What’s wrong with bots is they’re not ours

In the market we don’t yet have (but will), customers will have bots of their own (among many other tools—see #2, abover), and will be able to proffer their own terms of service as well (which are already in development). Image source: this video by Capgemini.

In Chatbots were the next big thing: what happened?, Justin Lee (@justinleejw) nicely unpacks how chatbots were overhyped to begin with and continue to fail their Turing tests, especially since humans in nearly all cases would rather talk to humans than to mechanical substitutes.

There’s also a bigger and more fundamental reason why bots still aren’t a big thing: we don’t have them. If we did, they’d be our robot assistants, going out to shop for us, to get things fixed, or to do whatever.

Why didn’t we get bots of our own?

I can pinpoint the exact time and place where bots of our own failed to happen. In that moment, all conversation and development went sideways, away from developing bots of our own (hashtag: #booo), and instead toward big companies doing more than ever to deal with us robotically, mostly to sell us shit.

The time was April 2016, and the place was Facebook’s F8 conference. It was on stage there that Mark Zuckerberg introduced “messenger bots”. He began,

Note his use of the second person you. He’s speaking to audience members as individual human beings. He continued,

At this point Mark is still talking about a new communication channel that could relieve the typical pains of being a customer while also opening the floodgates of demand notifying supply when it’s ready to buy. Now here’s where it goes sideways. This was a promise of pure VRM: tools for customer to deal with vendors, at scale, across whole market. For example, to issue a service request, or to intentcast for bids on a new washing machine or a car.

Alas, Mark took this promising possibility and… well, read on:

By “you” Zuck now means developers. He continues,

See the shift there? Up until that last sentence, he seemed to be promising something for people, for customers, for you and me: a better way to deal with business. But alas, it’s just shit:

Right there the opportunity was lost. Also all that promise at the top of the hype cycle.

And the world went along with it. Because, well, there’s lots of business left on helping sellers sell. For example, read how Aaron Batalion uses the word “reach” in ‘Bot’ is the wrong name…and why people who think it’s silly are wrong, written not long after Zuck’s F8 speech: “In a micro app world, you build one experience on the Facebook platform and reach 1B people.”


What we needed then, and still need, is for reach to go the other way. There are lots of market opportunities in the numbered list here. But we could also use bot design that would let lots of developers compete to give us better ways to reach businesses—at scale for each of us.

Nothing new about that. Today lots of developers compete to give us better ways to use the standards-based tools we call browsers and email clients. We should have bots as standard and competitive as those things.

In Market intelligence that flows both ways, I describe one such approach, based on open source code, that doesn’t require locating your soul inside a giant personal data extraction business, which is Facebook’s (and damn near everybody’s) sell.

Here’s a diagram that shows how one person (me in this case) can relate to a company whose moccasins he owns:

The blue tube with the QR code on it is the moccasin’s pico, which stands for persistent compute object. Think of it as a cloud on the Net for a thing in the physical world: one that becomes a standard-issue conduit between customer and company. Picos are ways to give every thing smarts of its own in a cloud of its own, whether it has brains onboard or not.

“The unusual problems of the internet of things call for unusual solutions,” Jonathan Zittrain wrote in From Westworld to Best World for the Internet of Things, which ran in last Sunday’s NYTimes. This solution is especially unusual, because it works both for things that have no brains of their own and things that do have brains, and are too easily get conscripted without knowing it into vast “bot armies” out to do bad things in the world.

A pico of this type might come in to being when the customer assigns a QR code to the moccasins and scans it. The customer and company can then share records about the product, or notify the other party when there’s a problem, a bargain on a new pair, or whatever. It’s tabula rasa: wide open.

The current code for this is called Wrangler. It’s open source and in Github. For the curious, Phil Windley explains how picos work in Reactive Programming With Picos.

It’s not bots yet, but it’s a good place to start re-thinking and re-developing what bots should have been in the first place. The time is ripe to tstart developing there, and stop working only inside giant silos.

This began as a post by the same title blogs.harvard.edu.doc on June 7, 2018.

Author of The Intention Economy, co-author of The Cluetrain Manifesto, Fellow of CITS at UCSB, alumnus Fellow of the Berkman Klein Center at Harvard.

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store