There is a phrase often falsely attributed to Winston Churchill (it came from George Santayana’s book, The Life of Reason) that goes: “Those who cannot remember the past are condemned to repeat it.” It came to mind while attending Intel Vision this week.
I entered the last huge technology wave soon after graduating from college in 1977 when Apple II launched. I’d just finished learning how to code in Fortran 4, COBOL, and APL, only to find all three languages had become obsolete. Thirteen years later, I was working at IBM when the wheels came off that cart driven by both internal mistakes and the rise of the PC that IBM drove (it was even called the IBM PC). A few years later, Windows 95 came out and set the stage for the collapse of a whole bunch of computer companies. Nearly 10 years later, IBM exited the PC business, having gone from pretty much owning the world’s computer business to becoming a bit player. Looking back, we should have defended the mainframe (which is, even today, one of IBM’s most lucrative and unique offerings) instead of throwing it under the bus.
Revolutions have consequences. The AI revolution is not only potentially much bigger than the PC server revolution, but it is also moving much faster. It feels like the advancements in AI are turning the old revolution’s years into months, and it’s still accelerating. This doesn’t mean Microsoft and Intel are slacking. On the contrary, both are working incredibly hard and are doing amazing things to advance their AI efforts, but I’m seeing some troubling similarities between what was done in the 80s and 90s by IBM and what Microsoft, Intel, and others are doing now. And neither company is anticipating its efforts will go off the rails like IBM’s did.
Let me explain.
IBM DOS – Microsoft ChatGPT
What most people don’t recall is that IBM had an OS that was internal and that those in IBM believed was better than Microsoft DOS. But these folks weren’t consulted. Instead of using its own OS, IBM licensed one that Microsoft had recently bought. Microsoft and IBM came from different ages, and the two teams quickly grew to distrust each other, leading to a dramatic separation just a few years later over OS/2. Microsoft had concluded, accurately I think, that IBM was pretty clueless as to what the market wanted.
Today, Microsoft has licensed its first AI effort from OpenAI, which is actually a worse partner to Microsoft than Microsoft was to IBM. The firm is attracting an amazing amount of litigation, its CEO is actively working with Jony Ive to create an alternative to the Windows PC, and I’ll bet that OpenAI folks are kind of thinking about Microsoft like Microsoft thought of IBM because they rightly think they know more about AI than Microsoft does.
In hindsight, IBM should have either used its own OS or bought Microsoft instead of licensing DOS. Granted, that likely would have made Apple more powerful, but Apple had fired Steve Jobs and was going through its own issues. There was no other powerful alternative company, suggesting that Dell and other companies like it wouldn’t have emerged because licensing from a company you were competing with (like IBM) just wasn’t done back then, at least not to the level that they did from Microsoft (pre-Surface).
Given IBM’s backing of OS/2 and that Windows NT was a clean-room version of OS/2, this would have resulted in the Windows 2000/XP wave happening around five years earlier. With Louis Gerstner’s focus on marketing, IBM would have retained far more of its power and wouldn’t have suffered the catastrophic brand damage it did. Regardless of the outcome, it would have slowed down the change and allowed IBM to evolve more successfully to address that change.
Microsoft vs. Netscape
When Netscape came to market, Bill Gates recognized the risk it represented and pivoted the company quickly to address it. First, he bought the Spyglass browser and then created Internet Explorer. Microsoft did some questionable things, but I watched the resulting trials. The things Microsoft did didn’t work (though attempting them alone was illegal and did get them into trouble), but it won because Netscape was poorly managed. Google was also a small company that took on Microsoft, but it wasn’t poorly managed, so it prevailed; granted, Bill Gates was gone by then, but when the battle first started, like with Netscape, Google wasn’t in Microsoft’s league. It is now, and no amount of throwing chairs by Microsoft’s then-CEO changed that (sorry, Steve).
This time, Microsoft is doing a lot of things right, like hiring an expert to run its AI division much like it hired Satya Nadella to successfully advance its cloud efforts. But as with Google, it lacks the ability to push back on any major AI effort that seeks to eclipse them. The entire Wintel segment is a pale representation of what both Intel or Microsoft had in the 1990s when it took on all comers, and it seems completely unwilling to fund marketing to a level where it can defend against the increasing negative noise coming out of AI efforts in general and ChatGPT in particular.
This is leaving a massive potential opening for the next Microsoft/Intel to eclipse them. Bud.ai was just highlighted at Intel’s Vision, which feels a lot like Microsoft DOS did back in the day. Were it or anything like it to take off, Microsoft not only lacks the marketing capabilities to push back, but it also doesn’t even seem to know it needs them. Microsoft isn’t the only one. Eva may be even more like an AI OS, but this is hardly an exhaustive list.
Horse Drawn Carriages to Cars – PCs to?
Cars started out as carriages with motors. Part of their evolution had companies making engines and frames while others (like Fisher) made the bodies. Now, cars, like almost everything else (other than RVs), are made more like a cohesive whole. Thanks to Intel’s Dennis Carter, we care about what goes into our PCs far more than we care about what goes into any other technology we buy. What processor(s) are in your car, in your TV, in your appliances? You don’t care and really don’t want to care. But Intel’s Dennis Carter and Andy Grove made us care about the ingredients in our PCs. They made us want Intel Inside and got us to pay more for better tech we didn’t really understand.
Dennis Carter’s effort kept Intel powerful and relevant for decades (we still care about what is inside our PCs) and hampered efforts by both AMD and Qualcomm to displace it. The AI PC that Sam Altman and Jony Ive are creating doesn’t have Intel inside at the moment. Right now, it is more likely to follow Apple’s lead than the IBM PC vector given Ive is the brains behind the hardware and Altman has been talking about creating his own AI chip for some time.
This has the potential to be the next Apple. Recall when Apple went into smartphones and how poorly that worked for everyone that was dominant before. BlackBerry isn’t a phone company anymore, Palm is gone, and Microsoft and its partners are (with the exception of Lenovo and Samsung, which jumped to Google) out of the phone business. Wintel took terminals out at the knees, much like the iPhone took out prior powerhouses like Nokia. This same event is re-emerging, and I don’t see anyone anticipating it other than those creating it.
Why should AI PCs, which don’t really need keyboards or screens, look like PCs any more than the cars of today should look like carriages? Instead of creating the Model T, what if Ford had created the Mustang back in the day? We are talking about a remarkably similar speed here.
NVIDIA thinks we’ll have AGI in five years. If true (it’s hard to push back on NVIDIA at the moment when it comes to AI), all the generative AI stuff will be nearly instantly obsolete, and much of the existing computer world will feel like I did when I came out of college. Fortunately, I never really wanted to be a programmer anyway.
Wrapping Up
The AI revolution will be far bigger than the PC server revolution was. It is not only moving a magnitude faster but also still accelerating. This means massive disruption and most of the companies tied to the old model are likely to fail unless they properly anticipate the risk this change makes.
Yes, Microsoft kicked off this current wave, but I could argue IBM really kicked off the PC wave, and look how that ended up for IBM.
Those who don’t learn from the past are condemned to repeat it. What scares me is that I don’t see anyone bringing up these prior failures and putting in place efforts to ensure the past doesn’t come back and kick our collective butts.
Marketing is one of the things that could stall a collapse, but companies regularly cut marketing when a problem like this emerges. They shoot themselves in the foot rather than shooting at the problem. Sadly, I expect the same thing to happen here if what I fear shifts buying behavior away from the current status quo.
By the way, I don’t think smartphones will survive this pivot, either.
- Why Intel’s FAB and Foundry Expansion May Save the Company - April 19, 2024
- Intel Vision: How the AI Wave Could End the Wintel Age - April 12, 2024
- With AI, Communication Skills Become Vastly More Important - April 5, 2024
