The beginning of the end of applications as we know them.

How GPTs will become the predominant way you interact with computers.

Do you remember 2008 when the iPhone launched with an initial 500 apps in its store? As of this year, there are 2.24 Million Apps available with ~$155 Billion paid to developers (Apple taking their customary 30%, of course).

I believe in the years ahead, November 6^th will be viewed as an equally impactful moment in the evolution of computing. The day that the application as we know it changed forever.

So what happened? OpenAI announced at their inaugural Dev Days event a number of great cost savings, feature and API enhancements, and one MASSIVE capability that has me buzzing.

First, the quick updates (all of which are noteworthy in their own right).

-          GPT-4 Turbo with Vision: This is about giving the AI the ability to understand and interpret images, not just text. It can describe pictures, analyze their details, or read documents that include images.

-          DALL·E 3 Integration: Developers can now use the DALL·E 3 technology in their own apps. This means they can automatically create images or designs, like generating unique graphics for advertisements or products.

-          Text-to-Speech (TTS): This feature turns written text into realistic-sounding speech, useful for things like reading out text in different voices, making audiobooks, or assisting visually impaired users.

-          Price Reductions: The costs for using various OpenAI services have been reduced, making it more affordable for developers to use these advanced AI technologies in their projects.

Now for the biggy:

GPT’s – Roll your own Assistant

Open AI introduced the concept of “GPTs” (Generative Pre-trained Transformers). Tailored versions of ChatGPT for specific purposes. This allows both non-coders and developers alike to create special-purpose AI apps using GPT technology tailored to their specific needs. The “GPT” term is something you’re going to hear a lot more of, in the same way “there’s an app for that” became common vernacular.

So what are they?

In essence, GPTs are like having a highly capable digital personal assistant with highly specific domain knowledge. It can handle a wide range of tasks, from simple conversations to complex problem-solving.

-          They can remember previous parts of a conversation, so you don't have to repeat yourself.

-          They can write and run Python code, which is like being able to do complex calculations or solve tricky problems on the spot.

-          They can pull in information from outside sources, like grabbing data specific to a particular field or topic.

-          And on top of all that, they can TAKE ACTIONS by using existing APIs. That might include everything and anything you would do in an app on your desktop or phone today.

o   Processing an order or refund in a chat

o   Coordinating calendars and creating meetings

o   Tracking stock levels, reordering supplies, or analyzing sales trends.

o   Handling common queries, routing tickets, or providing after-hours support

Most interestingly, is that in many cases, you don’t need to be a software engineer to create one.

In fact, for one of my clients, I was able to take their entire publicly available SDK documentation and create a natural language bot to help their customers integrate with their platform. It took me literally 5 minutes.

90-second Assistant creation guide

If you’re a ChatGPT Plus subscriber and want to try it for yourself, you should have the ability right now. Simply go to ‘Explore’ in the top left of ChatGPT.

Here’s a simple example where I describe an idea I have, a Business Gift Suggestion Assistant, that I can provide either an exported file of LinkedIn contacts or specify a LinkedIn URL Profile. It comes up with gift suggestions along with a personalized message and even a customized card.

Think how long it would take and how expensive it would have been for someone to create an app like that on Android or iOS previously.

Want to try it yourself? No problem, because not only can you create them for yourself. You can share them with others. Check out my Business Gift Guru.

Where do things go from here?

At the end of this month, OpenAI is allowing developers to monetize GPTs based on how many people use them. Creators will be able to publish their GPTs on the GPT Store and earn a portion of the revenue generated by the usage of their GPTs. OpenAI CEO Sam Altman mentioned that they will be paying people who make the most used and most useful GPTs with a portion of their revenue. The exact payment scheme has not been fully articulated, but it is clear that creators will be compensated based on the usage of their GPTs.

If you thought the gig economy and side hustle culture was heating up, wait till you see every influencer and enterprising individual create their own Assistant.

And this ultimately is why I think we’re about to see the demise of the application as we know it. Today an app is a neatly encapsulated user interface that you’ve been trained to go to on your phone when you have a need in mind. Each app usually has a different interaction method, interface, and capabilities. The app developer has created a ‘point of view’ of what the user is trying to achieve with the app and competes to offer its services to the broadest possible audience by marketing, store placement, and word of mouth.

Personalization in recent years has gone some way to tailoring experiences, but ultimately you’re still working within the confines of the app developers’ constraints and motivations with the application.

What if you never needed to install an app again?

What if all you needed to do was express intent?

Rise of the Assistants

Today, interacting with an app is “the tax you pay” to achieve the outcome you’re seeking.

• I don’t want to spend 3 hrs comparing flights on Expedia for a European trip. I want the most affordable and convenient flights for my family to get to our destination.

• I don’t want to browse doctors on my health insurance website to see what expertise and experience they have. I want the best-damned doctor for my individual age & gender profile, medical history, and coverage

• I don’t want to learn about and wade through Turbo Tax to figure out what laws and advantages apply to me and my business. I want to file my taxes on time, accurately, and to my best benefit.

And that’s what’s coming. . . Monkeying around with hundreds of apps on your phone and computer is going to seem as archaic as the typewriter.

The assistant of the future will know you. In all actuality, it’ll likely know you better than maybe you know yourself.

It’ll know what you like, what you don’t like, who you work and socialize with. It’ll know what brands you prefer, where you travel through the world, and when. It’ll know what you own and where you can get the things you need or want. It will know how much money and other resources you have at your disposal. It will see, hear, and sense everything around you, both in the digital and physical world. It will work tirelessly to achieve each and every outcome that you ask of it.

The “point of view” will shift from that of the developer who created the app to your personalized assistant who will “compose” an outcome on the fly for the outcome you’re seeking to achieve. The results of that outcome will depend on your modality. If you’re driving in the car, you might hear it give an overview of the trip it’s booked in your ear on your morning commute. If you’re walking down the street with your AR glasses on, it might see and highlight a great new Indian restaurant that you might like. Your tax preparation might be sent to you as a document for your final review without you lifting a finger.

So what happens to all the existing online businesses and application providers? It’s adapt or die time. Their services are still absolutely crucial, but over time they will need to let go of “owning the user experience” or at a minimum, make it services available in parallel. And they will, because 8 billion agents buzzing around trying to fulfill their owner's needs programmatically is a lot more efficient than trying to grab your attention with a dumb banner ad and hope you click to install.

They will need to expose their capabilities to AI Assistants through Microservices. The user’s AI Assistant will seek out the best providers in real time (maybe based on cost, subscriptions the user has, or other preferences) and cobble together the optimum outcome based on the user’s intent. Each service provider participant in fulfilling that intent will receive appropriate attribution and their cut of the transaction.

A much more efficient model, and one that might see huge increases for the service providers who are first to adopt.

While I don’t think this is going to happen overnight. . . it might. For many companies, their services are already architected in such a way that instead of exposing their capabilities to just a Mobile or Web interface, they’ll just need to expose it *thoughtfully* to AI Assistants.

The big winners in this game are likely going to be the Siri’s, Alexa’s, Google Assistants, and Cortana’s of the big tech companies that will effectively be your trusted AI Assistant. You will “trust” them to act on your behalf, and the stickiness and brand loyalty they’ll drive will be based on an intimate attachment you establish with them by sharing every facet of your life and trusting them to navigate and broker effectively and safely.

They’ll offer assistants of different capabilities and different price points to consumers. They will also be the ones (along with the NVIDIA’s of the world) who will benefit from all the service providers who are utilizing these services on their cloud computing platforms.

I don’t think this is at all far-fetched, and I think Nov 6th is the start of it all. We’ll see in the weeks and months ahead how quickly we’ll move to this new paradigm. While there will likely be some hurdles along the way, and clearly a need for better rails for privacy, legal, safety, and other societal concerns, I think this is an inevitability, and I think it’s going to happen quickly.

In my day job, I’m already helping clients prepare for this new era, and together we’re learning, experimenting, and already seeing very real and impactful optimizations, returns, and new opportunities.

If any of this resonates, or you’d like just to learn more, I’d love to hear from you.

Thanks for taking the time to read, and let’s look forward to a world of magic where your intent is manifested at the speed of thought, or at least, for now, by how quickly you can chat with your AI Assistant.

Looking for more?

Check out some recent brief notes from the last couple of weeks.