Disclosure: Microsoft is a client of the author.
Microsoft is rolling out an update to Windows 11 that adds ChatGPT. Though it’s mostly a front-end for the Bing search engine at this point, eventually it could change how we interact with our computers. Think of it as another step in a long evolution that began with Microsoft Bob, then Clippy and Cortana, that is now on track to provide the capabilities those earlier efforts never did.
Evolving the PC from what it is — basically a terminal with localized processing — into a far more capable and comprehensive tool will be a significant change that we are only just starting to experience. PCs could evolve into something closer to a smartphone, which could lead one or the other to become obsolete, much as MP3 players largely died when smartphones gained the ability to play music.
From Bob to Cortana
I was at the Microsoft Bob launch; it was pretty sad. Microsoft got ahead of its skis and promoted Bob as the next user interface after Windows. But developers were not impressed, and the effort turned into one of the company’s biggest failures. The problem was three-fold. First, the technology wasn’t ready. Second, developers were not fully on board with GUIs over command lines. And the target audience for the product (who loved it) was largely ignored.
Bob was popular with many older folks intimidated by technology; they found it easier to work with Bob than Windows. But for most PC users, Bob was like regressing to training wheels. It was inefficient, somewhat painful to use, and fell far short of its promise.
Then came Clippy, an automated assistant that could help you get things done. It was a good idea, but again, the technology wasn’t ready and most users found it annoying. I tried the F1 Clippy alternative, and while I rarely used it as intended, I did like the little guy waving at me from time to time like a low maintenance companion or digital pet. I still miss it.
Next came Cortana, which should have been a game changer given how the game Halo had defined what Cortana should be. This led to dashed expectations because, if you played the game, your expectations were way beyond what Microsoft could deliver. Even though the company could have closed that gap with generative AI, it never did.
But generative AI, and ChatGPT in particular, have the potential to actually deliver on those earlier attempts; I think it will change the very nature of computers.
The generative AI PC
Generative AI is conversational by nature. PCs historically are command-driven. When working on a spreadsheet or writing something, you don’t have a discussion with the PC. You just tell it what to do. Generative AI can be collaborative, it can be a coach, and it can volunteer help (much as Clippy tried to do) when it sees you struggling. And once it learns your preferences, it can offer help in ways that don’t annoy you.
Over time, it can learn how you like to work, how you prefer to receive help and even where you most need that help.
As a technology, it lends itself to verbal interaction instead of typing (which will also force us to reconceive the office). Desktop PCs are constrained by the input methods of keyboard and mouse (and laptops are also constrained by their screens). Head-mounted displays are limited because working with a keyboard and a mouse often means you need to see at least one of them to assure hand placement. But if you’re having a conversation with your PC, do you really need a keyboard or mouse anymore? And if you don’t need them, doesn’t a PC become more like a smartphone with a head-mounted display?
Do we even need a display if we’re interacting with the computer by talking to it? Many people already interact vocally with computers. Just this morning, I asked Alexa about the weather, got a shipment update, and interacted with it on some programing; none of those tasks required a display.
And ChatGPT is much more capable than Alexa.
Two different paths?
I see the coming evolution taking two potential paths. You might have one device while mobile and another when stationary at home or at the office. The mobile one would likely be more smartphone-like, very portable, always connected and personal. The static device would be more like the Halo rendition of Cortana (with an interface that’s friendly and engaging), integrated with other technologies at home or work, and capable of moving with you as you rove.
This won’t happen overnight, but it could happen pretty fast — if the market decides existing hardware designs are out of date and non-viable compared to generative AI computers.
With generative AI, keyboards and mice become increasingly redundant, displays become more fluid in design, and the PC itself becomes more personal (and even more reliant on the cloud). This suggests to me that PCs and smartphones in the post-generative AI world will be as different as today’s PCs are from the punch card writers and readers used by early mainframes.
There is a strong probability that the company that gets this right first will own the market, much as Apple did with the iPod and iPhone. If you think generative AI is disruptive now, just wait. The real change and disruption are still in our future.
