That’s a weird claim but let me break down my thinking for you.
First, the pendant. It might not be a pendant, it might be a watch or a pair of spectacles. But it’s a device with a microphone and some form of cameras that has a connection to your phone. It acts as the eyes and ears for your AI app. It will start to feed your AI app with a river of real time data from your daily life. The things and people you are looking at, your habits, where and how you spend your time, your conversations and so on. And you will be able to talk to it about things you are doing or are involved with in real time, in almost any environment.
We know that big tech are already working on wearables like this. For example, ex Apple design Jony Ive is working on something like this for ChatGPT.
The tech will need to find ways to massively compress the data it is collecting. But as with most things the Pareto principle will hold: most of the time nothing interesting is happening and the best most informative stuff will happen during a minority of the time. Also, of course, issues like privacy (both yours and others) could stop people adopting this tech, but assuming they do…
Why will you want to feed this personal information to an AI app? It’s because the more in-context information it has about you, the more useful it can be to you. People are already getting value from feeding Claude and ChatGPT things like their heart activity data from their smartwatches. But at the moment most people are limited in the personal information they can give an AI app. Mostly just whatever is on their computer or phone.
You might think you can just tell an AI about yourself. You can write out a description or even upload your journal. And to some extent that can work. But the problem is that we are limited in our self knowledge. We just aren’t aware of all our tendencies, habits, and behaviours. And when we are we might over or under exaggerate them. What you think is important, but your own thinking might be the thing that is imprisoning you inside your problems. And an objective look from the outside might be what’s needed in order to solve them.
Now the level of personalisation that this could enable is what makes it a game changer. Having real, objective and thorough data of your daily life is the difference between having a description of a crime and an actual security camera recording of it.
Since the Industrial Revolution we’ve been focused on making technologies that can scale, rather than making technologies that are particular to the individual person. Yes we’ve had personal computers, personal smartphones and personal games devices but they have all still run generic, universal software. Not software that is designed specifically for us.
But AI is smart enough to adapt itself to us, given enough relevant information. And the next generation of wearable tech will add that ongoing information. All that’s needed is a way to make sense of it.
And this the second part of my prediction: we’ll develop software that has a logic that bends itself to the unique circumstances of the individual. Not all problems yield to the same cookie-cutter solutions. Usually if you solve a problem in your life, the solution is very idiosyncratic to you. This is a very different approach from, say, receiving generalised advice from a self-help book or YouTube video.
Aristotle identified three types of intelligence: (1) Episteme, which are the universal laws and truths that science aims to discover (2) Techne, the knowledge of how to make things, and (3) Phronesis which is more like practical knowledge of what to do in particular unique situations. In the modern world we don’t often think about that third type as a unique form of intelligence. Its sort of like street smarts. But it’s more broadly the ability to devise new solutions to very particular problems.
But this is what we could soon enhance our own abilities in using AI.
What works or doesn’t work in your life tends to depend deeply on your personality and the circumstances of your life and environment. And requires thinking that can take those into account.
I see no reason why AI can’t develop this kind of intelligence. This is a new class of thing. Somewhere in between a therapist and a scientist who has reverse engineered your life.
The therapist is actually a good model for this. Except it’s a therapist that can objectively crunch a lot of numbers: taking into account a vast amount of information about the patient’s life.
One therapeutic framework that AI might find useful is that of Solutions Focused Therapy and related theories. These focus on the individual and look for solutions within the patterns of the individuals’ own life. They don’t assume that the problem itself is the focus (hence the name ‘solutions focused’). Instead they focus on instances of when the problem doesn’t occur. What is it about those moments? They can hold the key to the solution.
So the more an AI knows about the individual circumstances and life patterns, the more likely it will be able to spot those moments and what they have in common.
For most of history, practical wisdom was the one intelligence you couldn’t get from a book. It had to be grown slowly, out of your own life. A device that watches that life closely enough could grow it for you.