the X moment for AI–solve for X
One of the analogies people all over have been using to try and distil what this moment feels like for AI is ‘the X moment for AI’. This is “the iPhone moment for AI“, or “the Web/internet moment for AI”. Bill Gates speaks of AI in as analogous to when he very first saw a GUI.
I have my thoughts about the best analogy I’ll return to in a moment, but as someone once said about models, all analogies are wrong, but some are useful. What’s interesting perhaps is not so much why they are wrong, but how they are.
What are people trying to get at when they use an analogy like this? I think on one level it’s very simplistic. The Web and iPhon are technologies that have become seemingly ubiquitous (the Web took several to get any sort of significant adoption let alone ubiquity, the iPhone somewhat less). An AI is a technology that has arrived and seems to be on its way to ubiquity.
But that’s such a banal observation as to be essentially useless.
Can these analogies be more useful than that?
Well, the Web and iPhone are platforms, not (just products or technologies). Their importance, particularly in the case of the Web is what they enable others to do–not only their users but developers and publishers and other creative people and organisations.
Does it make sense to think of AI in similar terms? To date, perhaps not so much. AI has been a bespoke technology incorporated into individual products, from self driving vehicles to expert systems. A competitive advantage of individual organisations.
But with generative AI, this idea of AI as a platform makes more sense. The same underlying technologies are enabling a broad array of products via APIs.
Which suggests the different nature of the opportunity in the space–AI enabled products.
What we’re seeing right now is for the most part existing products and solutions incorporating generative AI, large language models and the like, often simply by plonking a chatbot into an existing UI.
What comes next? Certainly a lot of UI experimentation, as these interfaces move beyond simply text based chat. We’re already seeing that, with Adobe’s Firefly being perhaps the most mainstream example.
But beyond that, if generative AI is to be genuine platform, we’ll see AI-native applications, and business models that emerge unique to these.
Perhaps generative AI is a sustaining innovation, something that makes existing products and businesses better–makes tasks easier, less time consuming, less expensive.
But maybe this is the silicon chip moment for AI, something even more foundational than the iPhone or even the web. The silicon chip was profoundly transformational, allowing computing power to be embedded ultimately just about everywhere. The transistor had an impact none of those involved in its invention or even first decades of development could begin to imagine.
And I ave a feeling the same will be true of generative AI.
Great reading, every weekend.
We round up the best writing about the web and send it your way each Friday.