When it comes to discussions of all things GPT, there’s an interesting divide between how tech professionals talk about it, and how leaders do. For the former, ChatGPT and GPT models are very exciting — it’s a moment of “Oh my god, isn’t this just the coolest thing ever?” Meanwhile, in the general populace, perhaps ten to twenty percent actually know what ChatGPT even is.

For leaders, the value of any invention is about translating it into something that provides real value to the customer. And through this lens, what the technology actually is doesn’t matter — we’ve gone through this cycle many times with innovations and how mature it is. At times like this, I like to reference Simon Wardley’s idea of value mapping for innovation.

When it comes to measuring how mature a technology solution is (and whether you can adopt it), you can divide the life cycle into Pioneers, Settlers, and Town Planners.

Pioneers are those who are at the forefront of creating new solutions. They’re the ones who are most excited about GPT models right now, screaming “Geez, we just created Fire 2.0!” Their minds are utterly full of possibility, stuff that most people can’t see. They’re creating custom-built technology, and they’re right on the frontier.

However, while pioneers show you wonder, they fail a lot, and this isn’t always great for providing value to the customer. Sometimes, a technology never reaches its full potential, but an interesting gimmick that’s just beyond a prototype. It usually has bugs or some other failings, but it’s arguably useful.

Settlers are the people who steal from the early adopters to scale those ideas into something workable, something a larger audience can use. They build trust and understanding. In short, they make the future actually happen. 

Settlers shift what the pioneers have done, and turn it from being something custom-built to an actual product or rental service, with most of the rough edges buffered out. At this stage, not everyone has adopted the technology, so it still provides a business edge, but it’s not completely fringe, either.

Town planners are the people who operationalize the idea, and at this point, the technology has become commoditized. At this stage, the technology is no longer novel, it’s expected, part of any business’s SOP. It’s a highly-trusted thing that can be used to make things faster, better, smaller, more efficient, more economical. The technology is part of SOP.

Once the town planners have the technology, it’s foundational. The pioneers then build upon this foundation to create the next innovation. Think of when the cart was invented, and then this led to the horse-drawn cart, then the car, then self-driving cars. 

Each group steals from each other. The town planners steal from the settlers, who steal from the pioneers, who build on the work of town planners.

An example of this process: Cloud computing

Cloud computing is a great example of the pioneer to town planner journey. Just over a decade ago, it was still very much frontier tech, something in the hands of the pioneers. It really took a while to become mainstream, and even now some people struggle with it.

But with cloud technology, it’s become expected, in the hands of the town planners. Now pioneers are building on top of it with things like serverless and Kubernetes. Serverless was a huge leap forward in cloud computing, redefining how you do things with event-driven architecture, a whole new way of thinking about the previous technology. And because it’s the new frontier, people are out evangelizing it, and not getting far with it.

ChatGPT and GPT-4 are still frontier tech

From a pure technology perspective, the technology is incredibly advanced. From a perspective of technology impacting profit and loss, it’s still very early. 

Technologists are very much still in the honeymoon phase with GPT-based technology. Everyone’s out there exploring it, trying to break it. People are doing things like showing things like taking a photo to GPT-4 of what’s in their refrigeration, and asking it to come up with food recipes in 60 seconds. 

That’s very cool, and very novel, and you can see where it’s going. But what is really going to get it on to the market, where our aunts and uncles and everyone else are going to see the benefits of it? What will it take to see it in really mainstream applications? And most likely when that happens, they won’t even know it’s ChatGPT or GPT-4, since it’ll be feathered into some other use case. They’ll just know they’re getting a really great experience all of a sudden.

There’s a pattern with emerging technologies, where companies that win focus on the customer and work backwards, looking for opportunities to leverage the tech that marries with that. If you do it the other way around, it doesn’t quite work as well. 

Right now, I’d say companies are waiting for technology based on GPT models like ChatGPT to mature, let the marketplace soften the edges of it, and make it easier to consume. The technology still feels a bit rough in terms of the way the APIs work, and what you can do. There is still a bit more time where an ecosystem needs to be formed around it.

For leaders, picking the sweet spot to adopt

That question on when to adopt, for leaders, is very real. When it comes to ChatGPT and GPT-4 applications, you don’t want to get in too early that you’ve actually got to pioneer the space yourself, unless you’re going to be an AI company. And if you want to be an AI company, and build the product, go right ahead! That’s going to be an AI company. But that’s a gold rush right now, and everyone’s going after that, and there’s going to be a lot of fallout.

You can already see some of the products that are trying to incorporate ChatGPT into their products so their customers can start consuming it right now, Microsoft being a prime example, to varying levels of success. For some of these companies, it seems like you want to slap the AI inside, to say you’re on the forefront.

On the other hand, you don’t want to be too late where you’re just another competitor to an overfilled space, and having the technology doesn’t provide a differentiator at all. The only way to tell that is to make sure you’ve got an eye on the space. 

Right now, I think the best thing leaders can do is be a student of the industry, get educated on what’s going on, and keep tabs on all the blogs. 

Before you start thinking about AI, check out your other tech fundamentals, first

There are a lot of folks who are getting wrangled up in things like ChatGPT, which is a bright and shiny thing. However, all day long we see companies who are struggling with getting the basic fundamentals of cloud computing as part of their culture and operating principles, let alone something like serverless.

One of the reasons for this is the inertia of legacy practices. In large enterprises it can be very hard to throw out the old and adopt the new. Cloud practices are here and now. The reality is as an organization if you don’t have the fundamentals of cloud computing down, you’ve got a long way to go before you should be playing with generative AI.

You should also make sure your organization has the agility and structure to handle the adoption, the necessary skill knowledge in your teams – and most importantly, leaders have the situational awareness to connect emerging technologies like ChatGPT to customer value.