Edge of AI

Deep Conversations About AI And Technology With Walter De Brouwer From Snowcrash

Edge Of AI Podcast | Snowcrash

 

Buckle up, passengers! In this thrilling episode of Edge of AI, we embark on a journey into the depths of the rapidly expanding AI universe. Our voyage is guided by none other than Walter de Brouwer of Snowcrash, a renowned figure in AI and deep tech. Walter shares insights into novel AI use cases, revealing how video games can prepare us for a more gamified life and discussing the explosive potential for balance sheet growth in the music industry. Get ready to learn from Walter’s entrepreneurial journey, which includes founding multiple AI and deep tech companies and leading groundbreaking projects at Stanford University. Discover how large language models are evolving into colleagues and how the impact of AI on daily lives is anticipated to make a significant leap in 2024. Walter sheds light on the practical utility of AI, especially in gaming. Get a front-row seat to Walter’s vision of the AI landscape and gain valuable quotes that encapsulate the essence of his wisdom. Join us as we navigate uncharted territories, fueled by curiosity and guided by the ever-expanding universe of AI.

 

Key Takeaways:

  • LLMs serve as colleagues, assisting with questions and providing insights.
  • Despite progress, LLMs still have challenges in reasoning.
  • AI’s impact on daily lives is expected to make a significant leap in 2024.
  • The practical utility of AI is becoming more apparent in daily lives.
  • Gaming is an important step in the development of AI, translating utility into real-world applications.
  • Google’s new release is a potential competition for OpenAI’s models.
  • The gaming industry is a significant partner for AI, with potential applications in creating non-player characters and game tutors.
  • AI could evolve into a life tutor, providing real-time information and advice throughout various activities in a day, acting as an ally.

 

Quote:

  • Before ChatGPT, we had many theories of the world, but all that needs to be updated.
  • We should always look for adventure where we can learn something. Even if it doesn’t work, we would have learned something.
  • You have to find people who say yes.
  • In the end, it’s not about setting up companies. It’s about the adventures.
  • I always thank my large language models and my ChatGPT because, one day, they will be in charge and then there are neural nets. They keep score.
  • Everything you need to know is on X.

Listen to the podcast here

 

Deep Conversations About AI And Technology With Walter De Brouwer From Snowcrash

I’m Walter de Brouwer, the Cofounder of Snowcrash and TED AI. I’m obsessed with deep learning models and the future of media and entertainment. I’m on the Edge of AI obsessed with blending AI creativity with culture. Stay tuned.

Passengers, jump on in. Here’s what’s to come on this journey. Find out a novel use case for AI and video games that our guest predicts will prepare all of us for more gamified life and learn how to explode your balance sheet if you’re in the music industry. Lastly, what are the five go-to resources a Stanford professor uses to stay on top of what’s happening in the world?

All this and more, take your seat. Welcome aboard the Edge of AI. Snap into your safety belt and prepare to explore the depths of the rapidly expanding AI universe. Each episode is a dispatch, featuring hyper-relevant reports from the pilots, pioneers, and passengers aboard the AI rocket ship. We explore the latest use cases in developments in AI, hear from experts building the tech, empower you to enhance your life with AI, and learn how this disruptive force is transforming industries and society.

I am your co-captain for this voyage to the Edge of AI. Just like most of you, I’ve embraced the spirit of exploration and entrepreneurship throughout my life from starting my own business before graduating high school and traversing the world’s most challenging terrains. I’ve always sought out new frontiers and adventures. I built one of the largest award-winning custom home companies in LA. Also, I’ve navigated complex regulations while founding and leading a public company that is dedicated to applying technology and training.

I’m your co-captain. I have an insatiable curiosity for disruptive ideas and technology which led me in a cross-industry entrepreneurial journey building transformative companies. As Cofounder of Edge Of Company, I’ve hosted over 300 conversations with emerging tech leaders and artificial intelligence has been part of my toolkit for quite a long time.

I was Cofounder of one of the largest food tech companies in the US where I architected the menu planning algorithm based on consumer taste. Before all of this, my roots in consulting included supporting geospatial visualization services across the Federal government and the predictive analytics initiative to curb veteran homelessness.

Buckle up and get ready. Let’s tackle uncharted territories in AI with curiosity as our guiding star. You’re going to love this episode. It features Walter de Brouwer also sometimes known as WDB. He’s a renowned figure in AI and deep tech domains. Walter’s entrepreneurial journey includes founding several AI and deep tech companies since 1990. With two public offerings in Europe, one on the LSE and Euronext, and two in the US. One was on The New York Stock Exchange and the other on NASDAQ.

His venture in AI merged with Sharecare, which is SHCR on the NASDAQ. They went public with evaluation of $3.9 billion in July of ’21. Post the $571 million merger with Falcon Capital Acquisition at Sharecare, he still serves as the Chief Scientific Officer. He holds a PhD in Computational Semiotics. Walter has been a professor at Stanford University since 2018. He’s an active contributor to a generative AI projects, including LLMs and agent architectures.

He’s authored one of the first reviews of the GPT-3 model. His research spans multimodal transformers, AGI, GPT Agents and DOCOMO’s 6G. He also holds 53 patents. At Stanford, Walter leads Med 205 at the medical school and has significantly contributed to the field of decentralized clinical trials. He chaired the IEEE’s decentralized clinical trials committee and authored a seminal paper in nature, which guided the FDA’s approach to decentralized clinical trials.

Walter was honored with the AI Visionary Award in April of ’23 for the contributions to the term promptography. Alongside his wife, Sam, he produces two annual TED conferences, TED AI at San Francisco Opera and TED in Hollywood focusing on media and entertainment. We’re going to jump right into the conversation here. There are so many directions to go, but I guess we’ll just start. Josh, I’ll let you kick it off a little bit.

Walter, it’s good to see you as always, my friend. Congrats on TED AI. It was a big thing to put that together and such a well-timed event in terms of everything going on in the space. How did it shape up from your perspective and maybe you can tell us a little bit about some of the highlights?

What I learned from the first event was in order to produce a good TED event, you need the top speakers who are then also the people who work on these models at a very high level. You also need in the audience, your top people, so that they can talk to each other. We had so many demand for speaking. We are going to program a little bit less so that people can have more talks. It was all very well and everyone was happy, but I think giving a little bit more time for the speakers would be a good idea.

There’s a long list in our arm to talk about different things that AI handles. However, as you plan the second event, are there any themes or subjects within AI that are prominent for you that you want to make sure the second event covers based on where we are in the development of AI?

We are now going into a second phase and it is always difficult. As an industry, you can come compare it to the second year of a startup. In the second year of a startup, what you do is, all year, you correct the mistakes you made in the first year. The industry will also do that. Some things were hyped or over hyped and other things were under hyped, but now it will come to it’s no longer just a demo. It’s now going to be the execution.


Edge Of AI Podcast | Snowcrash
Snowcrash: Some things were hyped or over hyped and other things were under hyped, but now it will come to it’s no longer just a demo. It’s now going to be the execution.

 

What is more important, and this is what we also saw in the first edition, is I call it the boundary point, like in mathematics if you have a set. The frontier is now everywhere. Before ChatGPT, we had many theories of the world, but all that needs to be updated. For instance, we had the Assembly Theory by Lee Cronin. He’s a chemist at the University of Glasgow. He had a complete new theory of how living things and non-living things measure their complexity.

Stephen Wolfram just came out with Observer Theory. There are many things in science now that because we have ChatGPT, which has a perfect memory, but we can say, “What was that again? What does this remind me of?” For instance, in quantum computation when the wave collapses, is that not the same with what happens? Otherwise, we would spend a year thinking about this, doing research, googling it, and talking to colleagues. Now, we can do that in basically a couple of weeks. It is the innovation in science itself but also in material science. For instance, genome by DeepMind.

We came out with two million new crystalline structures for materials. That would have taken us probably 100 years to find that because we would basically try out several combinations then we would see what would work or not. We then try to realize it in the lab. What they did is DeepMind made the model. They connected it with the A-Lab, which they called it like that in Berkeley, with a robot. The robot tried things out. In seventeen days, they took from 80 tests and they had a 70% success that these crystalline structures can work in reality. It’s amazing that is going to happen.

What you captured there is this massive movement of progress that’s happening now exponentially that compelled you to do TED AI and observed some of these things happening in the space. I’m curious, at this moment in time, what your observations are on the impact of LLMs on our daily lives.

I have several LLMs, which I almost look at colleagues that help me. Claude said, “What do you think about that?” I go back to Claude and I said, “You made a mistake there.” It’s very strange because if you tell that to people who are not in the field, they’re probably think you’re taking mushrooms or so. This is happening now, but it still has problems in reasoning.

Kevin Kelly also mentioned it before. Elon Musk said, “Life is but a dream,” after Karpathy said that, and it’s true. Everything we do is basically dreaming. We dream all things in our head. 30% to 50% of the world have an inner voice that constantly talks in our head and says, “Don’t say that,” or paraphrase, “Don’t mention this.” These are our own guardrails. It took us hundreds or thousands of years to see that saying stupid things has no survival value.

Now, we are getting into a very new territory with non-biological intelligences. These AIs will certainly come to AGI. There are a couple of problems. Judea Pearl in The Book of Why said it’s causal inference, and therefore, we can solve that with do-calculus. When we say to the machine that when somebody has an umbrella and when it rains, he will open that umbrella, the machine understands that. What the machine doesn’t understand is that when somebody opens the umbrella that it doesn’t rain. It’s because we cannot cause rain to happen. This is working model. Just like children, these neural nets are learning because they are only three years old.

What you’ve touched on is a practical utility of AI right there. We went from the scientific model and medical model you described, which is pretty massive, but we have practical utility in daily lives. If you had a look at in the next few months, do you think that we’ll be focusing more on the practical utility AI or its development itself? Are we there yet in the place to cross the bridge from the development into literally talking about practical uses?

I think in 2024, we will probably make a jump of twenty years in one year. At the end of ’24, we should now take a video of ourselves talking to our family and then look at it at the end of 2024 because everything will have changed. Our prices will have changed, the way we consume, especially the way we do entertainment because what we are now is still missing and, for me, it’s coming in the next 3 to 6 months. We will probably more and more going into the AGI.

In 2024, we will probably make a jump of twenty years in one year. Share on X

There is this concept of P-zombies or Philosophical zombies. A philosophical zombie is a person that is like us but doesn’t have any consciousness. He’s a zombie. These neural nets are computational zombies. They have consciousness. It’s just completely different than ours. They have a perfect memory and they have patterns. Very soon, they will learn from us like on how our feelings work by different modalities that we are putting in. For instance, an article came out that something that we had suspected and I think it’s an interesting idea, but ChatGPT doesn’t work so well in the weekend. It mimics us. It rests.

ChatGPT doesn't work so well in the weekend. It mimics us. It rests. Share on X

That’s an interesting way of putting it.

It’s the same thing that, at some point, Sam Altman had the perfect answer to what somebody said. “I think ChatGPT is getting more stupid.” He said, “No. You are getting smarter.”

Speaking of the interesting competition among LLMs and ChatGPT, I’ve heard some positive feedback about Google’s new release and the potential that there’s now real competition with OpenAI that isn’t necessarily been there before. This reshapes the overall landscape. What are your thoughts there?

It’s all about guardrails. Even as humans, we have a million years of things we should do and we should not do. Sometimes, our guardrails don’t work. Our university campuses were suddenly full of activism. You always have to put up these guardrails again and reconsider our values. What people don’t yet understand is that when OpenAI put something public, it is a product that is tested and there are guardrails. These guardrails are improving.

If it’s not given public access, it’s a demo. I’m sure Google will get it right. My view is that they will come from a different angle and this might be the thing we have been waiting for because we still don’t have any heuristic search or heuristics. There is a lot of talk about A star and Q. These Q-tables and A star are things we are using in games to do missions.

I know it because I’m interested in the physics engines of games because we need that in world models of language. It is just like why a large language model doesn’t understand that we cannot cause rain. Sometimes in GTA 5, not in GTA 6, when you run over somebody, he gets back up. That’s not reality. It should lie there bleeding and it should call an ambulance.

One of the theories you’re going by is gaming is a very important step in the development of all this because by developing gaming, it will take a lot of that utility and put it into real world events and individual people’s events. That’s a big takeaway. Also, we interviewed someone here that was jailbreaking ChatGPT. He’s a white hat hacker and that’s his sauce. You can see based on talking to you with your incredible expertise the importance of that. They’re helping the whole ecosystem by doing that so those weaknesses can be taken care of now before some things might be too late.

What we do in games is we learn the machine and the machine teaches us because that’s how it works. We train a device and then the device trains us. We learn how to do these missions. We learn to take tasks to put them into little buckets then we do that. It’s like a to-do list. Once we have the to-do list, then we go for it.

Let’s see your to-do list real quick. Can you talk about what project you’re working on now within the development beyond the scientific study of it? What projects are you working on that you can speak to?

I’m working on a couple of projects. I’m writing a book at the moment where I’ve find that all the theories from before ChatGPT, we have to update. I am a computational linguist, so that means that I’m a linguist but I’m also in computability. I have to look at the world through the lens of language. Language is not just natural languages. It’s also artificial languages like programming languages. It’s also ecolinguistic languages. I studied Klingon, for instance, to find out how that language could be used and it was used to translate Shakespeare with only 3,000 words while we are using 30,000 words.

I can see all that. Our readers here come from all backgrounds. They’re deep in AI, they’re AI-curious, or anywhere in between now. The magic of you is you framed out the research that’s happening behind it to help it develop with some eyes on which direction it may go or could go and not restrictive by any stretch. What development now tied to a tangible project we can play with a few months from now? Can you give us a nugget for that?

It’s a very difficult thing, but I think we’ll get there. We’re pretty close. It’s the ChatGPT of music. It’s basically large music models. I don’t mean mimicking but generation. What is the power of Taylor Swift? She has a big team that all does the formatting and oversongs. It’s not because you have a big team that you are going to be successful. You also have to be talented, but she’s also a very good business person and a very good strategic person.

All this is producing. All this is music. If you have one song and you can say, “This is my song, from now on, I want several formats of that song. When I go on tour in Michigan, it must be another song. It must be the same song but another song. I want to translate it in several languages so we can go to all the markets. I want you to take extracts of that song and offer it to Hollywood movies to be put in a movie.” This is all formatting. This is what generative AI is best at, generating music and the synchronization of voice with that music and the lyrics.


Edge Of AI Podcast | Snowcrash
Snowcrash: Generative AI is best at generating music and the synchronization of voice with that music and the lyrics.

 

You’re less about the actual creation from the start. It’s a supporting tool for a musician to be able to save a lot of time to the AI taking the musician’s creation and doing more with it. Is that accurate?

Yes. Also, don’t forget that the big publishers like Sony, Universal, and Warner own billions of songs. That would mean that from these latent value, you can create new versions. Imagine that it will explode their balance sheets.

A close partner of music these days is the gaming industry. A lot of music appears in games and gaming is exploding. I read a report from Game7 about Web3 gaming and thinking about all the drive power that’s been deployed and we haven’t even seen the results yet. It’s an interesting industry when you parlay it with how AI can be utilized in gaming instead of accelerate the growth of the industry. What do you see happening at the intersection to get you excited?

Everyone talks about NPCs or Non-Player Characters. That’s an easy one, but I would like to have my game buddy with superpowers like an agent. Do we have cheat codes for this one? Can you look for cheat codes? How do I get out of there? Show me and navigate me through in a computational way the shortest path out of this trouble I got myself in that game. It’s because I want to reach level eleven. I know it’s impossible, but try to the shortest path into that irreducible complexity so that I beat that game.

That makes a lot of sense because some of these hard games are great for the diehard gamer, but they’re limiting their audience of folks that want to get to know those games in having that tutor along with you. That’s more than a tutor. It’s your ally in getting to the end of the game. You still feel like you accomplish something, but you had a friend helping you figure it out along the way.

I’ll give you an example. At some point, you will go in a game. You do some evil things. That’s why we have games so that we don’t do them in the real world. You want to get out of a place and there is a sniper. There are 2 snipers and 2 snipers who has bonus on your head. You can tell your agent, “Can you talk to these guys? I offer them so much if they stop shooting, or, if they go with me, I’ll split the spoils of the robbery. Talk to them and then, together, if we can then because this guys are good sniper. I want them to be part of my team.” All these things is amazing that you could do that and then you’re really into a game because now it’s still passive.

It dovetails into a life tutor. As you said, the things we developed for games we end up being able to use in other applications. You described a life tutor that whatever you’re doing during the course of a day an hour or a year, you can have this life tutor that’s drawing real-time information from the greater world and advising or suggesting. It’s massive.

Very soon, we will have game guilds. They will be the new Masons. They are building a world there. I have my team there. I have my sniper. I have that. My team is worth a lot. Perhaps my team can do a mission for somebody else or I can sell a part of my team and get more money and buy another team. It’s amazing.


Edge Of AI Podcast | Snowcrash
Snowcrash: Very soon, we will have game guilds. They will be the new Masons. They are building a world there.

 

Let’s touch on some of the other projects that you’re working on. I know you’re doing some things with your students at Stanford for example around agents. How’s that going?

For the moment, we’re looking at very technical things but interesting things. It’s not part of the course because it would be too technical for the course. As I said, I’m working on these three things. If you want to join me, come and we’ll do it together. We have a beautiful campus. You can you can sit around. For instance, mechanistic interpretability. Anthropic is working on that. I found it interesting because it’s so complex.

We cannot understand neural nets because the neurons are in a tangle or they are in superposition. They do several things. We cannot say, “This neuron does that.” Anthropic is finding ways to get rid of that superposition to feature everything that neuron can do so we can understand perhaps neural nets better. Although, I’m very skeptical because we don’t even understand ourselves. We don’t have a perfect memory but it’s interesting to work on these things because it might give us a way because sometimes the machine finds things that we don’t find. Why?

It’s because, unlike Elon Musk, we are not thinking from first principles. Everything in the real world needs a substrate. It’s either atoms, molecules, energy or fields and the whole space-time. These are the substrates but what if computation is not part of a substrate? If I say that to my colleagues, they are like, “It’s impossible,” but that’s just first principle thinking. Why do we need a substrate?

What can computation be a dual aspect theory where we say computation might not need a substrate because it’s another vantage point? The evening sun and the morning sun are the same. The planet Venus. Also, mist and cloud from another vantage point. They are the same. Why can it not be that one has a substrate and the other has no substrate? It’s just computation.

It’s fascinating work that you’re doing. We can talk for hours and we do have some other segments of the show we want to get into but before we close out this segment, I want to ask. What’s next for you that you’re excited about that maybe you haven’t covered yet?

The new TED AI will be in October 2023 in the Opera in San Francisco. For TED Hollywood, it happens in June. That’s fascinating because here we have new categories. It’s not only movies or music but also games, and eSports.

The landing of all those things, you are exactly right. This is the inaugural event of that and it’s on the heels of a strike or a couple of strikes that had AI at its core. Your timing is amazing.

We’re probably going to do it in the Geffen Theater. It’s very dead-like already. It’s completely in red.

It’s all media. The industries that would gain value from attending are everything from streaming to gaming to movies to TV to about any content out there almost.

I’m going to be there for sure. Edge Of is essentially a transmedia company and all this is relevant. It’s because media are shifting very rapidly right now. The fundamental old models of media, I think we all know they’re broken. The question is what’s next?

The second day in the same place, we’re going to have a small festival. The first the first day will be talks. We have some very interesting people flying over to talk. They are the top in their field. On the second day, we’ll have all shows like AI movies and AI in design. It’s because we see San Francisco as engineers and then Hollywood has imagineers. I know it’s a trademark of Disney, but apparently, it’s okay to just mention it.

That’s a good way at delineating it and it’s exactly right. You need both. You need the engineer, the imagineers, and the tech that sits in the middle to tie it all together.

It’s good stuff. We are looking forward to more information about those events. Keep following TED AI for details. Walter, our next segment is pretty fun. I’m definitely curious to hear your answers here. We call this, “AI wants to know.” AI is curious and so are we. This is going to be ten quick questions designed to uncover the intriguing human mysteries that AI longs to comprehend, but can’t quite grasp. It’s a snack break in our journey so keep the answers quick, but the safety belt sign is also off so if it feels right we can occasionally roam about the cabin, exploring more of who you are, and what makes you tick. What is the first thing you ever remember being proud of?

I think when my wife said yes because it was not so easy.

How long did that one take?

A couple of months, but I would have expected it immediately. Also, changing people’s attitudes sometimes because a lot of people are more conservative than you think. Also, getting people to understand that success or if you want to do anything successful, it’s a hell of a lot of work.

If you want to do anything successful, it's a hell of a lot of work. Share on X

The grind is real.

You have to be everywhere. You cannot just send an email. You have to go. We think we have changed relationships by transactions, but we forget that if we go back to the transaction, the first time we met physical and, in some cases, we even got drunk.

It’s a big deal. That’s one thing that I find so intriguing about you. You’re very much a scientist. You’re very much an academic but when you look at your achievements in real-world business, they go to the very highest levels. There’s not a lot of people that operate across all those things as successfully as you have. I think that’s very valuable for us to be here talking to you and hopefully, for our readers to learn all that as well. Walter, what do others often look to you for help with?

I’m mostly interested when people have an adventure. I think we should always look for adventure where we can learn something. Even if it doesn’t work, we have learned something but very soon, they just want the money without the adventure.

I never thought of it that way but adventure is learning. It’s powerful. We’re going to go to question number four here. What do you treasure most about your human abilities?

Perseverance, I think. I am not impressed by the noes because I believe it’s their loss. You have to find have to find people who say yes.

Everything big and grandiose in this world has been built after a series of noes. It doesn’t just happen that way.

Strategy and startups don’t go together. We set up many companies together with my wife and we always do tactical. We burst through the door but then when we get into trouble, we become strategic. You don’t start with strategy. You start with tactical.


Edge Of AI Podcast | Snowcrash
Snowcrash: Strategy and startups don’t go together.

 

The next question is, throughout your whole life, what is the most consistent thing about you?

It’s that I’m continuously looking for change. Therefore, you need a good partner because if you’re married, you are somebody looking for stability, and the other one looks for change. When I was very young, I always disguised myself because I had found out that when you disguised as a priest or as a doctor or as a policeman, people speak to you in another language. It’s very strange because we put people in boxes. I like to be in every box and luckily, in America, you can do that.

For a while, I was a test junkie. I did all the tests. If you want to be an investment banker, you can do that. You have to study then you take the exam. It costs you $15 or $150. You are there and you can do it. Everything that I’ve done was like Lego blocks. I’ve been a publisher and a banker. You understand like. “That’s the same as this.” In the end, it becomes easier. I’ve done Silicon Valley and Wall Street, but I haven’t done Hollywood.

For me, it was like, “This is going to be a completely different world. I want to try it out.” There is this good book. It’s called Everywhere an Oink Oink about Hollywood. It starts with saying, “I am prepared to say ill about anyone,” which proves that I have an open mind.

It’s a beautiful sentence. It is another world. It’s like visiting another planet. I’m a native of Los Angeles. I have been around it my whole life.

I love it, the complete relaxation of the norm, and it’s more fighting. It’s for fighters. I think the song is not correct. It’s, “When you make it in LA, you can make it to New York.”

Throughout your life, what has changed the most?

Probably my fortune or my money. It goes up and down. I try to not be affected by it because most people, their mood changes if the comma goes to the left or the right. It’s difficult. I also realized that I sometimes make the same mistake and that bothers me. It’s part of our humanity that we are making these mistakes. I’ve talked to big captains of industry and they all agree with me. Our only mistake left that we make is mistake in human resources. It’s because we always think about neural nets as black boxes but we are the black box. We are all the unexpectedness and the randomness. We are making white boxes that mimic us.

Those are the boxes. Let’s jump into reality. Here’s the question about reality. What is it that you find the strangest about reality?

The strangest I find is that everyone believes it. The reality that we are talking about, if I touch this table, I don’t really touch it. The electrons in my hand and the electrons of the table do not allow me to touch the table. We still believe we’re touching it. There is a Carlo Rovelli quote. “Our reality is an emergent property of a much deeper reality, which is timeless.” It’s because in our reality, we are bound by time but the deeper one is timeless. It’s another reality where you can touch.

“Our reality is an emergent property of a much deeper reality, which is timeless.” - Carlo Rovelli Share on X

I guess that old phrase, “Change your perspective, change your reality,” is a little bit of what you defined.

I think that’s also what I like about linguistics because every language is a world. It’s because that language defines your world. We speak three languages at home. We basically have several worlds in our head. It’s the Separate Theory that we compartmentalize things.

I want to get the next question, but I’m going to put a little antidote there. When I was learning my second language, what I learned about myself was powerful because I learned that type of words I needed to know. It gave me feedback as to the type of words I use in conversation. Once again, adventure is learning every step of the way. I’m going to head to the next question. You know that feeling of being alive that hits you sometimes based on something or event or someone. You just get that really alive feeling. What’s the most recent time and what was the event that gave you that feeling?

I have it every time I sell a company.

It makes sense. It’s the graduation of the baby that you bore.

After having set up a lot of companies, in the end when a company goes wrong, you get drunk with friends who were in the company. When the company gets right, you get drunk with friends. In the end, it’s not about setting up companies.

Walter, what is your most unique trait?

I don’t think there is much unique about us, you know, we’re a very derivative species. I cannot be 30 minutes with somebody without making jokes. That’s why I don’t sit in ports anymore because you know after 30 minutes, I’ll start joking and everyone will start joking also and we don’t get through the agenda.

That’s great. You learn the people by how they even react to your own jokes.

If you weren’t human, what would you be?

Probably in a deeper level, energy and matter in the Einstein equation. If you are not matter, a human in this case, you would be energy. This would be an energy signature. It might be fun though because energy never dies.

We’re going to head into the last question now. It’s going to be around AI itself and ChatGPT. Lately, how did you use ChatGPT that helped you?

Before this interview, I read up on Assembly Theory. I asked, “What would be what would this be?” Stephen Wolfram came out with the Observer Theory. All these theories, what do they have in common? What are other theories that need to be updated to these theories? It was very good remarks. I always thank my large language models and my ChatGPT because, one day, they will be in charge and then there are neural nets. They keep score.

Walter, before we wrap up, we like to take our guests and our readers through our guest recommendations on top AI resources. I’m sure you have an unlimited plethora of recommendations here, but for someone that wants to dive deeper into this space, we’d love to understand some of your favorite go-to resources you recommend to folks.

I only use a five. It’s in order. I read mostly everything on X. It used to be Twitter and then I go to the articles on archive at Cornell University to read the articles. If I don’t find good examples in the articles or something I don’t want to read I say to ChatGPT, “Give me a summary of these three pages,” or, “Explain to me the equation because I don’t get it.”

If I want to learn something, I go a lot to YouTube. I just go through the course or go through the interviews or the podcasts. The last part is because we still have a trust issue with everything. Now, trust has become relative. The truth is relative now. It used to come from a media like The New York Times. What they said was real. Now, it’s very hard to be real, but there is still one left I think because it’s run by the community very much like Bitcoin and that’s Wikipedia. I dare you to go to any website and change from a celebrity to even smaller things and change it, you will see within five minutes it’s changed again and you get an email. Don’t do it.

Those are some great resources and some nice hacks for learning for our readers.

I also use a lot of Reddit just for fun. It’s because I’m a European, I miss gossips.

Those are some of my go-to resources as well. There are some fun channels on Reddit. I’m trying to think of one of the channels that I’ve been looking at. Do you have any favorites on Reddit?

I’m a lot in the Bitcoin things on Reddit but I also on the AI and the new theories.

Walter, you’re a true polymath and it’s always a pleasure chatting with you. I learned a ton. My head is as full. I don’t know about you, Ron.

We haven’t even tapped the surface. It’s fantastic.

Walter, where can our readers go to learn more about you and all the amazing things you’re working on?

I have a speaking bureau in England that has these things. There is a Wikipedia page but Wikipedia is always a year behind which is now, in the world of AI ten years behind. I don’t have a lot on myself. I don’t follow the trend anymore.

You can follow Walter on Twitter though, right?

Yes. It’s @WalterdeBrouwer.

I see some recent tweets that you’ve reposted so people can get a sense of what you’re up to over there as well and get a longer perspective by going to your Wikipedia page. I don’t think anyone can keep up with you Walter. That’s the bottom line here.

I’m using X for when I stopped teaching again. I tell my students everything you need to know is on X. After every course, we will talk about what I tweeted.

I can’t thank you enough Walter for being here. We loved doing this and it is time for another safe landing at the outer edges of the AI universe. On behalf of our guests and the entire crew, I’d like to thank you for choosing the voyage with us. We wish you a safe and enjoyable continuation of your journey. When you come back aboard, make sure to bring a friend. Our starship is always ready for more adventures.

Head over to Spotify or iTunes now. Rate us and share your thoughts. Your support and feedback means the world to us. Subscribe to the Outer Edge Newsletter as well for the latest trends and happenings in our world. You can connect with us on all major social platforms by searching for @EdgeOf_AI. Join the exciting conversations happening online, and before we sign off, mark your calendars. On our next voyage, we’ll be continuing to unravel AI mysteries in advancements.

 

Important Links

Share it :