An article at Marginal Revolution explains what happened when a website refused to sing the praises of the Tesla Model 3 as a result of its poor braking performance:
This week Consumer Reports changed their review to recommend after Tesla improved braking distance by nearly 20 feet with an over the air software update!
The author is giddy at Tesla’s response time, but I’m not sure I share his excitement. Yes, it’s great that Tesla could fix the car’s performance that quickly, but I am not sure how I feel about my car getting over-the-air (OTA) updates for safety-related items like braking. If an update can improve your braking, it could also worsen it. I’d hate to get in my car on Monday morning only to run into the back of a bus because it downloaded version 5.7.1908 in the middle of the night and doubled its stopping distance.
There’s a broader point here when it comes to the quality of software and the ability to provide constant updates. The optimistic view is that updates allow continued innovation and enhance our ownership experience; the pessimistic one is that they allow companies to ship not-ready-for-prime-time software in the knowledge that the they can always fix it after release. It’s like an engineering Peltzman Effect, where the mitigation provided by easy updates causes the engineers to undertake riskier behavior. Just as Peltzman has been proved right in so many fields, I suspect we’ll find he’s already being proved right in software engineering.
Caution about this problem is I suspect yet another reason why the Industrial Internet of Things (IIoT) is not living up to its potential. Manufacturers do not want their shop-floor equipment updating overnight, and neither are they that interested in the constant drip-feed of marginally useful features that characterizes many consumer devices. They just want things that work, and that will stay working, and IIoT suppliers should look more to their industrial roots than to consumer-driven fads.