On building AI fit for purpose

not a replica of human inefficiencies

“I wouldn’t do it like that”

ONE OF THE MOST interesting aspect of artificial intelligence and robotics is that we will develop solutions that perfectly fit the problem space. And we may not only surpass the limitations of the human mind and body, but extend the senses as well.

Almost everything in the world today has been designed presuming a human will interact with it.

How would we set a broken arm differently if doctors had three arms and hands?

What would student classrooms look like if we had telescopic vision and supersonic hearing?

There is a natural tempo to people’s movements. What if our muscles contracted at a pace around ten times faster? Or slower? What would that mean for the expectation of what we could achieve in certain amount of time?

If it was slower, there might be types of manual work like surgeries that would be impossible to perform. If it were faster, other assumptions would change. (No one would complain about doing the dishes after a dinner party…)

Now, what of our brains? If we could remember 1,000 times more material on average, what do we do today that would change? And if we could only remember a 1000th as much? Could we drive cars or fly planes — or would we be constantly forgetting to look in the side-mirrors when changing lanes?

Next is mental speed. (Not just the size of storage, but the power of the processor.)

What if we could think 10,000 times faster? Or 10,000 times slower?

If faster, we wouldn’t just think of the same things more quickly; we’d think more too. That one feature — more — happens to be critical. It opens up the realm of very exciting new possibilities. To get a sense of why and how, imagine if we thought 10,000 times slower. Could we do much of what we do today? Our early ancestors be unable to hunt. We’d be vegetarians. Instead of domesticating animals, we’d be proud of our civilization’s accomplishment of domesticating carrots.

That’s what’s so thrilling about artificial intelligence and robotics. We will be able to have devices and computational reasoning that can vastly surpass people’s minds and body movements. We will be able to do new things.

Our solutions will be geared to what’s needed, not wed to the human form that used to be sole entity capable of handling it, be it making a decision or physically acting in the world.

But getting there won’t be easy. And humans will revolt at the idea that their form and function — albeit accidents of evolution — are being overtaken.

As in the image above: many minds working together. Yet justice is broken, peace is desecrated. And look out for that innocent kid…