Heuristics for Falsifiable Predictions

Evans’ Procedure for predicting the value of new technology, and its applications to CarLabs

Sam Havens
Jul 20, 2017 · 4 min read

This essay is an attempt to turn Benedict Evans’ recent article, Not even wrong — ways to dismiss technology, into a procedure. It attempts to answer the question posed at the end of the following quote:

“It is unquestionably true that many of the most important technology advances looked like toys at first — the web, mobile phones, PCs, aircraft, cars and even hot and cold running water at one stage looked like faddish toys for the rich or the young... But it’s also unquestionably true that there were always lots of things that looked like toys and never did become anything more. So how do we tell?

My current understanding of the argument Evans makes

This is a process for predicting the future of a new technology.

  • Step 1: Imagine that the technology has gained widespread adoption.
  • Step 2: Compare the reality of step (1) to current reality. The changes will fall into various categories; the two we care about are (a) changes in human motivations and desires and (b) technological changes.
  • 2a: If the change was in human motivation and desire: stop. You are now in “not even wrong” territory — that is, you will not be able to falsify any predictions that come from this line of thought.
  • 2b: If the change was technological: is there a roadmap to get there from the present? If no, stop — you are in the same place as mobile phones were in the 1940’s. Reexamine the technology periodically.
  • Step 3: There is a roadmap from the present to the world in which this technology is ubiquitous. Great, it could happen. Does the technology matter enough that it will happen? It is easy to fall into the trap of thinking about something you can build with this technology that people might want. Don’t do that. As Evans says, “try to separate the fundamental capability that’s being proposed from the specific uses.” That is, if you are thinking, the ability to do [x] means you could make [y], focus on x, not y.
  • An Example: Edison thought the phonograph would be used for distributing sermons. It turned out that the ability to easily record and distribute sounds was very desirable, but not for the predicted reason. If someone were trying to predict how much to invest in Edison at the time, they would have been better off focusing on the capability, and not the intended application.

Lets apply this procedure to CarLabs…

Step 1 — It is a common, if not ubiquitous, experience to interact with car manufacturers, sellers, repair shops, and vehicles themselves via a conversational (or other intention-estimation) interface. Of course, this interface was built on top of the CarLabs platform.

Step 2 — What’s different between our world and Step 1 World? Well, consumers already engage in these behaviors using GUIs, so it doesn’t require a fundamental shift in consumer desires, as long as these new experiences allow them to do what they wanted to with less friction*. Messaging already is ubiquitous, so no change needed there.

It seems likely that the world in Step 1 has more voice interfaces than our world. So, we would need voice interfaces to be more useful, and probably cheaper*. Given the current adoption trends, and the muscle Amazon and Google are both putting behind them, this also seems plausible. That is, there is a technological roadmap that involves incremental changes in technology (Step 3).

In the automotive vertical, this would also require a technological change: information would need to be more accessible via API. Inventory, specials, vehicle specification, OBD2-type information, and service scheduling need to be accessible (at least to CarLabs) via API*. I can talk about why in another article, but I believe we are already moving down this path.

So this identifies where the key assumptions are:

  • Conversational experiences powered by CarLabs will allow consumers to do what they wanted to with less friction.
  • Voice interfaces will be more useful and cheaper in the future.
  • All automotive data will be available to CarLabs via API.

Analysis

I may be making the exact mistake Evans is trying to warn about. That is, the CarLabs platform is about providing the ability to build and measure conversational experiences in a specific vertical. Maybe it’s not the application to the automotive vertical that is valuable. If we really do have a procedure for creating conversational experiences in a vertical, maybe that is the thing of value.

Notes

  • A version of this article appeared in the CarLabs blog, Lab Notes.
  • I don’t know Benedict Evans, and I don’t presume to speak for him. I admire him, and I hope to one day get his feedback on this piece. I know he is less optimistic about voice as a platform, favoring augmented reality. However, I believe the two are far from mutually exclusive, and that AR’s natural interface is a combination of video and audio.

)

Sam Havens

Written by

I used to teach and study math and physics, now I do Natural Language Processing.

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade