Well, it’s finally happening. The long awaited “agent revolution” is here, in the form of Muse, Meta’s two-week-old entry into the AI platform wars. The early reviews are in, and they’re quite positive (“blown away!” gushed the Times). Muse quickly hit No. 1 in app store downloads. Wall Street cheered, giving Meta’s stock an 11 percent bump.
Watching all this, I’ve never felt more like the wheezy Lorax, scolding from the stump of the last Truffula tree:
Do not use these agents, do not I implore!
For while they may charm you, it’s not you they work for!
To understand why I’m so adamant, let’s review the core business model of the four biggest consumer technology platforms. Big tech companies are aggregators: They jealously aggregate our attention (Facebook, Google), our purchases (Amazon), and our device usage (Apple). They turn this aggregation into money by distilling our interactions into advertising products. All four do lots of other things, but for each of them, advertising provides the foundation of their profits. This, in turn, drives the decisions they make about which products to launch, and how those products work.
Agents pose novel challenges and opportunities for these aggregators. As you might imagine, agents do not care about ads designed for human eyes – a serious problem if, like Google and Meta, you’re in the business of selling human attention. And agents are scary good at finding the best deals, another serious problem if, like Amazon, you’re in the business of dynamic pricing. Lastly, agents can live above the level of a device – meaning they pose a considerable threat to the device-based lock in so favored by Apple.
Put simply, agentic AI has the potential to disrupt big tech’s oligopoly, bigly. Let’s break down how each of the four are responding.
META. Proving yet again it’s the king of copying trendy products (in this case, OpenClaw), Meta launched Muse on Sept. 8th, and since then the company has been on a mission to prove Muse is a hit. The company’s PR team is working the phones, dispensing “exclusive,” off-the-record metrics to gullible tech reporters. Given that Meta is in the business of collecting your attention and your data, its no surprise that Muse presents an unparalleled opportunity for Meta to mine information that was previously impossible for it to obtain. For now, Meta has stated it will not use data collected by Muse for its advertising business. It will, however, use that data to train its AI platforms. But if you think Meta will never put ads in Muse, I’ve got a bridge to WhatsApp I’d love to sell you.
For Meta, agentic AI is a chance to move upstream from its own applications, and begin to capture consumer demand at the level of the entire Internet. No wonder Meta is investing $145 billion in AI this year alone!
AMAZON. Like most tech companies, Amazon has a long history of blocking any “bots or agents” that seek to interact with its platform. But with Muse, that policy has become existential. Amazon quickly blocked Muse, and has stood up its own AI agents, though for now they only work on Amazon properties. But blocking outside agents will only hold for so long – and might just motivate agentic shoppers to move their purchases to places that welcome the bots (Shopify, for one, has already integrated with Muse). Amazon is in a bind – and in coming months, the company will likely seek a pact of mutually assured destruction with Meta where the two companies can split the profits earned by allowing Muse into the e-commerce kingdom.
APPLE. As you might expect, Apple is responding to the rise of agents by leading with a prophylactic spin of protecting your privacy. Anticipating the launch of Muse, Apple last week launched a newly redesigned version of Siri as an AI agent that can access deeply protected silos of data no other company has been allowed to touch. Here’s developer Ian Collins on why that matters:
Until now a customer’s bot was something a company had to stop. Bots were scrapers, crawlers, fraud. [With Siri], Apple has put a bot on the other side of that line, fetching a customer’s own information out of third-party apps because the customer asked, and the app companies agreed. If a hotel company is comfortable letting Apple’s AI read its app on your behalf, it can no longer tell you that an AI touching customer data is a privacy problem. They already let the bot in.
Which leaves the question the app makers would rather nobody asked. Why does Apple get to do it and you don’t?
Good question, and one we’ll get to later. But for now, Apple is responding to Muse by out-Musing it: It’s betting that as the default AI agent for iOS, Siri will win. If Muse starts beating Siri in the race to aggregate customer attention, don’t be surprised if Apple finds an excuse to cut Meta off at the knees. It’s done it before.
GOOGLE. It took Google two years to respond to the threat posed by ChatGPT, but when it did, it pushed all in. Integrating AI chat into its core Search property has driving advertising revenues to an all time high. Expect the company to do the same with agents. Google doesn’t have an answer to Muse, yet, but it certainly will, and soon. Along the way, it’ll protect its massive AI ads business by charging agents the same way it charges humans: A click is a click, after all. To be honest, of all four, I’d be most likely to use a Google agent, because of the company’s roots in the open web. But would I trust it? Nope.
So let’s review. Meta has launched Muse, a data-hungry agent that works for Meta. Amazon has blocked Muse, and launched its own on-site agents who work for…Amazon. Apple has (re)launched Siri, who works for … Apple. And Google is about to launch its own agents, who, of course, will work for Google.
Which leads me to ask: Why on earth would we trust any of them?
The genius of AI agents is their potential to work for us, and only us. I wrote about this more than three years ago: The only good agent is an agent who, like a magic genie whose gaze binds it to its master, is beholden to you alone. An agent built by Meta, Google, Amazon, or Apple may claim to work on our behalf, but its true master is the company that built it. Those companies’ business models will determine what actions their agents will perform, and to believe otherwise is to ignore centuries of business history.
This is why I’m cheering the rise of truly independent agent companies like Instinct or Town. They’re building a future I can get behind. I hope they get a chance to shine, but I’m not certain they will. It’s a huge lift to build a world-class product that can compete with the tech oligarchs – and it’s even harder to refuse buyout offers reaching into the tens of billions of dollars. Godspeed.
—