Assimilated AI

Beyond the novelty

Generative AI is an unprecedented technical innovation. Whereas other past technologies such as electricity, motors, internet, etc. all required specialized knowledge for working with early iterations, Generative AI looks and feels like a usable product, out of the gate.

In a previous article, I have highlighted the many uses and disruptive capabilities of Generative Artificial Intelligence (Generative AI) in generating original text, image, audio, and video content.

These unique capabilities are exactly why we can strap a simple chatbot user interface in front of Generative AI, and voila! Anyone with a basic command of natural language can start interacting with it right away.

Having easy access to Generative AI, and its conversational nature, has led us to the false thinking that these AI Models are finished digital products, and with just a bit of prompt engineering training, everyone should start using them as-is.

There is a lot of adoption happening right now with Generative AI. But, let's not forget that these are ‘self selecting’ early adopters, who are typically very forgiving of any shortcomings in the initial stages of new innovation.

Even in my personal experience, I have implicitly learned how to work around the inaccuracies and hallucinations of the Generative AI models I use frequently. I accept there will be errors in the AI's responses and keep refining my prompts, knowing that further iterations of the models will improve in accuracy, over time.

To gain traction beyond the first wave of early users, new technologies have to demonstrate reliable features that consistently perform as expected, before being eventually assimilated into everyday products.

After growing rapidly for the first 6 months, user visits to Generative AI sites, and even ChatGPT, have recently started to plateau. The next wave of adoption will require something more substantive than just the novelty of chatting with a humanlike bot.

For Generative AI to gain sticky adoption, the underlying Foundation Models need to be purposefully assimilated into the features of our everyday digital tools and applications.

Who will drive Generative AI assimilation?

In the enterprise, my bet is on Microsoft. With the most widespread installed base of operating systems, workplace productivity apps and a partnership with OpenAI, they have the lead in the business world.

Now Microsoft has a history of being clunky and buggy in early versions of products - remember Azure vs. AWS in the early days? or video calls on Zoom vs. Microsoft Teams a few years ago? - but eventually, Microsoft does get it right.

Microsoft's Generative AI Copilot is no exception. Copilot's features, built on OpenAI's models, have been quickly plastered over all Microsoft apps, but feedback on actual utility has been lukewarm.

That said, no one is giving up on their daily office work in Microsoft Windows and Office 365 apps, so the iterations of putting Generative AI into product features will continue. Eventually, some features will find a product-market fit and stick.

Microsoft is also experimenting with Neural Processing Units (NPUs) to introduce PCs that will have Copilot and AI embedded directly into the machine’s hardware.

In consumer technology, my bet is on Apple. In keeping with their company ethos, Apple will take a human and design-centric approach, to seamlessly incorporate Generative AI into their products.

Apple's challenge in the past has been the closed nature of its products and technology ecosystem. However, when it comes to Generative AI technology, Apple knows that it's playing catch-up and needs a different approach.

In less than a year, Apple has quickly pivoted to using the same playbook as Microsoft, first announcing a partnership with OpenAI in June 2024, to most recently following the same partnership roadmap and now becoming an investor in OpenAI.

With OpenAI partnership under its belt, Apple is already moving forward with introducing new Generative AI features.

If anyone can hide technical complexity and bring delightful user experiences to life, it is Apple. From the beginning, Apple is taking a unique design-first approach to assimilating Generative AI into their products.

It won’t be all on just Microsoft and Apple...

Remember that with most general-purpose technologies (electricity, automobiles, motors, etc.), for every popular mainstream product using the technology, there is also a marketplace for tons of other specialized and niche products built on the same technology.

So dear builders and founders, please keep playing around and tinkering with Generative AI!

Maybe your experiments with Generative AI will lead to building something cool and unique...and even if it doesn't, at least the learning process will help you unpack how this innovative technology works, demystify it and perhaps even discover novel use cases.

MAIN TAKEAWAYS

1. Broader adoption of Generative AI will happen through assimilation into mainstream products.

2. As a general-purpose technology, Generative AI is unprecedented because of its mass availability and accessibility.

3. It behooves all of us to keep learning, testing and experimenting with this powerful new technology.