest View

Notes on Startups, VC & Entrepreneurship

“Talk is cheap, but camaraderie builds” - Jackson Cummings (Head of Wellington Access Ventures)

$10bn, 100 Person Company?  

I got into a pretty spirited debate over the weekend that the future state of entrepreneurship across the country will be of the leaner, but larger organization. My take is that it will become increasingly easier for smaller teams to build business to the scale of existing public companies. A 100 person company that does $1 billion in revenue, and is valued at $10 billion feels possible in my lifetime. 

That being said though, that thesis is still being proven out. For example, the top 25 US publicly traded companies have an average revenue per employee of $1.3M. OpenAI which is reportedly at $2bn of ARR is right around the $1.3M ARR per employee mark, and Midjourney which has famously spurned venture capital to date is 44% higher at $1.8M ARR per employee.

Why is this important? Increasingly, I liken aspects of the job of a venture capitalist to that of a biologist. In biology, you can’t just go to your peers and say, “I have made a groundbreaking discovery, the Tyrannosaurus Rex is alive and well” and expect that they will believe you at face value. 

If you are a bad VC, the above behavior is totally acceptable, almost encouraged. It’s encouraged because we are in the business of attempting to predict the future, and being in the business of attempting to predict the future - it’s much easier to be declarative and hand wavy in your predictions, than it is to be specific and nuanced.

Alternatively, a good VC is like a good biologist and a good biologist comes armed with specific evidence to support their discoveries. As such, OpenAI & Midjourney (and by extension all of their competitors) become bellwethers for the limits of AI productivity. If the companies that make up the infrastructure of the forthcoming AI application wave cannot achieve outsized productivity gains, why should anyone else? It will be interesting to see where these two companies are a year from now, in addition to seeing how the sentiment on AI as a broad based productivity lever changes over time.

Self-Driving Present

One of the most shocking things to me about technology discourse and coverage is how little air time and press coverage the driverless cars in SF receive. We quite literally have cars driving themselves around the streets of Phoenix and San Francisco, and from my vantage point - they seem to be doing pretty well. The self reported/researched data that Waymo has put asserts that, “When considering all locations together, compared to the human benchmarks, the Waymo Driver demonstrated:

• An 85% reduction or 6.8 times lower crash rate involving any injury, from minor to severe and fatal cases (0.41 incidence per million miles for the Waymo Driver vs 2.78 for the human benchmark)

• A 57% reduction or 2.3 times lower police-reported crash rate (2.1 incidence per million miles for the Waymo Driver vs. 4.85 for the human benchmark)”

While in the grand scheme of things, the surface area of what Waymo is doing is relatively small (i.e testing is concentrated in geographic areas with the most favorable variable conditions); the long term effects of self-driving cars are fascinating. 

In a world where self-driving cars are prevalent throughout the country you can envision a few things:

• Lower auto insurance rates as the likelihood of severe crashes goes down

• Secular decline of auto repair shops driven by A) lower crash rates and B) lower auto ownership

• Increased consumer services spending driven by not needing to pay a monthly car loan

In any event - the steady simmering of self driving cars continues to be an area to watch, especially if you view them as an extended computing platform and not just a strictly transportation utility/improvement