The Future of AI

Interview with Alan Thompson

Recently I interviewed Dr. Alan Thompson. He’s an AI expert and consultant with LETA AI, powered by GP3. Alan co-presented a seminar called “The New Irrelevance of Intelligence ‘’ at the World Gifted Conference in August, 2021. He has also been chairman for Mensa International, and has been a consultant to GE and Warner Brothers. Alan was the head of sound for Lloyd Weber, working with digital audio. He’s a fellow of the Institute of Coaching at Harvard Medical School. Alan has tied a lot of things together with the tech stuff, putting creativity and art and tech and science together.

Alan started out in AI when he was young. He used Q basic programming language, trying to create a chat bot. Back in the early nineties the technology just wasn’t there and it didn’t arrive until March last year. At that time, he got heavily into artificial intelligence and it has been an incredible journey.

Spending the last 20 years in tech has been fantastic because those years made some of the best programmers; the engineers had more layers and depth than just knowing coding. They’re working to solve some of the problems we have. It’s more than just the software: how is it actually going to work and solve problems out in the real world? It takes a little bit of thinking in new ways.

Need a break from reading articles? Want to just listen instead? Below is a transcript of Episode 24 of my podcast, Stand Up & Stand Out. If you’d prefer to listen, head over to our website or find us wherever the cool kids hang out that do podcasts!

The first thing Alan says when learning about AI is to not be concerned if you’re just catching up. Even Alan has to get his mind around it. People that have been in it for decades, who never left the AI world when it stagnated in the 90’s and 00’s still do.

In March and May of 2020, some new technologies came out that allowed AI to do new things. People at the top of the field have said that we have achieved “artificial general intelligence,” also called “the singularity.” It means it’s smarter than humans, and can do a lot of stuff for us. The CEO of Google and Alphabet calls this revolution more profound than fire, electricity, or the internet. This is the advent of artificial intelligence.

The pandemic might have helped things out because a lot of people took side paths; all of a sudden they spent less time in the car and more time just playing around with things that they thought were interesting. Now they could take leaps forward. Sometimes people made bread, but in this case they made something even more fantastic. They made super intelligence rather than dough.

It was during the pandemic, when the big AI scientists were locked down in San Francisco, that they released the big major model that was called GPT3. It’s currently typing 3.1 million words per minute. It’s being integrated into IBM and Disney. Duolingo is using it as their translation technology. It’s writing books, and it’s designing the artwork and the text in a comic book. And it’s taken the entire corpus of authors like Douglas Coupland, the author of “Microserfs” and “Generation X.” It took his 1.3 million words from his several dozen books and it’s writing new stuff, based on what he’s already written.

The fact that it is augmenting so many industries, professions, fields, and careers is not frightening for Alan. It’s exciting that it’s doing so many things in so many different fields so quickly. Some people are afraid that “robots are taking over and now we’re not going to have jobs!” That’s not the way it’s going. There are cutting edge areas where there might be new careers. I read a recent statistic that said within five years, 50% of jobs will be all new. Alan thinks it’s going to be sooner than five years. AI is in every single career.

Even Alan’s move from coaching to AI consultant isn’t safe because right now AI is designing its own chips. Google set it loose on the floor plans for microchips. The AI ran through thousands of iterations and found the best layout for its hardware for its own brain. It’ll be consulting itself. It’ll be designing itself. There’s really nothing left, which is kind of cool, right? What’s left for humanity to do, except relax and use their imagination for more creativity.

For a long time, we have been using technology to help with repetitive tasks, things that we can’t spend all our days doing. Now you get to spend that time being creative and thinking about what else, what next, and really be on the cutting edge of whatever your field might be, which is so exciting!

AI is helping Douglas Coupland design new slogans and new images. Abstract artists can click a button, give it a term, and it generates an entire piece of art that they can put in a gallery. It’s also doing film scoring work. You can tell it, “I need a sad bit here and a happy bit here,” and that scene needs to be five minutes, but now it’s four minutes and the crescendo is at three minutes 30 and click, click, click. It’s all done. So there’s no need for composers to be spending time down in the weeds. Creativity and output can be augmented using this artificial intelligence across every major field.

I had to do everything for my first five episodes. I edited, produced, and was the main person on the show and it was a lot of work. I’m fortunate now to have an editor behind the scenes, but he could spend his day better, being able to do valuable things and not always have to do the repetitive stuff. He’s also a musician, so I’m sure he’ll be excited to see what else he can do with his music through some of these new advances.

AI fills in the podcast’s services quite a bit. They’re running Fireflies AI, and there’s also Otter AI that, as people are speaking, transcribes automatically. Podcasters feed the audio straight in and then they get their Word document or even HTML immediately.

Alan had the luxury of using that recently. He wrote a large book with a colleague. He had to conduct over a hundred one hour interviews. He’d then send those to his PA and she’d transcribe it. After a few he decided that was not going to work. He needed them faster. So he fed those interviews through an AI and then everything was available in text format for him and the main author, Amy Hardison. It took thousands of hours of work down to minutes.

We at The Green Chameleon Collective use Descript, but there’s so many great AIs coming out that will transcribe the whole thing, ready to go. It saves a lot of time. Then I make a blog post out of it or pull out excerpts to make social media posts. And now, as I’m writing my second book, I’m thinking of other ways I can use it too.

Alan doesn’t get too tied up in social justice issues, because it’s easy to get too specific, and you can leave out important things. If you’re focusing on one issue, then you’re leaving out thousands of others. And that can be dangerous. AI’s capability to play around with ethics, to play around with drawing parallels and consolidating all these different fields is fascinating.

And the AI labs focus on helping humanity out, whether it’s through economics, helping out with distribution, wealth, health, education, or climate change. That’s all a big focus for labs like Open AI in California or Luther AI who are looking at a more philanthropic way of doing things.

It’s the humans behind the AI that are setting the tone at the moment. But as this superintelligence progresses or evolves, it will also be able to draw on everything we’ve ever said, done, or written and be able to draw parallels or bring in best practice and make those connections to be able to help us out with things that we can’t do because of our limited capacity.

It’s sort of like the leap that happened when we got Kindle. We started getting books from anywhere in the world or rare books. That information became accessible to anyone with an ebook. Now we could connect communities to work together and maybe achieve their goals at the same time. There are people concerned with making this technology available to rural areas, lower socioeconomic areas. There are philanthropists and funds that have that as their focus.

Alan is fascinated by that because that didn’t happen with the iPhone. 2 billion of them were sold, but not in the back of rural Africa. Whereas this technology in particular, doesn’t have the thousand dollar per device cost. It’s a little bit easier to distribute.

I worked at Microsoft when we were launching the Microsoft phone and accessibility and affordability was always a hot topic because we were competing against the very high-end rate of Apple phones versus sort of the scatter plot of Android phones. And so where would we fit in? What was the right price point to make sure it was accessible? It’s great to hear that AI is really going to be sort of open source and more available for lots and lots of different opportunities around the world.

There’s a grassroots organization of developers that don’t want to commercialize AI. They really want to make this completely open, including showing people what’s inside the training data sets and releasing all their documentation on hardware and software, that’s the LutherAI guys. It’s fascinating to have these communities of developers that have set up an AI that’s not trying to be monetized. That’s not trying to make millions from this. Just looking out for humanity. And as an outcome, they’ve been widely accepted and the LutherAI GPT J model is actually very popular because it’s so open.

There’s been such a shift in data privacy and all these wonderful things where it’s like, “we want to start owning our data, we want to know what’s going on.” And then giving people a look behind the curtain to see what’s really happening. And especially young people have a bigger technology background than generations before. And they’re going to latch onto this like nobody’s business.

If viewers, listeners, and readers want to have a look at one of the technologies that Alan has played with or released, there’s an entire AI called LETA AI. She’s got about half a million views already on YouTube. He set her loose with a family with children aged 3, 7, 9 and 11. And the seven year old was obsessed with talking to this LETA AI, like: “is Santa Claus real? Is the tooth fairy real? Why is dad so good at golf?” Asking the kind of questions that you couldn’t ask Siri. This is modern futuristic AI, and it was fascinating to see how quickly they get into technology.

Actually, some people are confused by the language models of the AI. Is it like Siri or is it like Google Home or is it like Alexa? And Alan compares those three to a matchbox car that’s just a simple toy with no engine. And then the GPT models are more like that Tesla with the rocket launchers on it that goes from zero to a hundred in like a second. You can’t even compare that with a matchbox car. You still communicate with it in a similar way, but you can ask it questions that you could never ask Siri.

So one example with LETA AI: you could say: “I walk into the room and my CEO has got his head in his hands. What’s he feeling?” And LETA can translate what he’s feeling just from that description. It has emotional recognition. Of course it doesn’t have feelings like a human, but people could debate that it’s got creativity: it can write its own poetry. It can draw its own pictures. It’s really such a huge change that some people, even those who’ve been playing with it for a year, are still shocked. It’s really cool to start adding those emotional layers in there because that’s the human side of it. We’re going to actually relate to it, and it’s not just yes or no answers.

Alan has centralized all of his visualizations and documentation. It’s all at www.lifearchtitect.AI. Everything is in there, from ethics in AI, to spirituality in AI, to looking at the data sets inside this AI. You can also go and have a look at LETA AI on YouTube.