,
14 tweets,
5 min read
Read on Twitter
Tweetstorm time! Six months ago I thought Waymo was the clear leader in self-driving technology. Now I think they've gotten themselves stuck in a technological cul-de-sac. Here's why my thinking changed... arstechnica.com/cars/2019/02/g…
Following @ericries and @claychristensen, I've long been most bullish about companies that get a basic product to market quickly and then iterate to make it better over time.
Tesla and Waymo are both pursuing incremental strategies, but they're very different. Tesla is incrementally going from "level 2" to "level 4." I've long thought (and still think) Tesla's strategy was unworkable.
Waymo's strategy was to start in Phoenix and expand geographically over time. Six months ago seemed to me like a smarter, more workable form of incrementalism. If you haven't figured out snow, you can not offer service in snowy places until you figure it out.
When Waymo touted testing of fully driverless cars in November 2017, that seemed like confirmation that their strategy was working. I assumed they would be able to turn those driverless cars into a commercial service in a few months, then expand to other cities in a year or two.
But I kept hearing rumors that Waymo wasn't testing driverless cars in significant numbers—rumors that were confirmed by @amir's reporting in August. theinformation.com/articles/waymo…
Then in December, Waymo launched a "commercial service" that was only open to existing members of its early rider program. It had two Waymo employees in each vehicle, which means they were losing a lot of money on every ride.
The thing that worries me isn't just that they missed their self-imposed deadline for a commercial launch, it's that they seem to have no idea how far they are from launch. They thought they were almost ready for driverless operation more than a year ago. Now who knows?
And that made me think harder about Waymo's strategy, and I realized it might actually not be incremental enough. Startups without Google's billions were starting with much simpler, safer, cheaper services. arstechnica.com/cars/2018/06/h…
In December I wrote a piece arguing that Waymo was making a mistake by trying to go straight to a full-speed taxi service. At the end of the article I mentioned in passing that Waymo could wind up like Xerox. arstechnica.com/cars/2018/12/w…
Then I decided to read @hiltzikm's book about Xerox to see how well that analogy held up. And boy does it ever! Both companies invented technology far ahead of its time. Both companies wasted years on quasi-commercial user testing before launching a proper commercial product.
As the commercialization process dragged on, both companies saw employees to quit and start their own companies: Adobe and 3Com were started by PARC veterans, and another guy quit PARC to lead the Microsoft Word team. Nuro, Aurora, and Otto, were all created by ex-Waymonauts.
More fundamentally, Xerox built a $16,995 computer because they thought it was beneath their dignity to build a smaller, less capable, cheaper computer. Apple had no such hangups so they won the market with the Macintosh.
I suspect Waymo is making the same error—that they think (in @olivercameron's words) "we might get laughed at" if they started with something simple like @voyage's retirement community taxi service or @nurobots's grocery delivery service.
