Shots Fired Between Tech Behemoths in the AI Dance Battle
(Image Prompt: tech executive Sundar Pichai doing a nerdy LARP involving hand to hand battle with battle axe and spear 3D HD photorealistic with a grenade in the air about to hit the combatants)
This week we saw the battle for AI dominance heat up. No, I’m not talking about Meta’s Chief Scientist Yann LeCun call out OpenAI for failing to contribute to AI research but instead focusing on “high wow-factor flashy demos and product development, at the expense of research, so as to attract investments.” That was technically last week.
This week the reigning principals themselves, Satya and Sundar, weighed in on the skirmish where Satya extended an open invitation for Google to dance. In an interview with Nilay Patel, Microsoft CEO Satya Nadella pulled no punches saying how long he has waited to mount a realistic challenge to Google in the search category.
Today was a day where we brought some more competition to search. Believe me, I’ve been at it for 20 years, and I’ve been waiting for it. But look, at the end of the day, they’re the 800-pound gorilla in this. That is what they are. And I hope that, with our innovation, they will definitely want to come out and show that they can dance. And I want people to know that we made them dance, and I think that’ll be a great day.
And of course that prompted this user to provide a generative image of Disco Satya:
Microsoft has a lot to be optimistic about. Search is profitable for them but they have very little market share. Still, according to the article, search accounts for $11B in revenue per year, and any gains—however small—will be taking share from Google. Satya is rightfully beaming at the asymmetric benefit to Microsoft at the cost from Google:
“Very few times in history do opportunities like that show up where you suddenly can start a new race with a base where every day is incremental gross margin for you and someone else has to play to protect it all: every user and all the gross margin.”
This is the video of the interview here:
Alphabet responds.
Many have been waiting for Google to respond since OpenAI launched ChatGPT last November. One Xoogler credited with being the creator of Gmail observed last December:
“Google may be only a year or two away from total disruption. AI will eliminate the Search Engine Result Page, which is where they make most of their money. Even if they catch up on AI, they can't fully deploy it without destroying the most valuable part of their business!”
Speculation on Google’s reaction reached fever pitch two weeks ago when the NYT reported that Google called a Code Red. As if in a sequel to the nerdiest action movie ever made, co-founders Larry and Sergey were called back from retirement to return for one last job. Supposedly, some screenshots shared to reporters validated that Sergey file a changelist to add his username to the code related to LaMDA, Google’s natural language chatbot.
The thought that Google needs to “catch up on AI” is comical. Google Brain and DeepMind are two of the most advanced research and publishing labs in AI today. But it would be naive to equate that with the ability to ship products. History is filled with graveyards of large companies that developed revolutionary technology and failed to deliver on their potential. Xerox Palo Alto Research Center (PARC) invented the mouse and the graphical user interface (GUI), but startups Apple and Microsoft commercialized them. The AT&T commercials that described the future are eerily accurate, but AT&T captured very little of the value despite its own Bell Labs having made much of them possible.
Leading in AI research is very different from being able to release a product or satisfy Wall Street. On the latter two, Google stumbled. Google CEO Sundar Pichai was forced to enter the AI fray this week by announcing Google’s upcoming conversational AI bot built upon LaMDA called Bard.
In Google fashion, they didn’t release the model for people to try it, but will be providing access to “trusted external testers” to provide human feedback. But the market reception to the announcement, which depicted Bard as providing a false answer in the demo, was not kind: Alphabet shares dropped 7% on Wednesday. To be fair, the S&P also dropped about 1%—but still.
Here’s some in-between spaces observations about the sparring that occurred this past week:
Launching is hard. It’s even harder for incumbents.
I describe Satya and Sundar as “sparring” because neither company has released what they’ve announced. On this point, I give Meta a lot of credit for having released their AI chatbot Galactica designed for scientific research before ChatGPT came out. Never mind that it was withdrawn after three days for providing misinformation, which understandably strikes a raw nerve for a company like Meta. Launching something controversial is much harder for large incumbents that have much broader surface areas for attack.
Ironically, however, some attack is from within. Most people understand how large companies have difficulty innovating. They lack speed, it’s hard to adapt, fear of cannibalization, innovator’s dilemma, etc. But an often overlooked reason is that there are teams internally that will actively prevent you from launching. If you look closely, you will find them.
That is likely why Larry and Sergey had to get involved. Professional CEOs often don’t have the mandate to ignore whole parts of organizations that have been built under their watch.
I think the fact that Google rushed out this announcement is a positive sign. Even the botched demo of Bard giving misinformation is to be expected. Google hasn’t had to move this quickly with such urgency in a long time. Given how many people were recently hired (analyst estimate about 63,700 were hired since the pandemic), I suspect only a few people at the company has ever been involved in an important product release. The fumbled announcement and market reaction are all signs of Google shaking some rust off its processes. No big deal.
Startups offer leverage even to existing incumbents.
It’s almost unthinkable that the balance between the two behemoth incumbents in tech, Microsoft and Alphabet, was shaken up by a startup. Okay, OpenAI is no longer a startup, but territories that have long been settled after decades of hand-to-hand combat between incumbents became disrupted by a grenade launched by a company that didn’t exist prior to 2015. And the response to that volley from Alphabet, who many regard as the most successful startup ever built, promptly wiped out over $100B in market cap. The stage is set for tectonic shifts in the technology landscape, all because of a single startup.
The other way startups offer leverage is that it provides a buffer for large incumbents. Meta and Alphabet will be directly attacked on the basis of AI safety, alignment, bias and misinformation. Microsoft, on the other hand, can outsource the monumental task of managing hallucinations to its partner OpenAI. That luxury will end once Bing and Edge integrate with the model that follows GPT3.
Satya is a killer.
One thing that stuck out in his interview, other than Satya’s unconstrained glee, is that Satya does not shy away from zero-sum language. We are so used to hearing people deflect questions about competition by talking about focusing on their own mission, creating value, and lifting all boats. To be sure, Satya slightly nods in that direction, but then squarely shows he’s looking to expand the size of his pie.
You can’t blame him given the decades of Google products taking share from Microsoft:
Gmail from Outlook and Hotmail
Chrome from Internet Explorer and Edge
Google Docs from Microsoft Office
Chrome OS from Windows
Take a look at that list: these are side projects that are peripheral to Google’s main business taking share from Microsoft. In fact, each of those are heavily subsidized by Google’s healthy margin from search. That’s why Satya can hardly hold back his excitement at the chance to strike at Google’s jugular. Reducing Google’s margin in search will help Microsoft compete in these other areas, including the increasingly important battle between Azure and Google Cloud. Satya is a fiery competitor, and that’s what might make him the best professional tech CEO in history.
Honorable mention: Reid Hoffman.
I can’t help but think there’s something here. Reid was an early investor in OpenAI, is on the Microsoft board, and the acquisition of Reid’s LinkedIn allowed Microsoft to have a stake in a big part of the internet economy where Microsoft was behind. I haven’t heard anything specific, but would love to hear more about his role as dealmaker here. It just makes sense.
Every once in a while, a company will make an acquisition whereby the successful integration of a single individual justifies the acquisition. Apple’s acquisition of Next Computer brought Steve Jobs back to a struggling Apple. Google’s acquisition of Nest could easily be justified with a successful integration of Nest founder Tony Fadell, known as the “father of the iPod.” One could argue that Brett Taylor was twice the target of a solo acqui-hire with Facebook’s acquisition of FriendFeed and Salesforce’s acquisition of Quip, and each time Brett ended up as the CTO of the acquirer. Microsoft’s acquisition of LinkeIn almost fits this pattern but for the fact that LinkedIn is actually a very valuable asset.
The unexpected return of consumer surplus.
With rising interest rates and the cost of capital going up, consumers will no longer be in a position to have VCs fund their on-demand needs. But that does not mean everything will get more expensive. One of the most important deflationary forces is competition in technology. I’m optimistic to see competition again in areas where we’ve suffered from winner-take-all dynamics.
I remember being at Google in 2005 and hearing murmurings of Microsoft launching a search competitor. While outwardly we expressed confidence, I do recall an underlying anxiety of being crushed by a large incumbent. Once Bing launched, that anxiety faded, and I can only speculate that the paranoia has been dormant—until now.
The truth is search has not kept up with consumer behavior and expectations. Not that search quality has suffered, but that the way people use search has changed over the previous decades. Satya hit the nail on the head by changing the paradigm from a “search” engine to an “answer” engine. It used to be that Google allowed us to find the right website. Today, only in some cases are we looking for a website, most of the time we are looking for an answer. Google’s answer box—or featured snippets—comes closer, but delivering a list of links as an answer to a question feels increasingly antiquated.
Competition between strong, well funded tech companies means there will be goodies in store for us, their customers. The first round of the AI skirmishes finishes with the following score: Satya-1 and Sundar-0. I give the most-valuable-assist to Reid Hoffman, and the winner in the next few years will be the consumer.