It was an honor to be on the Smart Driving Car podcast with Cade Metz, Fred Fishkin, and Alain Kornhauser for a thoughtful discussion of what makes driverless different than an Internet app. Inspired by Metz’s recent New York Times article, The Costly Pursuit of Self Driving Cars Continues On and On and On, the focus of the discussion was on the challenges of crossing the chasm to mass adoption of driveless. Metz is also the author of the book, Genius Makers: The Mavericks Who Brought AI to Google, Facebook, and the World.
As alluded to in the above video, driverless is similar to cable television, broadband, or video on demand in the salad days of those markets. All of those technologies had to slog through on a market-by-market basis, learning the nuances of their particular Operational Design Domains (ODD).
Driverless – The New VOD? #
Let’s look at one of these, the VOD business. 25-years ago, it had its own Level-5 hype of “any movie, anytime, anywhere”. In many ways, technology was the easy thing (well, not really). But, even if the technology had been perfected 25-years or even 20-years ago, the Level 5 vision couldn’t have been implemented because of business reasons.
For example, the content owners didn’t necessarily have rights from the various talent to distribute content in an on-demand setting. It took a while for them to rewrite agreements for an on-demand world. Of course, it wasn’t always worth it to go back and get on-demand rights for older content as the legal fees couldn’t be justified. This was particularly true in the early days when there were not that many eyeballs.
Further, on the distribution side, the studios and content owners had a spider web of content agreements that maximized their profit by windowing the content (i.e. studio release, premium channels, cable, broadcast, etc.). They had to carve out an on-demand channel as a new distribution channel that grew their pie, while not upsetting any of their existing distributors. This literally took a couple of decades.
Finally, the content had to be delivered securely (content owners had grave concerns about piracy) and with high-quality (Kevin Costner wouldn’t want Buffaloes that skipped because of poor digital encoding). Of course to do all that in the early days was expensive from the cable distributor’s standpoint, as meant expensive storage and bandwidth.
There were limits on the number of available movies available at any time, as well as how many movies could be simultaneously delivered. Particularly in the early days, each VOD network was isolated, often proprietary to a given ecosystem, and not part of a national network.
Then there was the consumer experience, people understood trick functions (pause, rewind, fast-forward) from the ubiquitous VCRs and emerging DVDs. The selection paradigm was the physical video store. For various reasons, it was a huge challenge to replicate the video store experience. It probably took an outsider, Netflix, with an approach unencumbered by legacy to finally create a VOD interface that worked.
So despite the many waves of hype and disillusionment, on-demand video slowly advanced. It did so by starting with a narrow operational design domain (e.g. procuring niche content, such as documentaries, where rights could and would be granted). Over time, new models, such as subscription on-demand, which decoupled the purchase and viewing decisions, broadened the base.
The Year 2000 All Over Again #
As the Internet became a viable distribution medium, and as new content creators and distributors came into being, new forms of on-demand became ubiquitous. Finally, in 2021, we are almost living up to the hype of VOD in the late 1990s.
The driverless industry feels similar to the year 2000 for VOD. There are some hyped-up visions of what 2040 might look like. Driverless is sure to start with narrow ODDs (e.g. autonomous container movement at ports, first/last mile shuttles, etc.) and broaden over time. “Fasten your seatbelt” is probably the perfect metaphor for the driverless road ahead.