If you were wondering how Microsoft feels about artificial intelligence (AI), CEO Satya Nadella is here to put your mind at ease. They enjoy it. In the company’s annual report, he writes a letter to shareholders praising artificial intelligence (AI). It’s easy to see why he thinks this is the most significant development in computer science in over a decade.
While Microsoft’s bottom line is healthy, the company may be idling its time away. Their countless product trials have failed to break into their target markets, and their efforts to break into the mobile, search, and hardware sectors have all stagnated or failed.
However, their cloud business is booming, so they’ve been tailoring the firm and its offerings to that success. As lucrative as it is. However, there is only so much innovation potential, and even that success is starting to seem stale.
They must have tracked industry developments for several years to determine which ones were beneficial enough to implement. Internet socialization? Not at all; far too much hassle. Fitness? Their infrastructure is easier to understand. Blockchain? Flawed and unnecessary. Metaverse? Super comical.
Microsoft waited patiently, bobbing up and down like a surfer in flat water. Then, the AI wave came up from underneath them, and they began to paddle desperately.
In his yearly message, Nadella says:
The next generation of AI will profoundly impact all areas of software and industry, including our own. Microsoft is still influential 48 years after it was founded because the corporation has successfully adapted to every major development in the computing landscape, from the personal computer and server to the World Wide Web and Internet to the cloud and mobile devices. We are at the forefront of this new period and are doing it again now.
After that, there are a couple dozen specific instances where AI is implemented throughout their divisions, products, and plans. Microsoft isn’t just dabbling here; they see it as the future of computing for consumers and businesses.
It’s not just a tool, like a silicon improvement that doubles the efficiency of data centers or a battery that doubles a phone’s life. It acts as a transformer, in a sense. Human-computer interfaces (such as keyboards, mice, and touch displays) have significantly shaped the history of computing. To see, hear, understand, and make sense of our intent and the world around us, we believe we have finally arrived at the next huge step forward—natural language—and will swiftly go beyond.
Imagine being at the helm of a huge computer business like Microsoft amid an upheaval of this size, and you can practically see the stars in his eyes. Microsoft has previously experimented with the notion of moving beyond the mouse and keyboard. Still, its natural language interfaces (like Cortana) and alternative technology (like HoloLens) have not progressed beyond parlor tricks.
However, they supported OpenAI, the industry leader in natural language processing, whether by luck or insight. The technology appears revolutionary, and the way the cookie has cracked, they are in a good position to give Google a black eye. Although having conceptualized AI’s enabling ideas in-house, Google has been caught off guard by the industry’s quick transformation. They are trying to recover, but the corporation has never been good at rallying behind a single idea, and this time may be no different.
Both Microsoft and OpenAI will benefit from this partnership. OpenAI has gained a customer or investor with practically unlimited money and a genuine interest in using AI technologies. By passing off the market-leading product as its own, Microsoft avoids the embarrassment of seeming, as it truly is, well behind the curve in AI development. Microsoft is probably training its foundation models in the background to protect against betrayal, but Nadella doesn’t acknowledge it since the partnership’s forward pace dwarfs its efforts.
Just picture that Microsoft was left out in the cold while Google benefited from a lucky arrangement with OpenAI. Microsoft would be in a far worse position than Google because it would have to rush to create LLMs that are just a quarter as good, and for every month it took Microsoft to catch up; Google gained another million users.
It’s no surprise that Microsoft is spending heavily to shore up its position and, if feasible, increase and deepen its collaboration with OpenAI.
But Nadella did sound a worrying note when he described the second of the two breakthroughs defining this era of AI: “the emergence of a powerful new reasoning engine.”
In the same way that you wouldn’t expect a calculator to reason when you tell it to multiply two integers, the current crop of AI models doesn’t “reason” about anything.
Nadella is neither a fool nor an ignoramus, of course. What he’s claiming is true; these systems do things that are essentially indistinguishable from thinking, and he knows it. It’s almost amazing to have a machine summarize a lengthy written document for you and have it do so in iambic pentameter when, until recently, only humans could do so.
It turns out that some reasoning problems may be reduced to statistical tasks since language exhibits predictable patterns. This is already impressive enough that we don’t need to add any extra pixie dust.
However, Microsoft’s use of such rhetoric reflects the somewhat unwarranted trust AI systems have inspired in its backers. They have a lot of potential, but they’re still in their early stages after only a few years. Although their capabilities will increase, we may not discover their limits until they have caused significant damage.
Overconfident and misinformed uses of the systems we have today pose the greatest hazards, as AI ethicists have repeatedly warned. This is not because some future catastrophes or imaginary systems displace whole sectors. Even if AI models cannot harm themselves, one CEO with starry eyes may damage the company.