Tomorrow I’m speaking at PSFK’s CXI2018 conference on innovation In the new consumer experience in New York City. The title for my talk is “What people want from tech”, a topic I recently discussed with PSFK founder Piers Fawkes:
One of the frontiers that I’m excited about is the opposite of where everybody’s going. Less tech, smarter people, better fitting components, higher quality design, quieter products…
I don’t want an app that “tells me by taking my geolocation at all times when to pick up the milk and butter or whatever, and eggs.” I know how to do that. I don’t need a $2 million venture-backed company that’s going to leave my fridge stupid after it fails to tell me these things.
We need to quiet down our machines. They need to break less. We are in the most disposable period of time. It’s not just environmental. It’s that people will not have the capacity to purchase all of these things continuously again and again in the future. These components will become very expensive, too.
What we really want from technology, in other words, is less technology that’s cutting edge for its own sake, and more long term tech that integrates more seamlessly and unobtrusively with our lives. I discussed this theme from another angle in a March PSFK interview with Piers, pointing out the limitations of artificial intelligence:
The best stuff handles 70% on Google Search and then gives us back 30%. We choose from that. A bot that makes a choice for us can fail disastrously. I should be able to text back the response, and the bot should be able to get it or store a variety of responses.
AI is a fake term. There’s narrow AI, there’s some general AI. There’s not any superhuman AI out there yet. It’s a forgotten term, cybernetic interaction or feedback loop. Films have given us expectations of perfect machines. We forget those machines were scripted. None of it is real.
Consider Google Duplex, which many see as a major breakthrough in AI. But Google’s own write-up of its technology notes that “Duplex can only carry out natural conversations after being deeply trained in such domains. It cannot carry out general conversations.”
In other words, Duplex is a good example of applied Narrow AI, which doesn’t try to do everything — just things with very few variables: Location, date and time, number in party, and so on. But each domain is very different. There are a whole universe of differences between booking a hair appointment at a salon, versus booking a table at a restaurant. While Google Duplex is an impressive step forward, it’s impressive because it’s restricted to a narrow domain.
This is where we’re really at with AI technology: We’ve been promised too much from Artificial Intelligence for a very long time. So when companies under-promise by restricting AI to a narrow domain — and in Duplex’s case, making it seamless and unobtrusively useful — they’re more likely to be successful.