IBM and Microsoft were once close partners. They divorced badly, but then eventually came to the realization that collaboration was important. Now they’re collaborating again. But Microsoft, which stunned the market by licensing OpenAI, and IBM which created the market with Watson and now watsonx, are approaching the market very differently. These differences are important.
With over a century in the business, IBM realized early on that ethics and quality matter. While it has moved more slowly than Microsoft, its products are known for their extremely high quality, security and reliability. Conversely, Microsoft is all about speed and cost, but as Boeing recently demonstrated, a focus on speed can have a catastrophic impact on quality and financial viability.
Let’s talk about the balance, or lack thereof, between speed and quality.
OpenAI
At the heart of this difference is how Microsoft was able to move so quickly into a market that IBM dominated for years. Microsoft used a similar strategy to the one it used to move aggressively against Netscape Navigator at the birth of the Internet. In that case, Microsoft licensed the Spyglass browser and then evolved that platform into Internet Explorer. Due to a mix of good execution by Microsoft and horrid execution by Netscape, this resulted in Microsoft owning the browser market (which it later lost to Google, but that is another story).
OpenAI is a mess. Its board was concerned that CEO Sam Altman had gone off the rails, so they fired him badly, resulting in that decision being reversed and the board (at least those that wanted him out) leaving. They were replaced by people more loyal to Altman.
OpenAI’s AI had serious quality issues, and Altman, repeating a common theme of older CEOs, fell in love with a movie character and attempted to recreate it without the permission of the actor who provided the voice to the character. Stealing a voice, particularly a woman’s voice in the era of Me Too, is extremely problematic. Putting the company at risk to fulfill a personal need highlights a tendency to abuse power and is remarkably like what Mark Hurd did at HP and resulted in Hurd’s termination.
No one has much control over Altman, including Microsoft. This puts a cloud over the technology and suggests that future Altman decisions could harm both OpenAI’s and Microsoft’s efforts.
Boeing and Why Focus on Quality Is Important
Boeing’s recent problems are a case in point. Before acquiring McDonnell Douglas, Boeing’s quality led the aerospace industry. McDonnell Douglas focused on speed and price, which resulted in cheaper airplanes that tended to have serious problems. Those problems devalued the company and allowed Boeing to buy it.
But then McDonnell Douglas’s practices infected Boeing which now has deadly problems. This has put Boeing at risk of failing and resulted in a lot of resignations and the degradation of Boeing’s brand. Boeing is one of Microsoft’s largest and most important customers.
Should OpenAI’s problems compromise Boeing, the result could be catastrophic.
Quality vs. Speed
Let’s look at what happens if you speed up a process like coding without first addressing quality. Let’s say you create ten things an hour and have a 10% failure rate. That’s one failure an hour and is easily addressed. But if you speed up the process, as manufacturing companies tend to do, to 10,000 things an hour, you have 1,000 failures an hour. That will overwhelm quality control and a lot of the defective parts will make it to market.
This is why you want to focus on quality before speed. Whether we’re talking about building aircraft or creating code used to manage a city or a massive power plant, you want to make sure the problems created fall within your ability to deal with them.
IBM’s focus on quality and security before speed better assures a low failure rate result. This is particularly important if you are writing other AIs, because a quality problem could explode if it is pushed into other AIs with critical responsibilities.
Wrapping Up: The Industry Is Not Focusing Enough on Quality
Much of what we see in the industry is a massive focus on speed. Microsoft is hardly alone in this focus. It’s widespread across the industry. IBM is the standout exception for aggressively focusing on quality, reliability and security. It’s in the company’s DNA. Microsoft is more focused on speed because that is in its DNA.
Microsoft does have quality checks in its solution, but since it doesn’t control OpenAI, it’s dealing with the symptoms and is not in an acceptable position to address OpenAI’s quality issues at their source.
One solution might be to use Microsoft’s Copilot for creation and IBM’s watsonx for quality control since quality control should be separate from the creation process. Another would be to use Copilot for things where quality is not critical but use watsonx where it is.
Regardless, looking at the advantages and risks of both approaches can better inform you of which tool works best with each project. No tool is good at everything, so knowing which tool is best is a good deal of what differentiates a novice from an expert. Given how fast AI is moving, you need to become an expert fast.
