Yesterday Facebook announced, as predicted, the launch of a range of tools to facilitate the development of “bots” on its Messenger platform. The argument being made far and wide is that bots are a replacement for apps which have become so numerous that their usefulness to users is plunging and most developers are no longer making any money.
It’s easy to see why Facebook is so interested. Unlike Apple or Google they don’t have a hardware and operating system platform with which to “own” the customer. Facebook needs to make its apps perform this function and bots give it the chance to make Facebook, and specifically Messenger, much more useful and immersive and in the process make hardware and operating systems much less significant.
But bots are only a transitional technology. The holy grail (told to me over 20 years ago by the head of Microsoft Research and still true today) is the Star Trek Computer. The film Her is the best modern take on that vision. That’s why Google (with Now), Apple (with Siri) and Microsoft (with Cortana) and even Amazon (with Alexa) have been pouring so much time and money into developing competent AI-driven assistants.
But the technology still falls short of the vision, so in meantime we will have bots – highly specific and constrained AI-driven chat bots which aim to do one thing (booking a hotel room or flight for instance) very well and reliably.
They will undoubtedly be a huge success – WeChat in China has already demonstrated that quite clearly. This post from Andreessen Horowitz has the best account I’ve found. How long the success will last rather depends on how quickly the more general AI being developed by Google, Apple et al gets good enough. Expect them to develop bot platforms of their own, but also to amp up their own investment in generalized AI. We all still really want the Star Wars computer, after all.