23 May How Microsoft’s AI Gamble Could Backfire (Again)
Microsoft’s history doesn’t inspire a lot of confidence.
| 5 min read
Microsoft is racing to lay the groundwork for generative AI. It’s the first in this space, but history suggests that there’s a huge risk of failure.
As expected, this year’s Microsoft Build conference is basically a sales pitch for AI. Microsoft spent the first day of this conference explaining how its Copilot AI “personal assistant” will revolutionize software development, corporate workplaces, and the average person’s Windows PC.
But a certain attitude underpinned this conference—Microsoft knows that it’s the first to truly embrace AI. It’s laying the groundwork for competitors like Apple and Google, but it’s also racing to reach customers before anybody else. And for those who have witnessed Microsoft’s many successes and failures, it isn’t hard to imagine how this gamble might explode in Microsoft’s face.
As Always, Microsoft Wants to Be First
Over the last few decades, Apple has gained a reputation for arriving “late.” It ignores new technologies for a few years. Then, it swoops in with a “perfect” product and hardcore advertising, effectively signaling that the new technology is ready for prime time.
But Microsoft tends to go the opposite direction. It likes to be the first to offer groundbreaking ideas or technologies. Microsoft obsessed over smart homes in the 90s, it predicted the future of wearable devices (particularly smartwatches) with its SPOT initiative, and it launched its first touchscreen Surface device (a coffee table, oddly) before the iPad’s debut. Plus, Microsoft popularized hands-free calling in cars, and it’s spent decades dabbling with ARM-powered PCs.
Many of these “firsts” are success stories. But there are just as many examples of failure. Microsoft is not a leader in smart homes, its smartwatch tech is forgotten, and three years after the launch of Apple Silicon, Microsoft still doesn’t sell a compelling ARM-based PC. Even the Surface lineup, which is incredibly successful and influential, has failed to meet its initial goal of offering a proper tablet experience (which is arguably a good thing—Windows 8 and Surface RT were crappy products).
Microsoft’s headfirst dive into AI could be a huge success for the company. And, based on the short-term stock market gains, Microsoft would be foolish to pause its AI rollout. But history shows that this “first” could end in failure. It all depends on Microsoft’s AI implementation, and of course, its luck.
Windows Users Are Often Microsoft’s Worst Enemy
Even when you ignore moral or legal arguments, Microsoft’s obsession with AI is very divisive. There are clear benefits to this technology—it can automate tasks, answer questions, and so on. But Windows users don’t like to have things shoved down their throats.
One of the only consumer-focused announcements at Build 2023 was Copilot, a “personal assistant” that runs on Bing AI (Chat-GPT) and pops up on the side of the Windows operating system. The idea here is pretty simple. Instead of opening an AI chatbot website, you can pull up Copilot at any time. It can also access your filesystem, meaning that it can organize documents, locate and share photos for you, or find files that you’ve misplaced.
Within minutes of this announcement, I saw Windows users complaining about Copilot. And some of their complaints are justified—Bing AI is a blatantly unfinished product, average users may put too much trust into the AI (leading to misinformation and reduced productivity), and because Copilot requires cloud-based processing, it could result in a massive reduction of user privacy and security.
Users also note that Copilot is “infantilizing,” sort of like a modern version of Clippy. Instead of making its products more user-friendly and actively fixing problems in Windows, Microsoft is treating this “personal assistant” as a Band-Aid. Personally, I don’t share this opinion, but it’s an important example of how Windows users hate being told what’s best for them.
Maybe these users are overreacting, and maybe they aren’t. After all, Microsoft has slowly brought back some of its worst habits—it’s filling the Edge browser with bloatware, breaking PCs with unnecessary widgets, bullying potential customers, and shoving ads into the Windows operating system. If Copilot isn’t useful to the majority of Windows users, these users will consider it an annoyance.
The problem, of course, is that natural language AI is a genuinely useful tool. We just aren’t sure how to use it outside of a few circumstances. If Windows users treat Copilot like Cortana, other tech companies will have the chance to overtake this race, and they’ll have the benefit of observing Microsoft’s example.
The Infamous Microsoft Victory Lap
When Microsoft is excited about a product or concept, it tends to get a bit self-congratulatory. The company has a habit of patting itself on the back and donning an “in your face” attitude before the race is finished. There’s Steve Ballmer’s infamous “developers” chant, plus the embarrassing attempt at “dancing” during the Windows 95 launch, but there are also several smaller examples.
One thing that came to my mind during the Build 2023 conference was Microsoft’s reaction to the iPhone. Then-CEO Steve Ballmer laughed at the iPhone and explained that Microsoft preferred its strategy in the cellphone market, which (at that time) mirrored its approach in other areas. Essentially, Microsoft believed that it could take control of an existing industry’s software market (such as the automotive industry) by getting all of the big corporations stuck on groundbreaking, Microsoft-branded software.
The first day of Microsoft’s Build 2023 didn’t really feel like a developer conference. Instead, it felt like a software sales pitch to corporations, which can use Copilot and other tools to increase productivity. This is understandable, but it’s also a sign that Microsoft may be overly-focused on the successful nature of its “process.” Ironically, Microsoft repeatedly mentioned how developers will define how we use AI tools—something that should probably be ironed out before the sales pitch and rollout.
And Microsoft did a bit of self-aggrandizing, too. The company wants everyone to know that it’s announcing “50 new products,” something that sounds impressive because it’s completely impractical from a development or journalistic standpoint. And, at one point, Microsoft paused to talk about how it loves to “do legendary sh**.” (Censored by us, not Microsoft. I’d give you the timestamped video, but the conference is still live at the time of writing.)
Admittedly, this style of “in your face” attitude is way more tolerable than watching a grown man flail, sweat, and scream in front of an auditorium full of people. But nonetheless, it’s a victory lap at the beginning of the race, which is something that should never inspire confidence.
Even a Failure Can Become a Success Story
I’m not trying to predict Microsoft’s downfall. Really, I hope that Copilot and the Bing AI live up to all of Microsoft’s promises. And if Windows users treat Copilot like Cortana, it won’t be the end of the world for Microsoft. If anything, it’ll be a sign that Microsoft was too early, and competitors will use Microsoft’s failure as a learning experience.
Plus, Microsoft’s failures aren’t always a permanent thing. The company crashed and burned when it went all-in on tablets—Windows 8 was a disaster at launch, the Surface RT proved that Windows couldn’t run on a power-efficient ARM chipset, and so on. But out of this failure came the extremely successful Surface 2-in-1 concept, which is now a staple of the laptop industry and has clearly influenced the iPad’s development.
We aren’t sure what’ll happen next. All we know is that Microsoft is laying important and risky groundwork that will accelerate AI development. Once the smoke clears, Microsoft’s successes and mistakes in this area will dictate the future of AI implementation. So, even if everything falls apart, it’s not a real loss.