In this episode of the AlchemistX Innovators Inside, host Ian Bergman sits down with Ron Green, Co-Founder and CTO of KUNGFU.AI, to explore how companies can unlock the real value of artificial intelligence. With over 30 years of experience in AI—from the early days of neural nets in the 90s to today’s cutting-edge reasoning models—Ron offers clear guidance for leaders facing pressure to “do something with AI.”

Here are the five key takeaways from their conversation:

1. Don’t Reinvent the Wheel

Ron’s first principle is simple: most businesses don’t need to build AI models from scratch. Instead, leverage off-the-shelf tools, APIs, and open-source models, and fine-tune them for your domain. Reinventing the wheel wastes time and resources when speed and ROI are what matter.

2. Proprietary Data is the Real Moat

The strongest competitive advantage isn’t the AI model—it’s your unique data assets. Proprietary data allows companies to train and adapt models for domain-specific solutions, whether in finance, healthcare, or logistics. Ron emphasizes that investing in data collection, quality, and organization today pays massive dividends tomorrow.

3. Aim for 10× ROI or Don’t Bother

With so much low-hanging fruit available, Ron argues that leaders should only pursue AI initiatives that promise at least a 10× return on investment. Anything less means you’re focusing on “nice-to-have” projects rather than high-impact opportunities that transform operations, reduce costs, or unlock entirely new products.

4. Build Safeguards Into Every AI System

AI is probabilistic, not deterministic—which means errors will happen. The solution? Human-in-the-loop processes, policy shells, and phased deployment. Ron shares a case study where KUNGFU.AI deployed an AI system for billions in loan factoring, testing it in “dark mode” before going live. The result: fraud dropped, decision times fell from 24 hours to 9 seconds, and no jobs were lost.

5. Success Depends on Culture and Alignment

The biggest reason AI projects fail isn’t technical—it’s lack of stakeholder alignment. Innovation is a team sport. To succeed, organizations need executive buy-in, cross-departmental coordination, and a culture that balances caution with curiosity.

AI adoption isn’t about chasing hype—it’s about moving cautiously, deliberately, and with impact. Ron Green’s advice offers a roadmap: start with proven tools, leverage your data, demand clear ROI, deploy responsibly, and align your teams.

Have a question for a future guest? Email us at innovators@alchemistaccelerator.com to get in touch! 

Timestamps

🎙️ Introduction & Ron Green’s Journey into AI (00:00:00)

🧠 From Neural Nets in the 90s to the 2012 Deep Learning Breakthrough (00:03:00)

🚀 Why We’re Still at “Day Zero” of the AI Revolution (00:06:30)

🏗️ Founding KUNGFU.AI & Helping Businesses Embrace AI (00:07:50)

🔑 What Leaders Need to Know: Don’t Reinvent the Wheel, Focus on Data (00:23:30)

⚠️ Risk, Safeguards & the Loan Factoring Case Study (00:31:25)

🤝 Culture & Alignment: Why Stakeholders Decide AI Success (00:37:59)

📚 Staying Ahead: Lab Days, Learning, and Curiosity (00:45:10)

💻 Remote Work, Documentation & Team Dynamics (00:47:30)

⚡ Rapid-Fire (00:52:20)

 🔗 Where to Follow Ron & KUNGFU.AI (00:52:50)

Full Transcript 

00;00;12;23 - 00;00;35;11
Ian Bergman
Welcome to season six of Alchemist X Innovators Inside the podcast, where we explore the world of corporate innovation and dive deep into the minds and stories of innovation. Thought leaders crafting the future. I am your host, Ian Bergmann, and if you're an innovation agitator like me, then this is where you want to be. Ron, how are you today?

00;00;35;13 - 00;00;36;27
Ron Green
I'm great. Thanks for having me on.

00;00;37;04 - 00;00;48;13
Ian Bergman
Yeah, thanks for joining us. It's, it's a Monday in February. It's 2025. So it's the future. And I'm still in that mode of it's the future. Because I can't believe it's already late February.

00;00;48;15 - 00;00;52;14
Ron Green
I know in in the year 2025 just sounds like the future to me, doesn't it?

00;00;52;16 - 00;01;10;28
Ian Bergman
It really does. Yeah. You know, I get that a little bit every year, right? And I actually remember I'm dating myself, but I remember I think it was actually 2005 where I was like, oh, this is the future, right? But 2025, it is the future. And we're actually kind of living it. There's self-driving cars, there's AI around.

00;01;11;01 - 00;01;23;18
Ron Green
I know, and if you do, you know, asked us maybe in 2010, would that be the case? I don't think many of us really would have realized or believed how much would have happened in that small amount of time.

00;01;23;21 - 00;01;31;24
Ian Bergman
Yeah. How much would have happened and how challenging it is to keep up and figure out kind of what to do with these tools that were given. Right?

00;01;31;25 - 00;01;37;18
Ron Green
Oh it's unreal. I mean, you know, I work in AI every day and I feel like I'm drinking from a firehose constantly.

00;01;37;21 - 00;02;07;21
Ian Bergman
Amazing. Well, I think that's a really good segue to actually introduce you to the audience. For folks listening, I am absolutely thrilled to welcome Ron Green to the pod. Ron is the co-founder and chief Technology officer at Kung Fu AI, a leading AI management consultancy and engineering firm. And now, Ron, you've got a really interesting background. I'm going to let you kind of speak about that a little bit, but as you say, you drink from the firehose every day and you contribute a bit.

00;02;07;22 - 00;02;14;05
Ian Bergman
I've seen your writings on LinkedIn. You're pretty good at kind of posting on about the news, but also adding a little insight.

00;02;14;07 - 00;02;40;13
Ron Green
Yeah, I've been engrossed and excited about this whole AI thing for over 30 years now, all the way back to when I was doing my undergraduate degree in computer science. And, it's actually kind of a funny story. I've told this before, but I was about to graduate with the computer science degree, and I was so burned out, I didn't know what I was going to do for a professional career, but I knew it wasn't going to involve software because I was just dumb.

00;02;40;15 - 00;02;59;21
Ron Green
And then I totally not. Yeah, exactly. And then I took this AI course my last semester and realized instantly that's what I wanted to do. And then, you know, I it's been really up and down since then. It's been a 30 year rollercoaster ride for me between, you know, grad school and professional work. And everything really changed in 2012.

00;02;59;21 - 00;03;01;29
Ron Green
And we've been off to sort of the races since then.

00;03;02;02 - 00;03;21;20
Ian Bergman
Well, I want to talk about that in a minute, but actually, I'd love to just start with a little bit of that casting back, like, you know, both of us were there. I think what a lot of people today would consider to be fairly early in kind of the the modern tech wave. What did I mean? As you were doing these studies you said 30 years ago?

00;03;21;20 - 00;03;23;24
Ian Bergman
So not kind of in the mid to late 90s.

00;03;23;27 - 00;03;49;06
Ron Green
Yeah, it was. So I did a master's in artificial intelligence at the University of Sussex in England in the late 90s. And back then you didn't even use the term artificial intelligence. You would talk about, you know, complex systems, dynamical systems, things like that. And we were building artificial neural networks. And at that point in time, artificial neural networks were decades old.

00;03;49;06 - 00;04;24;08
Ron Green
And in fact, you know, the the real core breakthrough in training these deep networks happened in the late 80s. And so we even knew about backpropagation. We knew all the nuts and bolts, how to train these. For the most part. What we didn't know was we lacked the necessary data by maybe, six orders of magnitude, and we lacked the necessary compute infrastructure and capabilities by, I don't know, a billion, trillion times.

00;04;24;09 - 00;04;41;11
Ron Green
I mean, it was just incredible. So we were building things that would show promise, and they would work on toy problems, but they just wouldn't scale for the most part to, you know, real world complicated production problems. Well, and it's really funny.

00;04;41;11 - 00;05;03;26
Ian Bergman
So you were running up against kind of the limits of available scale in data and compute then and, you know, fast forward a couple decades and some of the most interesting conversations now about the potential of AI are still about scaling limits. And, you know, is there a limit to what happens if you are able to throw more time, more compute, more data?

00;05;03;26 - 00;05;09;05
Ian Bergman
Yeah, at a problem. And it's really interesting. There seems to be some disagreement on what the answer is. Yes.

00;05;09;07 - 00;05;28;12
Ron Green
Yeah. And so I think, you know, there's kind of a subtle distinction there in a couple of ways. One is that I think half the people believe that if we had the ability to scale, compute and data sort of in an unbounded way, meaning let's just keep making bigger models, let's just keep feeding them more and more data.

00;05;28;15 - 00;05;56;17
Ron Green
There's really no sign that artificial neural networks, deep learning would saturate. As far as we can tell, they will be able to scale in a, you know, near unbounded way, even if that just meant they're just memorizing data, let's say. The other question, though, is like how intelligent will they be? Can they really produce, you know, and simulate or manifest the intelligence that we as humans think about?

00;05;56;20 - 00;06;28;18
Ron Green
You know, what that would be reasoning and analytic ability and things like that. And large language models which is everybody knows now like ChatGPT etc., that that's the dominant approach. And there are a lot of doubters about that poetic killer version of sort of neural networks, although I would argue in the last I think 90 days, some incredible things that have happened, the where I was pretty confident we could push this approach a little bit further.

00;06;28;22 - 00;06;40;12
Ron Green
Now I have almost, certitude that we're going to see some amazing things in the next three years pushing this paradigm forward a little bit more. Well.

00;06;40;14 - 00;07;00;18
Ian Bergman
I agree with you, although with some really big caveats. As we get into esoteric things like self-awareness and consciousness. But in terms of intelligence, I agree. But let's get there in a minute. I want to share a little bit more about your journey. So what do you do today and how did you find yourself in this absolutely fascinating seat that you are in?

00;07;00;21 - 00;07;27;08
Ron Green
I consider myself to be unbelievably lucky. So I'm a guy who got interested in artificial intelligence in the 90s. Way too premature, but did a bunch of startups that were, you know, some involving, you know, biotechnology, some involving like really bleeding edge advanced telecommunication stuff. There were smatterings of artificial intelligence and optimization and other types of technologies there.

00;07;27;11 - 00;07;51;25
Ron Green
But for the most part, we really couldn't push it that far. We just really couldn't push it that far. And then with the resurgence of artificial intelligence and deep learning, specifically in 2012, it really allowed me to go back to artificial intelligence full time. And so we founded Kung Fu AI, nearly eight years ago. At this point, all we do is artificial intelligence.

00;07;51;29 - 00;08;13;23
Ron Green
That's all we've done from the beginning. And so it was kind of funny. I really expected the world to recognize this revolution a little bit earlier. It took ChatGPT, I think, for everybody to fully appreciate it. But we've been building, you know, just amazingly capable state of the art systems in like computer vision, on natural language processing and things like that for almost a decade.

00;08;13;26 - 00;08;28;11
Ron Green
And I get to sit here as the CTO of this amazing company. I work with the smartest, kindest, most creative people I've ever met in my life, building AI systems for our clients and I. I think I would do it for free. That's how much I enjoy it.

00;08;28;13 - 00;08;36;00
Ian Bergman
I won't, I won't tell your clients what? But but actually. Seriously. So. So what's the one liner for your company and what you do for your clients?

00;08;36;03 - 00;08;58;19
Ron Green
We are a strategy, an engineering firm. We help our clients build their AI strategy roadmap and execute against it. So we don't come in to just tell you what you should do. We can actually help you build it. That means standing up a center of excellence, hiring an ML engineers. And if you want custom AI engineering, we have a team of world class engineers.

00;08;58;19 - 00;09;13;16
Ron Green
We've built state of the art systems across the entire spectrum. You name it, we've done it. And we just want to help our clients figure out how to embrace AI with their business and get really strong, solid ROI.

00;09;13;19 - 00;09;33;29
Ian Bergman
Yeah. And so this puts you at the heart of one of the really like toughest, to be perfectly honest kind of innovation and strategy questions of the day. Right. Which is a little bit defensive. How do I respond to emerging competitive and disruptive threats a little bit a little bit sort of customer centric like expectations of what is possible is changing dramatically in the market.

00;09;34;02 - 00;09;57;17
Ian Bergman
Right. There's there's a lot. So you're in a really fascinating place. I want to talk a little bit though, more about the timing. So you keep going back to kind of the 2012 era. And you know, I've been in and around technology and sure, some truly incredible things were happening in terms of machine learning ability to, you know, discern quality to run predictive models, etc..

00;09;57;19 - 00;10;28;13
Ian Bergman
But I will say, you know, I feel like sort of the, the public vibe and sentiment around, I really did shift with the transformer models with, you know. Sure. Actually, yes, chat GPT, but even that even the playground models. Right. That came out around the OpenAI playground, right, with GPT two, like that was the moment as I got into the playground where I started to think about, okay, this is a different form of intelligence than a predictive ML model.

00;10;28;15 - 00;10;36;03
Ian Bergman
I mean, has that changed dynamics in terms of how you work with clients like, or is this just one long continuous evolution?

00;10;36;06 - 00;10;59;17
Ron Green
No, it it really has. No. It was it was a step change in the market when, you know, I, I've referenced 2012 a couple times. That's obviously when this sort of modern phase of deep learning broke on the scene. And what was special about it was we could finally build computer vision systems that could understand images, not necessarily as well as a human, but in some cases pretty close.

00;10;59;20 - 00;11;25;07
Ron Green
And then shortly after that, we had natural language processing and it could start to kind of understand text, and you could do things like classification and sentiment analysis. And these were really powerful capabilities for one simple reason. It was the first time in history that we had perceptive capabilities in software that that were approaching human level capabilities on these different fronts.

00;11;25;07 - 00;11;33;13
Ian Bergman
What a fascinating statement. Perceptive capabilities approaching human level. Like I, I hadn't thought about it in those terms.

00;11;33;15 - 00;11;51;16
Ron Green
Yes. In it's you know, we I think we all of us think about intelligence when we think about AI. It's it's in the name itself. But what's really crazy is for the longest time in AI, in AI research, the things we thought were hard turned out to be easy, and the things we thought were easy turned out to be hard.

00;11;51;16 - 00;12;13;27
Ron Green
So in the 50s, we imagined if we could build a computer program that could play chess and beat a human, that program would have to have amazing broad intelligence, and it would probably be superhuman intelligence. Turns out, no, it was a really narrow domain. And and we achieved that and nobody thought we had, you know.

00;12;13;27 - 00;12;14;09
Ian Bergman
Right.

00;12;14;09 - 00;12;16;11
Ron Green
AGI at that point, artificial general.

00;12;16;12 - 00;12;19;29
Ian Bergman
That there was no Turing test associated with beating a grandmaster.

00;12;19;29 - 00;12;42;06
Ron Green
Exactly. And what it turns out is that, you know, things like vision are perceptive abilities. It is so effortless to us because of all of the millions of years of evolution that we we can't even introspect it. We don't know the process that we go through when we see an image and we recognize it. The same thing goes with speech and sound.

00;12;42;09 - 00;13;02;28
Ron Green
And so we kind of undervalued those capabilities. We didn't think they were that complicated, where on the other hand, chess was so incredibly difficult for us. We you have to think about the position of all the pieces. You have to think about. If I move this piece, how will they respond and how I respond to their responding on and on and on.

00;13;02;29 - 00;13;29;24
Ron Green
So in our minds, we kind of got it flipped. Yeah. It turns out just being able to recognize whether a photo is of a cat or a dog was incredibly difficult for us to do until the last 15 years. And then once we sort of locked onto these perceptual capabilities, things like the transform architecture, we mentioned large language models that unlocked a whole different front.

00;13;29;24 - 00;14;00;11
Ron Green
And now we got into sort of generative AI. And what is so special about ChatGPT was it was the first AI system that was sort of multifaceted, capable we could build systems before that. We're really good at maybe face recognition, but that's really kind of all they could do with ChatGPT you brought to the table was a single model that could do things like summarize text, generate poetry, tell you a joke, and many, many more.

00;14;00;17 - 00;14;06;22
Ron Green
And that's the reason it finally caught the world's attention and kind of sparked this. I take off.

00;14;06;24 - 00;14;25;24
Ian Bergman
And did become something of, it's clearly a consumer platform, but it became a platform that people could it platform in the truest sense, something that people could experiment with, play with, you know, initially through GPUs, etc., but, you know, find their own creative uses and outlets that you, you could never imagine as, as an engineer. That's right.

00;14;25;24 - 00;14;30;02
Ian Bergman
I feel like we're just at the, at the beginning of, of this.

00;14;30;04 - 00;14;37;15
Ron Green
I couldn't agree more. In fact, I literally have a saying that we are a daisy. We're not even at day one yet because what does that mean?

00;14;37;15 - 00;14;41;22
Ian Bergman
So okay. Well, I mean, and we're we're not talking Jeff Bezos here. What is what does this mean?

00;14;41;28 - 00;15;26;11
Ron Green
Right. The reason I say say zero is we're still tinkering around with simple ideas. We're still doing things like, what if we put approach A and approach B together? You know what? If we put the peanut butter and the chocolate together, right. And just try simple, simple, simple things. We'll be it. Day one I think once we are really starting to push the boundaries on the approaches, that and we've exhausted them in sort of a meaningful way right now, we're still seeing, for example, like with Deep Sik a few weeks ago, we did in Riddick, loosely simple, you know, reinforcement learning approach they took that has fantastic results.

00;15;26;14 - 00;15;31;13
Ron Green
And then we can go into that if we want to. But just really simple stuff is still moving the needle in a big way.

00;15;31;15 - 00;15;43;24
Ian Bergman
And so when you're talking about day zero, you are you know, you're talking about from a computer science and research and engineering perspective for sure. What about from sort of a societal and adoption perspective?

00;15;43;27 - 00;16;06;09
Ron Green
Yeah, that's a really good point. I think that most people are starting to use AI on a daily basis, even if they don't necessarily know it, but I think most people do. I think most people are using ChatGPT or Claude or something like that increasingly. But the capabilities of those systems are growing at a rate that I would be shocked if most people understood.

00;16;06;09 - 00;16;07;01
Ron Green
For example.

00;16;07;01 - 00;16;11;02
Ian Bergman
I don't I follow it every day and I don't understand. It's blows my mind exactly.

00;16;11;09 - 00;16;20;22
Ron Green
You know, last summer there were no reasoning models, and the reasoning model is one that doesn't just generate its output, it will basically talk to itself. It will think out.

00;16;20;22 - 00;16;23;12
Ian Bergman
Loud, take an agent approach within the own model.

00;16;23;15 - 00;16;49;10
Ron Green
Yes. And those those models have gone from being laughably bad at math and things like if you said, how many, how many letters are in the word strawberry, you just, you know, famously terrible, terrible ability to reason that now we have these models, just six months later that can solve, you know, math, Olympic level algebra and calculus and geometry.

00;16;49;10 - 00;17;25;14
Ron Green
And we have models that can do really, really, really complicated analysis and introspect on their own output. And what's so important about that is once these models have the ability to reflect on their own output, it's almost like recursion. It's like we can think and we as humans, we can think about thinking. And as soon as you take that one recursive step, then there's nothing stopping you from thinking about thinking about thinking about thinking ad infinitum.

00;17;25;14 - 00;17;27;05
Ian Bergman
It's just it's just compute.

00;17;27;06 - 00;17;39;27
Ron Green
It's just compute. It's, you know, there's just compute. And we're right there with artificial intelligence. In many ways. It's a little bit of an oversimplification, but it's a it's an analogy that I think is pretty strong for where we're at and where we're going right now.

00;17;39;29 - 00;17;58;05
Ian Bergman
Well, and I actually think it's really it's really well said and really important. So one of the things that I've really been struggling with actually, in my own conversations, my own world, is how to articulate just how profound some of the advances have been that have been tied up in kind of the, let's call it the large language model approach.

00;17;58;05 - 00;18;18;18
Ian Bergman
Right? Even even a year ago, I was able to really cleanly make the point that, hey, like, this stuff is advancing really quickly with generative imagery and video, right? I, I, I remember it's in front of so many audiences, I would be like, all right, let me show you, like the first thing I did with Dall-E or the first thing I did with Midjourney, right?

00;18;18;18 - 00;18;42;10
Ian Bergman
Which was really cool. And I nerd it out and then six months later it was better, and six months later it was photorealistic. And now, you know, have a video clip, right? That is almost indiscernible from something real. Although interestingly enough, it turns out when you generate stuff and then show it to a true expert in the field, they pick up on, on the interpolated inferred made up stuff really quickly.

00;18;42;10 - 00;19;12;16
Ian Bergman
Yeah. But anyway, yeah, so so like with pictures it was easy. But now everybody nods their head and kind of accepts that. Yep. Okay. You want an AI generated Trudeau saying things done? It feels much harder to me to make real the advances on sort of the the logic, the reasoning and the text side. How do you explain to business leaders, you know, folks worried about strategic imperatives, competitive threats, whatever.

00;19;12;16 - 00;19;19;28
Ian Bergman
How do you explain to them what that the advancement means and what they need to do about it? It's kind of a big question.

00;19;20;00 - 00;19;50;18
Ron Green
You know, I would probably lay a foundation briefly on what's happening and why it's significant. And and very briefly, it's this in the last six months, we've realized both in sort of Frontier Labs privately and now with open source frontier labs like Deep Sea Org, which are publishing their research, that some of these techniques that we've played around with before, we need to take a look at it again.

00;19;50;18 - 00;20;17;02
Ron Green
In particular, there's something called reinforce learning and was so beautiful about this is over the last decade or so, a technique called supervised learning has been by far the most dominant technique. These AI systems, they learn through examples. And so what we do is we'll we'll give it an input and the model will make some output, a prediction or an image, whatever, whatever it's designed to do.

00;20;17;05 - 00;20;37;18
Ron Green
And we will correct it and say, no, it should have been more like this. It should have been more like that. And if we do that with enough examples, that model doesn't just memorize the right answer, it learns to generalize and it can handle inputs it's never seen before and give the right output. That's great, but that means we have to know the answer for everything.

00;20;37;18 - 00;21;12;07
Ron Green
That means we have to know how to solve everything we've ever shown it. And where we're going now with reinforcement learning is, and this is really new, we have realized in the last six months that if these models reach a certain level of capabilities, then we can use them with this technique called reinforcement learning, where we ask it to solve problems and we can essentially cold start it with math and, computer programing examples.

00;21;12;07 - 00;21;42;17
Ron Green
And it essentially develops this built in almost like analytic capabilities, these reasoning capabilities that we can apply to other domains. And what's going to happen in 2025 is there's going to be a the explosion of verifiable reinforcement learning models. And we're going to see math and coding. Those are just the tip of the spear. Everything else is going to start falling under this umbrella.

00;21;42;19 - 00;22;12;25
Ron Green
The legal domain clinical diagnoses insurance. You know I every every business process, everything you can think about. And so earlier when I was talking about perception and images and text. Well now imagine if you can apply domain specific reasoning. So every business is going to be disruptive. And if all of that wasn't enough, the beautiful part about this is we're now seeing that these models are integral to training the next generation.

00;22;13;00 - 00;22;29;14
Ron Green
So all of the great models that will be released in 2025 will be trained on the basis of the models from 24, which were trained from the models on 23. So we're essentially standing on the shoulders of the generation before.

00;22;29;15 - 00;22;36;02
Ian Bergman
Which is exactly human learning in a way. But compressed by orders of magnitude.

00;22;36;07 - 00;22;48;08
Ron Green
Exactly. That's a perfect analogy. And so we are going to have models with this ever sort of tightening feedback loop, at least over the next few years. It's going to be it's going to be insane.

00;22;48;10 - 00;23;13;11
Ian Bergman
So, you know, let's take this back to people and, and let's stick with kind of business leaders. Business leaders are struggling. The folks that I talked to write privately in their vulnerable moments, it doesn't matter whether we're talking fortune 500 C-suite or, you know, the director that just got hired into their new marketing job. They're they're struggling with uncertainty and, you know, kind of inability to predict the future.

00;23;13;13 - 00;23;35;28
Ian Bergman
How do we, as an industry, help folks think about navigating the this change? Because there's powerful new tools? I will say that, you know, I think there's still a very significant gap between what we understand they should be capable of and what actually kind of, you know, comes out in practical use. But there's very power, powerful tools. It's changing very, very quickly.

00;23;36;00 - 00;23;37;14
Ian Bergman
What do we do?

00;23;37;17 - 00;23;56;19
Ron Green
There's a great question in like, you know, not not to exaggerate, but I would I would say, with all you know, seriousness. That's literally why we exist as a company as kung fu AI. And it's really not that complicated in the grand scheme of things. It's a few things. One, when you were thinking about investing in, I don't go or reinvent the wheel.

00;23;56;19 - 00;24;16;21
Ron Green
If you're trying to build a language model, use something off the shelf. There are amazing open source things, and if you need them to do something special, like in a narrow domain, great, you we can fine tune them. There are things like rag that can can help with hallucinations. So first off, don't go. Don't go. Reinvent the wheel.

00;24;16;23 - 00;24;43;13
Ron Green
Number two, there are companies out there with products that are really, really good. And if long as those products can solve your your need, just pay for it. Go, go pay for ChatGPT right. If you need to do OCR, just go pay for the API. That's that's there where businesses should be putting their money and their investment is where they have a where they can have a competitive differentiation.

00;24;43;16 - 00;25;27;01
Ron Green
And right now that mostly involves one thing. What data do you have that is proprietary, that you can leverage in. And there's many ways you can do this, but that you can leverage that data to enable either new capabilities, new product features, a cost reductions, automations and predictive trending recommendations, things like that. It's all about your own proprietary data, and that's the beautiful thing about this moment in AI, where ten years ago, even if you had proprietary data and you wanted to use it, it was going to be expensive because you had to start from scratch.

00;25;27;03 - 00;25;51;22
Ron Green
Now, it doesn't matter what you're doing. There are incredibly powerful models out there that you can use as the basis for that, and you can fine tune them, customize them, do whatever you need to do on your own data. And then the last thing I would say is there are real needs. And I think this is going to be for the foreseeable future for what I call domain specific AIS.

00;25;52;00 - 00;26;21;11
Ron Green
That means, I'll give you an example. We built a system that can predict race cancer five years in advance. A computer vision system. Unreal, can do one thing. That's all it can do. But it's important that there is nothing else in the world that can do that. And there are unlimited number of opportunities like that for businesses to go and innovate and provide superhuman capabilities that you can't get ChatGPT or Claude or any of these other, you know, language models to do.

00;26;21;11 - 00;26;21;19
Ron Green
Okay.

00;26;21;19 - 00;26;41;00
Ian Bergman
So you're making, I think, two really interesting points here, like the first one, and this is going to sound a little dismissive, but I don't mean it to be actually, because I think I think it's so pure and important. And it's simplicity is don't reinvent the wheel. Like this is like innovation trap 101. If you're not an expert at building and training models, why bother, right?

00;26;41;01 - 00;27;08;21
Ian Bergman
That's right. But it is really important to remember in today's world we have a lot of companies that have, you know, a lot of in-house staff and engineers who are curious and would love to learn and build this stuff, but so don't reinvent the wheel. But the second point that you made, look at the thing that you have data and look at these painful maybe point problems or point solutions that are just hard and maybe, maybe you can do something kind of magical today.

00;27;08;23 - 00;27;30;00
Ron Green
That's exactly right. And it's really funny, the I think saying don't reinvent the wheel is a perfect summary, but there are so much pressure. I know I talk to CEOs and CTOs all the time who say something just short of this, hey, can you give me some AI? And I will say, well, what do you need? No go, I don't care, I just need some AI.

00;27;30;01 - 00;27;32;08
Ron Green
The board is going to eat me alive.

00;27;32;10 - 00;27;50;00
Ian Bergman
Well, and I think. But I think you know. So look, I see and hear that as well. And I think that actually really speaks though to this theme of like uncertainty. Fear. Like we don't know what to do. So right now that is a reactive statement. That is a like I need to show that I am, you know, keeping up with the market.

00;27;50;00 - 00;28;19;07
Ian Bergman
Yes. Which is just harder than ever to do. But you do want to go beyond that, right? Anybody can throw up a branded interface into a chat bot on their website or anybody can, you know, say, oh, hey, great. I deployed GitHub Copilot across my engineering team. We have AI, right? That's right. But I like what you were saying about, you know, maybe get that out of the way so you get the pressure off your back and then you say, what asset do we have in what is often usually our proprietary data?

00;28;19;07 - 00;28;23;19
Ian Bergman
And what is a problem. Yeah, that maybe we can see if we can do ten x better.

00;28;23;21 - 00;28;41;07
Ron Green
That's right. And in fact, you know some people may think that I'm being dramatic if you're not getting at least a ten x ROI on whatever AI initiative you're looking at, don't consider it. And the reason is there's so much low hanging fruit right now, most companies have done almost nothing.

00;28;41;07 - 00;28;56;14
Ian Bergman
So what is interesting that is a that is an argument for try fast, fail fast, try fast, fail fast. The world is still world of startups and entrepreneurship and how but like it's an argument that you will find something don't just stuck right and we'll.

00;28;56;14 - 00;29;17;22
Ron Green
Get stuck don't just, you know throw 100 projects at the wall. Well, another thing that's really a little bit of a slippery slope that people have to understand is unlike traditional software, you know, modern AI is probabilistic. And so we'll see people who will go do their first AI project and they'll say like, oh, we got to like, you know, 60% accuracy in the first week.

00;29;17;22 - 00;29;23;00
Ron Green
And they're like, how long could this possibly take? It gets really, really tricky going from this.

00;29;23;00 - 00;29;25;00
Ian Bergman
You need five nines. You're in a different world.

00;29;25;00 - 00;29;50;29
Ron Green
Yeah, exactly. I mean, and it is just really, really difficult. Look at all the advancements we had with self-driving cars. It'll look like in 2018. You would need a driver's license in 2020. Well, it's 2025. Nipping every corner case is really, really, really difficult. So don't underestimate the complexity. But at the same time, don't do anything that doesn't give you differentiated value.

00;29;50;29 - 00;30;16;06
Ron Green
And that's all that's all predicated on your data. And I have to say one more thing, which is we see this all the time where companies have a very, very clear initiative that they could attack. It would be differentiated and they have the expertise, but they weren't collecting the data, or they collected the data and they threw it away, or the data is there, but it's fragmented and only some of it was collected.

00;30;16;06 - 00;30;19;10
Ian Bergman
Well that's everyone I mean, that's every organization I've ever worked on.

00;30;19;14 - 00;30;36;28
Ron Green
It's it's everywhere. And so, you know, that old saying about like the best time to plant a tree is 20 years ago. Second best time is today. Yeah. Go analyze your AI opportunities. Go look at your data. You know, what's the reality of your data? And if it's not there. But you have big initiative that you want to go after, that's fine.

00;30;37;04 - 00;30;43;26
Ron Green
Start getting your data in-house situation under control now and then you can go attack these in a year or 2 or 3.

00;30;43;27 - 00;31;04;21
Ian Bergman
I actually love that. Right? Like, I think it's so important to remember that, you know, when you have your base infrastructure, when you have the platform in place, whatever you want to say, you can, you know, build out the solutions and I the problem is that's kind of the unsexy stuff, right? What's sexy is the the customer facing the investor facing the, the something deployment.

00;31;04;24 - 00;31;25;09
Ian Bergman
But luckily a lot of I mean, you know, most most organizations have the ability to allocate capital and resources against against future benefit. That's right. I do have kind of you know, there is a hard question, though, that everybody struggles with in deploying AI, and that is, well, the classic innovator's question of what if this goes wrong? Right.

00;31;25;11 - 00;31;44;10
Ian Bergman
And so you're probably familiar with one of my favorite examples. This is Air Canada about a year ago, a little over a year ago, their, customer service bot. Yeah. The challenge that I have with this, I mean, for folks that aren't from I think I've brought it up on the pod before, but for folks that aren't familiar, Air Canada got sued and lost because their generative customer service bot.

00;31;44;10 - 00;32;09;19
Ian Bergman
I think they promised a fair that didn't exist and they tried to not honor it. Right. And you know, for what it's worth, I think the outcome of the judicial process was both entirely predictable and I think, right, you know, there was an agent of the company that made a commitment, and now they have to honor it. But what's tough to me about this, right, is like they did exactly what I would tell anyone to do in their position.

00;32;09;21 - 00;32;32;06
Ian Bergman
Right. Find a big problem. Multi-hour or long phone queues. Right. That was killing them at the time. You know, test and roll out technology. Try it and then, you know, iterate. Yes. But they got sued. There was a press cycle, you know, all of the things that scare boards and C suites and investors, it's going to keep happening.

00;32;32;08 - 00;32;35;21
Ian Bergman
So what do we do about the risk profile with rolling this stuff out?

00;32;35;23 - 00;32;59;05
Ron Green
Okay. Great question. I have many three different points I think that would be relevant here. One is that generative AI exploded on the scene in late 22. And I think what everybody forgot when they were using it was you would interact with these language models and you would ask questions and refine it and, you know, essentially work with it in an interactive way.

00;32;59;05 - 00;33;24;08
Ron Green
And people just forgot about the interactive part and they went and said, let's just plug in generative AI systems and have them just produce content and give it right to the end user. And we had come for AI for the last, I guess, three years now. We've been really, really pushing back on this approach. Generative AI is incredibly powerful, but like humans, you know, it can make mistakes, it can hallucinate.

00;33;24;08 - 00;33;47;10
Ron Green
Even the reasoning models that are available today are far from flawless. So we always advocate if you're doing a generative approach, make sure you have some type of human in the loop so that you're more augmenting your human and and increasing sort of the leverage versus just completely taking away humans job and automating something. So they didn't do anything wrong.

00;33;47;10 - 00;34;11;05
Ron Green
They just misunderstood sort of the corner cases. If it was right 99% of the time probably wasn't enough. The good news is the models are increasing at a just breathtaking rate. A lot of the sharp edges are getting soft and and we're going to see that more of that. Now. The other part of it though is so I'll give you an example like, where we put our money, where our mouth is.

00;34;11;05 - 00;34;33;22
Ron Green
We have a client, publicly traded company, non-disclosure can't say who they are. They do billions in loan factoring per month. We built a system, took about a year, and it was a combination of an AI system that learned from, you know, years and years and years of data. But then it had a shell like a protective policy shell put around it.

00;34;33;29 - 00;35;01;07
Ron Green
So it didn't matter what the I thought if certain certain rules and requirements weren't met, they were overridden with traditional software, you know, safeguards. Okay. So the model could make predictions and could have a confidence level that were operated on. But there was still this policy. Shell. Well, we still didn't roll it out like that. We then spent six months testing this in dark mode where we would have everything flow through it.

00;35;01;07 - 00;35;16;03
Ron Green
I'm talking literally billions and billions of dollars worth of loan factoring decisions, and we watched what the model would do, and then we watched to see how it ultimately how that that transaction closed was the money paid back.

00;35;16;03 - 00;35;17;08
Ian Bergman
Was the app totally.

00;35;17;09 - 00;35;44;26
Ron Green
You know, was everything transpiring the way the model predicted? Only after six months did we say we have the confidence, and then we just slowly turn the dial up a little bit, a little bit, a little bit with monitoring 24 seven monitoring that models now in production is trading hundreds of millions of dollars per month by itself. Frauds down for they've gone from a 24 hour to around nine seconds.

00;35;44;28 - 00;36;05;20
Ron Green
It's outperforming humans. And no humans lost their job. Now they take the humans and they put it on all of the decisions that the model can't figure out that are too complex. And to me, this is a is a just a real wonderful example of how to move thoughtfully and cautiously and not just flip the switch on.

00;36;05;20 - 00;36;07;10
Ian Bergman
Some and deliberately, though.

00;36;07;12 - 00;36;08;09
Ron Green
And deliberately.

00;36;08;14 - 00;36;26;21
Ian Bergman
Cautiously, but deliberately, which is, I think, really powerful. And you made a really interesting point. It amplified the capabilities of the humans in the job. I think there's this, you know, constant fear. Like, what do we do? Are we going to have enterprises that have six employees and a bunch of digital agents? That seems unlikely to me for so very many reasons.

00;36;26;21 - 00;36;51;26
Ian Bergman
Agree. Right. But I think that's a really interesting point. And and I love that idea that you can also do this. You can run this digital AI twin for however long you need, and you can test it. And, you know, presumably, obviously, we're not going to ask you to violate the confidence of your clients. But, you know, presumably it doesn't get deployed unless the data fairly clearly showed that there was a significant improvement in some of the factors that you identified, right.

00;36;51;28 - 00;36;54;10
Ian Bergman
Repayment rates, risk monetization.

00;36;54;13 - 00;37;21;13
Ron Green
That's exactly right. And and you know, and look, they were nervous and they were understandably very excited, but concerned that maybe all this AI stuff wasn't going to live up to the hype. And now I love it. You know, we get to see their, you know, quarterly press releases and they are all in on AI now. It's now it's clear that there's a path to not only achieve these really lofty goals, but do it in a sort of a safe, measured way.

00;37;21;13 - 00;37;32;07
Ron Green
And, you know, big kudos to them because they, they trusted us and, and were willing to to go the slow route, what we call dark mode to make sure everything was tested before we flipped the switch.

00;37;32;09 - 00;37;59;08
Ian Bergman
Well, a lot can go wrong very quickly at scale when things do go wrong, so I'm glad you were able to give them that advice. I guess, if we kind of step back when you're working with clients across different verticals, right? Finance and underwriting, health, retail, I don't know, wherever, whoever you might be talking to, in whatever vertical they're in, what characteristics or patterns do you see that distinguish?

00;37;59;08 - 00;38;14;10
Ian Bergman
And I ready company from those that maybe are really going to struggle. Right. We've talked about data hygiene and we've you know, we've talked about kind of focusing on problems. But is there something cultural you can see? Is there something in the executive mentality?

00;38;14;13 - 00;38;26;28
Ron Green
You know, the number one reason AI projects fail is really on the human side. It's lack of sort of coordination and Buy-In by all of the stakeholders.

00;38;26;28 - 00;38;27;28
Ian Bergman
So sounds familiar.

00;38;27;28 - 00;38;51;28
Ron Green
We'll see this really quite frequently. Yeah. Where maybe parts of the business are very forward looking and they're ready to go and they want to explore, but it gets completely undermined because, you know, this is a team sport, right. And you know, if it's on a board, if operations is on board, oh my gosh. You know, heaven forbid the the domain matter experts aren't on board.

00;38;51;28 - 00;39;11;08
Ron Green
You can't walk in and just throw some AI at a at a problem and expect it to succeed. So it's really kind of a combination of maybe underestimating the complexity on the data front and the training front and mistaking, you know, being halfway there in a day with thinking, you know, it's going to take two days. It's still pretty complex.

00;39;11;08 - 00;39;27;06
Ron Green
It's a little bit more art and science, but that is just execution. It's really about moving together and getting buy in with all of the executive team stakeholders that need to be bought in for it to succeed. That's the number one thing.

00;39;27;10 - 00;39;33;02
Ian Bergman
So all innovation is stakeholder management. I mean that like it really is. Yeah, of course it is. Right.

00;39;33;04 - 00;39;49;13
Ron Green
It really just comes down to people. I mean it's always people, right? It's just, you know, I like to joke, you know, life would be so easy if it wasn't for people, right? You know, because it's we're complicated, complex people and everybody's got their own, you know, agenda. You have to get alignment on these things.

00;39;49;16 - 00;40;12;06
Ian Bergman
So what do you say to the leader? It might be a C-suite leader, might be someone else, kind of a divisional a divisional leader. What do you say to that leader who has found him or herself unbelievably inspired by kind of the art of the possible, like the belief that something really cool can change in their business. And, you know, they ask the question A, is this possible?

00;40;12;06 - 00;40;14;19
Ian Bergman
And B, what do I what do I do first?

00;40;14;21 - 00;40;30;29
Ron Green
What you need to do is whether you're working with us or some other company that has real experience with building production, AI systems is just nine days between a proof of concept and getting something in production one.

00;40;30;29 - 00;40;31;13
Ian Bergman
Hundred percent.

00;40;31;15 - 00;40;54;20
Ron Green
Don't go build. I just say you can build a. I go do a thoughtful dive on what the opportunities are there. Like we talked about, some are going to be cost savings, some are going to be automation, some are going to be new product capabilities that that may be worth the investment. And then before you do anything else, go understand the data situation.

00;40;54;20 - 00;41;17;24
Ron Green
Make sure you have the necessary data that's of quantity, quality and distribution to solve what you need to build that AI system. And when I say distribution, I mean if you're trying to build like let's say a fraud detection system, but you don't have any examples of fraud, right? That's going to be a problem. It's going to be hard.

00;41;17;24 - 00;41;41;03
Ron Green
By definition, fraud is in the minority. So you have this skewed data set already that you're almost always dealing with. So I don't go racing out and do an analysis on that. And then and only then when you have a good sense of the ROI and the data quality, then you can start thinking about AI and technology and execution and things like that.

00;41;41;05 - 00;42;01;10
Ron Green
It's really it's really just making sure that you are sweating the things that would cause you to fail early. And so I bring it all the way back to what you said, which is like fast fail, go. You can go explore ten initiatives, but don't go spend a bunch of money building the proof of concept. Go spend the time to investigate them first.

00;42;01;12 - 00;42;04;15
Ian Bergman
You know. Do you ever read the book How Big Things Get Done?

00;42;04;15 - 00;42;06;02
Ron Green
Oh, I think I did, but years ago.

00;42;06;05 - 00;42;14;00
Ian Bergman
Yeah, well, it's it came out a few years ago and it's one of my absolute favorites. If you've read it, you may recall somebody, oh.

00;42;14;00 - 00;42;16;03
Ron Green
You know, maybe I, maybe I have it then it's.

00;42;16;03 - 00;42;35;24
Ian Bergman
It's, it's really good. I listen to it as an audiobook. I'm forgetting the name of the author, which is unfortunate, but we'll link it in the show notes. But it talks about both physical and digital, like major infrastructure projects, and it's absolutely fascinating. Like everything from the Sydney Opera House to massive digital systems to remodeling your kitchen in New York.

00;42;35;24 - 00;42;56;20
Ian Bergman
But like the key point to the book is exactly what you just said. Like, you take this entire book full of unbelievable stories of successes and failures in large, large, complex projects or smaller ones. But the key point in the book boils down to invest your time in preparation and learning right before you scale and go to production.

00;42;56;25 - 00;42;58;20
Ian Bergman
You just instantly reminded me of it.

00;42;58;27 - 00;43;27;18
Ron Green
Yes, I there's a there's an old famous computer science quote. I can't remember who's who said it, but it basically says every successful large system started off as a successful small system. It's just about impossible to go build some large, complex thing from scratch and have it succeed. And you can even see in AI, we've got these giant models now, but we worked our way up there were there were a thousand milestones.

00;43;27;20 - 00;43;39;06
Ron Green
And you back in the 90s, we were building neural networks. I literally was building neural networks that had like less than a thousand parameters, you know, which is so laughable now that we have well.

00;43;39;06 - 00;43;48;10
Ian Bergman
Let's give people context because what, what the estimates of like ChatGPT for are are, what, 400 billion parameters or trillions, I actually forget yeah.

00;43;48;10 - 00;43;52;20
Ron Green
I know it's it's in the trillions. Yeah. And importantly they're closed so nobody knows for sure.

00;43;52;20 - 00;43;55;19
Ian Bergman
But okay so let's take lama like the latest lama model right.

00;43;55;19 - 00;44;26;06
Ron Green
Was yeah that was 405 billion. Yeah. And and that is you know, open source that that model would be very complex. Lama is currently we know this for a fact because Yann LeCun the their chief AI scientists said this in December. Their training, their next generation model on 100,000 H100 GPUs, each one of those 100,000 GPUs from Nvidia costs 30 grand retail.

00;44;26;08 - 00;44;42;07
Ron Green
And so, you know, we to talk about unimaginable scale. But we didn't start that way. We worked our way up. And they're willing to make these investments now because they have confidence built up over the years. Even meta needed to start small to get to where they are now.

00;44;42;09 - 00;45;10;07
Ian Bergman
Everyone. It's such a great lesson. Well, okay, so as we kind of wrap up here, I have a really important two part question for you. One, how do you keep up with the sheer pace volume of news and, you know, advancement coming up and what advice do you give to, you know, people who can't swim in this every day in terms of what they should do to try and keep up with what's going on in AI?

00;45;10;10 - 00;45;32;02
Ron Green
I'm pretty proud of this. This is this is fairly radical. You know, we're a services company. So we are we're literally, you know, out there building custom strategy and AI solutions for our clients. And we decided I think it was about three years ago now. We were struggling to keep up. There were so many white papers. There were so there were so many new projects to try.

00;45;32;04 - 00;45;56;04
Ron Green
We did something radical. We took Fridays and we now call it Lab Day. Every Friday is dedicated just to learning. We do no client work. Ask billable at all. All we do is have guest lecturers come and present. You know, over video. We read white papers together. We play around with new models and new framework. We we co teach each other.

00;45;56;04 - 00;46;19;05
Ron Green
We pair, program, do experiments, and even that's barely enough. I mean, even that barely enough for us to stay up to date because it's just moving so fast. And then as far as other people wanting to keep up I would tell them, listen to podcasts like this. There are great sort of news digests. If you want a little bit more of a, a deeper technical dive, not that much deeper.

00;46;19;05 - 00;46;40;22
Ron Green
Check out our podcast. It's called Hidden Layers. We talk about every month. We have a podcast where we cover all of the important news within AI. So you can give yourself 45 minutes, kind of get caught up that way. But sadly, if I'm being really, really honest, it's just a lot of work because things are moving at, you know, breakneck pace.

00;46;40;24 - 00;46;59;08
Ian Bergman
Well, they I mean, they really are. But I think you did hit on a couple things that I think are really important. Like one is, you know, find your source. Right. Whatever it is, you know, and at whatever frequency, find your find your source. But the other thing that you hit on, I love what you're doing, by the way, with, you know, taking Friday your version of 20% time.

00;46;59;08 - 00;47;15;19
Ian Bergman
Right? I love what you're doing. And I think I think there is something important. I feel like we are at a moment when curiosity and people who express and follow their curiosity can reap outsized rewards. Right? Sometimes it's just you've got to take the time and try it.

00;47;15;20 - 00;47;16;12
Ron Green
Yes.

00;47;16;15 - 00;47;31;25
Ian Bergman
I agree, whatever it is, right? And I love that you've institutionalized that. Honestly, I wish I could copy it in my business like we're in a completely different business. But I'm like, if only we could take 20% to hack around with ideas to, you know, debate. Oh my gosh.

00;47;32;02 - 00;47;41;08
Ron Green
It's really been incredible. It's been it's been impactful. And it really, really helps with the recruiting and retention. Yeah. You know it's been pretty magical overall for us.

00;47;41;11 - 00;47;50;14
Ian Bergman
Absolutely incredible. Well I want to spend a little time just kind of getting to know some of little anecdotes actually about you. And these things can be really fun.

00;47;50;15 - 00;47;51;00
Ron Green
Yeah. You got.

00;47;51;00 - 00;48;03;07
Ian Bergman
It. I'm just going to ask you to answer some questions rapidly. Don't think. Just answer them. Some of them. We'll just move on. Some of them we may talk a bit about. I'll start with an easy one though. You going somewhere? Are you a road trip or a flight person?

00;48;03;09 - 00;48;04;18
Ron Green
Flight. I want to get there.

00;48;04;23 - 00;48;13;03
Ian Bergman
You want to get there? Yeah. Not. It's not the. It's not the journey. It's the destination. Oh, no, I love that. How about remote in person work?

00;48;13;06 - 00;48;38;19
Ron Green
You know, I'm. I'm a blend. We went all remote during the pandemic, and now we've hired people across America, so we are remote. But, man, I miss I miss being in person. I miss the brainstorming. If we could, I would be I'd be remote Monday Fridays in person, you know, Tuesday, Wednesday, Thursday. Because I think there's like a superpower being in a room together at a whiteboard with the marker, taking turns, you know, coming up with crazy ideas.

00;48;38;21 - 00;48;56;07
Ian Bergman
I got to tell you, I want to I want to riff on this for a second because we struggle with this as well. We're globally distributed organization. We have kind of a bunch of people in the Bay area and then all over the world, and it's really hard in my experience, like, I don't feel like I have ever cracked the nut on bonding or planning.

00;48;56;11 - 00;49;02;01
Ian Bergman
Those are the two things bonding and planning that are so hard. How do you deal with those two?

00;49;02;03 - 00;49;21;27
Ron Green
It's tough. We we do a couple things on the bonding side. We have two team events a year, six months apart where we fly everybody in to Austin. We're based in Austin, Texas, we fly everybody in and we get together and we do several days of sort of team building stuff. That really helps because there's no substitution for being in person.

00;49;22;00 - 00;49;53;14
Ron Green
And then on the remote side, it's just about if there's anything working. Remote has helped me focus on it's communication. Written communication. So much stuff is in slack now, and so being able to think through how things can be misinterpreted or anticipating questions and addressing all of those initially really, really cuts down on the churn. But I'm telling you right now, I miss the days when I could say, oh, I just thought of something and step out of my office.

00;49;53;14 - 00;49;54;02
Ian Bergman
Yeah, I do.

00;49;54;02 - 00;50;00;25
Ron Green
To, you know, get the ear of 25 people, do a deep five minute dive, and we're done. And we just can't do that remotely.

00;50;00;25 - 00;50;21;10
Ian Bergman
No, it's it's schedule that structured. The context switching is tough. But that point that you made about documentation though I think it's I think it's mean such a hard discipline. It's one I really struggle with. But, you know, the CEO and founder of Automattic wrote a blog post a few years ago, maybe 5 or 6 years ago, that it was actually kind of foundational, I think, in terms of how I at least think about remote work.

00;50;21;10 - 00;50;22;11
Ian Bergman
And you said the exact same thing.

00;50;22;18 - 00;50;24;16
Ron Green
Okay, I'm going to go check that out, that.

00;50;24;22 - 00;50;35;08
Ian Bergman
You have it. Yeah, you really should. It was actually a little bit pressured. It was pre-COVID is a little bit prescient. Okay. So you're a tech guy Mac windows or something with the star and Nicks after it.

00;50;35;11 - 00;50;52;12
Ron Green
I'm a mac guy. I remember like it was yesterday when the original Macintosh came out in 84, and I, I just about fainted. I couldn't believe it. It was so amazing. And, I think it's got a it's got a, you know, a close, close place in my heart.

00;50;52;15 - 00;51;14;19
Ian Bergman
Man. I mean, I remember the Apple tunes and playing Sid Meier's Pirates, I think. Yeah, it was kind of the era I was in, but my first, my first computer was a centrist. 650 I kind of. Ooh, that was. Yeah, that was that would have been early mid 90s somewhere in there. And that was like, that was the first time when I had my own computer and I was like, I can do anything with this.

00;51;14;19 - 00;51;18;23
Ian Bergman
Yes, and I did. I coded some HyperCard games. Oh, you remember HyperCard?

00;51;18;23 - 00;51;20;15
Ron Green
Oh yeah, totally I did.

00;51;20;15 - 00;51;28;00
Ian Bergman
I put together a bunch of cards and workflows. I sold them on AOL. I think I made like $9 or something from my HyperCard shareware.

00;51;28;02 - 00;51;28;18
Ron Green
Yeah, but.

00;51;28;20 - 00;51;30;14
Ian Bergman
It's unreal what it opened up.

00;51;30;16 - 00;51;51;23
Ron Green
And the reason I was so crazy about the Macintosh was I had, I think it was a TI 99 and I would draw graphics, but to do it you had to like literally enter hexadecimal codes for like different bitmap parts of the screen. And, and it would take days. And then I saw the Mac and you just took a mouse and you click and drew and I was like, I couldn't believe that was even possible.

00;51;51;27 - 00;52;10;16
Ian Bergman
You know what I love about that story? Like it is, it is not every year that you get a use case that you know and you understand, and there is a truly like order of magnitude, step function, leap in capability to solve it. And that like we just tied this whole thing together, like your move from TI 99 to the Mac.

00;52;10;16 - 00;52;14;08
Ian Bergman
Yes. Right. Is I think, comparable to a lot of what's happening right now.

00;52;14;11 - 00;52;15;16
Ron Green
I totally, totally agree.

00;52;15;17 - 00;52;20;13
Ian Bergman
That's so much fun. All right, one more question. Dog. Cat or other.

00;52;20;15 - 00;52;27;26
Ron Green
Cat. Yeah, I, I like dogs, I love cats, I just think they're just kind of cool in a way that the dogs aren't.

00;52;27;29 - 00;52;50;05
Ian Bergman
They are cool. They are aloof. Yeah. Dogs are many things. I love them, they're not cool. Right. That's a really good that's a that's a great point. Awesome. All right. Thank you so much Ron for coming on. Innovators inside. There's been a fascinating discussion. Obviously you know we'll link out to your podcast. But for folks that want to stay in touch with you, with your company, where else should they go to follow you or you're on LinkedIn?

00;52;50;05 - 00;52;51;06
Ian Bergman
Where else should they look?

00;52;51;08 - 00;53;05;11
Ron Green
Yeah, they can find us on LinkedIn. Find me there. You can find me on Twitter. You can find us on Twitter, the podcast. And of course, if you want to get in touch with me, shoot me an email. Very simple on that. Ron are in a kung fu AI.

00;53;05;13 - 00;53;16;15
Ian Bergman
Amazing. Warren, thank you so much for coming on. The innovators inside have an amazing rest of your week, and I can't wait to hear kind of what new insight or new prototype you're coming up with next.

00;53;16;17 - 00;53;19;26
Ron Green
Oh thank you. And this was a blast. I really, really appreciate you having me on.

00;53;19;26 - 00;53;42;24
Ian Bergman
Awesome. Cheers. And that's a wrap for today's episode of Alchemists Dax Innovators Inside. Thanks for listening. If you found value in today's discussion, be sure to subscribe to our podcast and check out our segments on YouTube. Links and follow ups are in the show notes, and if you have questions you want us to feature in future episodes, email innovators at Alchemist accelerator.com.

00;53;42;26 - 00;53;47;16
Ian Bergman
Stay tuned for more insider stories and practical insights from leaders crafting our future.