AI is already pretty omnipresent. We're talking about AI models and code generation and image creation and chatbots everyday.
But what if people stop thinking about interacting with AI software, but rather take for granted that it powers a lot of what they see everyday?
Computing went through a similar change, where going from being a product to being foundational infrastructure. I think AI could go through a change like that.
From end product to underlying infrastructure
A possible next step for AI development is, instead of individual people accessing a software model, to leverage AI in novel ways to power other systems.
An enterprise doesn't need an actual person asking questions to an AI chatbot; they could have AI analyze and digest reams of data on the fly to find patterns relevant to their operations.
Imagine an industrial system with:
SaaS
IoT sensors
Equipment
Connected assets
Telemetry data
Monitoring systems
AI analytics
That's just an example, but AI models aren't necessarily end products here - they're powering insights about a whole ecosystem.
That's why I think these physical examples are exciting: the data sources in the physical world are enormous, and change constantly.
The potential for AI to analyze, understand, and predict this data at scale has applications in areas as broad as tracking inventory, or equipment or vehicle locations, or monitoring for issues.
It's useful beyond just the AI product itself, and that's valuable.
This is why I feel that venture building in industrial technology can have value in physical computing, and why Aperture Venture Studio looks at the space as interesting. These aren't just edge cases of the AI + internet revolution either.
The technical side isn't as glamorous as a new AI model, but the work required to power these use-cases touches on everything from the actual model to data collection and normalization, integration capabilities and more. It's one thing to have a flashy AI chatbot, it's another to have that same model actually do something useful for a client's needs.
The Other Interesting Side Of AI In The Physical World
The other interesting aspect is one I hinted at earlier: if we get to the point of AI being ubiquitous in the digital realm as well, creating text or images becomes less impressive - and thus, human work becomes more valuable.
There's only so much value in someone simply generating text, images or code - it's just a process that can be automated. What becomes more valuable is the judgment and experience, the taste and craft, the personality and nuance that only people can bring. It doesn't mean it'll totally replace traditional art work, but it'll evolve.
That's all to say that the AI era likely won't be choosing between "computers do everything possible" or "people do everything possible." In some situations, people are better, and in others, computers are. It's valuable in both roles.
I think that's one way to approach this space: not asking the questions of "how cool can we make AI," but instead examining where computers and AI can augment, empower, and even enable human work.
As always, the actual product depends on the problem being solved, but a product is ultimately a way for a company to solve a problem at scale.
It's not just about the software
The Other (Less Intriguing) Question You Might Ask About AI
Probably the most common concern about the future of AI is not about its capabilities being exceeded by a large language model or anything like that.
If anything, the opposite: that we get to a point where there's literally nothing left to do with AI. So of course we will have made that a thing.
I think the much more intruiging question is about how it'll revolutionize industries and change how we work beyond just, "it can replace chatbots and code writing."
Because when most people look at AI, they see something like this: neat, but not necessarily an end-all-be-all revolution.
Computing power became cheaper, and we still needed to attach it to something people needed.
That's not to knock on AI or its capabilities or anything like that, just reflecting on how AI development is likely to go for the foreseeable future.
There's literally nothing to do with AI left, and so we find new ways to do things.
So as much as always as I think the question, "how smart will AI get?" is fun to ponder, I'm personally not as interested in that as I am how it could change the systems, companies, and objects we interact with daily.
Like, how far can we take this?
Top comments (0)