A-I-A-I-IO (Premium)

During the Google I/O keynote this past week, CEO Sundar Pichai said “AI” almost 30 times, while other executives said the term over 110 times. That’s more than one “AI” for every minute of the keynote, so I assume it’s obvious that Google is all-in on AI.

But here’s the thing: Google has been all-in on AI for years, and it’s kind of amazing to me how the press so quickly picked up on internal Google complaints that the firm was caught flat-footed by Microsoft’s sudden AI-based Bing chatbot announcement this past February. (Though I admit that its hastily assembled first response was poorly done.) It’s not fair to pretend that Google is “behind” in AI when it is Microsoft, not Google, that is behind. After all, Google’s AI prowess was all created internally, while the most impressive parts of Microsoft’s work came from an outside company, OpenAI.

Regardless of your take on that, the logistics of this competition aren’t all that important: there are only three tech firms that are big enough and have the right assets in place to make this kind of AI play, and Google and Microsoft are two of them. (The other, of course, is Amazon.) And an honest (if early) assessment of this market suggests two things. First, that AI will indeed change everything. And two, that it most likely will not upset any of the major players and their relative market positions.

The “AI will change everything bit” is not controversial: all you have to do is watch a Google or Microsoft demo to see how AI will fulfill the very promise of personal computing, which was that we would have these inexpensive and easy-to-use tools that would help us get work done more efficiently. This is an important distinction that gets lost in the arguments about whether AI (or any technology from any era) will result in a loss of jobs. Of course it will. But we will likewise look back on jobs people had in some bygone time and wonder why those were jobs at all. The real goal here isn’t for AI or some other computer-based construct to replace us, it’s to help us as people achieve our own goals more easily. This is the double-edged sword of progress.

I’ve told a few stories that illustrate the tension of progress.

The first you will surely have heard: I hosted a pre-con session at Microsoft TechEd years ago in which system administrators and IT pros were alarmed by the push to cloud computing. “Do you mean to tell me that my last job as an Exchange administrator will be to migrate my company’s email over to [what is now Microsoft 365]?” one frustrated person asked me. And the answer is as obvious now as it was then: “Yes.” That company didn’t exist to manage an email infrastructure, the email infrastructure existed to serve the needs of the company, and if there is a cheaper, better, and more efficient way to do that, the responsible thing to do as an employee of that company is to recommend doing so and then assist in the transition. (And with regards to that gentleman’s job, come on. We all got into technology because it was dynamic, exciting, and ever-evolving. But sometimes we get soft, stop learning, and make the wrong decisions for one’s employers, which might include fighting to keep email in-house because that’s your job, ignoring what’s best for the company.)

The second came up this past week on Windows Weekly when we were discussing device sales. I noted that I cover PC sales every quarter, but that every year I write a comparison of that year’s PC sales and growth/loss with the previous 10 or more years. This write-up requires me to find an Excel spreadsheet that I make a copy of every January to update it for the new year’s sales numbers. And this activity requires me to learn how to use Excel again as if it were the first time I had ever used the product, because I just don’t use Excel all that often. And this task, adding one year’s worth of data to two columns of numbers and regenerating a colorful chart to make the comparison more visual, is an excellent example of how AI can make life easier. I’m a writer, not a number cruncher, and I shouldn’t have to waste time learning how to use a tool that I basically only use once per year just so I can get something done. Nor should I have to pay someone else to do this thing for me: AI can just do it because I can easily explain what I’m trying to accomplish. Here, AI—technology—is clearly and literally making my life easier and more efficient. This is what it’s supposed to do. (As discussed in AI Art (Premium), image generation can be used similarly, and there are of course the same concerns about taking a job away from others.)

My argument about AI not changing the relative market positions of the major players in Big Tech, but it is also more speculative because we don’t have as much prior experience to fall back on. But let’s work through this.

AI might be most correctly thought of as an amplifier, by which I mean that AI isn’t some new product line, service, or market, but rather an ingredient that will be applied almost literally everywhere, like adding salt to a meal to amplify the flavor.

For example, Microsoft is adding AI to Bing search because Bing search is an also-ran and its decades-long efforts to compete with Google toe-to-toe on features has failed. So Microsoft hopes that AI will amplify Bing. In this case, not just functionally but also with perception, that Bing will transition from a laughingstock to something that people respect and might choose to use.

There is precedence for this. Despite its incredible market power and consistent growth, Microsoft’s stock price flatlined for 10 years before the company shifted gears and announced its intention to become a major player in cloud computing. This transition resulted in the rise of what’s now called Microsoft Azure, and its heady growth numbers over several years—which was possible because the business started out so small—excited Wall Street, goosing Microsoft’s stock and propelling it to its current status as the world’s second-biggest company by market capitalization, a number that is achieved by explosive stock price growth. Microsoft’s power today was built almost solely on perception, and it wasn’t until this past year that its cloud computing promises were matched by reality: by my estimates, over half of Microsoft’s revenues now come from cloud computing. (That Azure growth has slowed dramatically in recent years is, of course, not coincidental.)

While it’s clear that AI will help grow Bing’s usage share, I don’t believe that Bing will ever seriously rival Google Search. First, it’s a bad brand, something that’s the butt of jokes, like Zune. Second, it’s not a very good search engine, and most who do try it switch back to Google Search just as quickly as they use Microsoft Edge once to install Google Chrome. From an AI perspective, what people really want is an AI-infused Google Search, and that is exactly what Google is going to deliver. And so we’ll see this brief moment of positive press for Bing followed by the inevitable discontent because, again, it’s still just Bing.

So, yes, Microsoft tweaked Google, something none of us expected—meaning, were are all caught flat-footed—but there is always payback for poking the bear. And that payback will be that Google adjusts. And it will adjust quickly because what Microsoft accomplished is a capability that Google already has. It doesn’t have to create something like the Bing Chatbot from scratch. Worse, for Microsoft, now that it has poked Google, the online giant will unleash a tsunami of features that will keep users in its ecosystem, features that Microsoft cannot match, like the improvements coming to Google Maps. Nothing Microsoft did will upset the relative positions of these two companies.

This works in the other direction, too. During Google I/O, Google showed off how it would add AI capabilities to the productivity apps in Workspace, and to Gmail. And I’m sure these improvements will be excellent, and in keeping with my notion that AI amplifies existing technology, making it more useful. But that part of the keynote was less compelling because Microsoft already dominates this market, and this is where its unique strengths will ensure that nothing Google does will shake up this part of the market either.

The secret sauce to Microsoft’s prowess with productivity apps and businesses is something called the Microsoft Graph. And like Google’s longstanding AI work, this is not a capability that suddenly came about as a response to anything, it’s work that Microsoft has been doing for years. The Microsoft Graph is a way to bridge data silos and present cross-company information when and where it’s needed in useful ways. It might be seen as a type of glue or connective tissue that doesn’t just tie together disparate sources but also finds the connections that matter. It’s like an Internet of data that is private to a company, a way for employees to find others within the company with specific skills, provide information about other people in a meeting, figure out schedules for any sized group of coworkers, analyze data intelligently to make decisions, and much more.

The Microsoft Graph is a competitive advantage for Microsoft, something that sits behind its dominant Office and Teams products and nascent contenders like Loop. It is in its own way an amplifier, and the addition of AI to the mix will only make this thing more powerful and more useful to Microsoft’s customers. This will ensure that businesses stay put.

The very nature of AI ensures that it will benefit all comers, which is why I don’t feel that the relative positions of each Big Tech player will change. Nor will it help too many smaller players enter the big leagues and compete more effectively than the dominant players today. If all businesses have AI, then all businesses benefit, and nothing much changes relative to each other. There will be some ups and downs. And then everything settles back into place.

The competing narrative here is that a smaller player like OpenAI acts as a disruptor and changes the market forever. But there is no market here. OpenAI will likely see huge success. But huge success for OpenAI probably won’t come from user-facing services like DALL-E. It will come from funneling its technology into paying partners both big and small. OpenAI, like AI, is basically just plumbing. This company isn’t going to create the next big search engine, let alone complementary services like email, calendar, and maps.

That’s my take, and we’ll see what happens. But I wouldn’t count out Google in search just like I wouldn’t count out Microsoft in productivity. Dominant players gonna dominate.