The Power of AI and what the Future holds

I am by no means an AI expert. And though recently someone referred to me as "entrepreneur and Product guy masquerading as an architect" ( true quote in a recent mutual intro email ), I really love to tinker to try to understand how things work and the impact of new innovations - especially in the most profound of ways.

This week, I've attended TED AI in San Francisco and over the course of two days, I've absorbed a ton of food for thought - thoughts that have changed my perception and understanding to date. Things that struck me as powerful, while simply said. I'd like to share this with you, and hope those that have not had a chance to attend the event, can appreciate where we are now.

Here's what I've learned.

CX - Computational Language and Model of the World

Ok, let me be frank. My love for math disappeared some time in college and I no longer have the pleasure or the patience to explore, so given the phrase of "Computational Language" as a Mathematical Applications in Computer Science - went right through me. Until I heard these words:

"Language encodes the model of the world!" Walter DE BROUWER

The words we use are not about communication. They are about the model of the world. Communication is just a small part of it. It describes our interaction, but other species are able to do this in just a few sounds. We, as humans, are much more sophisticated.

To appreciate our very own intelligence, as Walter De Brouwer described think of the words we use. Words like "gravity", "equilibrium", even the meaning of the word "sophisticated". These words describes the model of the world. And, we evolve that model. "Google" has a meaning that describes a function of search in the world. So does "Uber", "AirBnB" and others. No longer names, but models of large functions. And we continue to evolve our model of the world, inventing new names, adapting what we say and how we say it to draw new meanings.

What was achieved with AI is the ability for Computers to learn from the language model. Similarly to that of a young child, learning each word and meaning with some words, even us parents struggle to explain. How do you explain the word "understand" to a child?

So if we learn from the world now, is it feasible that our children may learn from the AI model? Neo in the Matrix programmed to learn martial arts in metaverse model no longer sounds like fiction. Ability to train the brain with imagery, like Number 5 in Short Circuit hungrily saying "input".

"Imagination is more important than knowledge. For knowledge is limited to all we now know and understand, while imagination embraces the entire world, and all there ever will be to know and understand." Albert Einstein

The Impact on Medicine

I've long been telling a joke "When I became a software developer, I started to get scared to go to the doctor." Oh, boy did I create some bugs in my younger days. All part of the professional experience of course, which taught me many things, but joke does stay. In engineering, we have the luxury to make mistakes and fix them. In medicine, mistakes can be quite fatal.

So, when it came to applying AI in medicine, I always felt that we wanted to apply AI in a way that gives us the right answer. A guaranteed correct precision, which is completely wrong. "I" in AI stands for intelligence and intelligence means learning from mistakes - an artificial brain in the computer applies intelligence given the many data points available to its disposal. What do we know?

We know some things are easy for AI to figure out with high reliability. We've also seen the way AI hallucinates. We know mistakes are not acceptable, and supervision is definitely needed... until eventually we reach the moment we as humans no longer possess the level of intelligence to quickly verify. What happens then?

Good questions, for there are no clear answers. At least not yet, but this is where Eric Topol, MD took the stage. He shared how there are over 12 million medical errors every year and how the diagnosis accuracy quickly drops the longer time it takes after the erred determination. And then, AI stepped in to show how Diabetes can be detected from X-ray, how Heart Disease can be detected from a Retina Scan, and how Kidney Disease can be detected from an EKG.

If all of this is done today. What can be done tomorrow?

Reflections

AI is in all caps for a reason. It is big, it has huge impact, we're on a cusp of things so profound and yet so real.

As I was left pondering about many of said words, I share deep phrases that stood out.

"AI is the new kind of camera" - as it captures the world

"Hallucination is the essence of creativity." - regarding LLMs hallucinating as if its bad. We human do it all the time in sleep, in daydream, and in wonder.

and lastly

"Song is a conduit of emotion" - Oak Felder