We’ve spent a lot of time lately talking about what OpenAI and Anthropic are doing with artificial intelligence.
We’ve also talked about how those companies, along with Microsoft and Google, are pouring hundreds of billions of dollars into the infrastructure needed to make AI more powerful.
Today, I want to check in on another company that has been surprisingly quiet in our recent conversations.
Meta.
While he’s not sparring with professional MMA fighters, Meta CEO Mark Zuckerberg has been busy penning a manifesto on superintelligence. Last Monday, he published it: a 6,500-word post titled “The Future Is for Everyone.”
And whatever you think of Zuckerberg, I think it’s worth paying attention to.
Because this isn’t a Facebook product announcement or a prediction about when artificial general intelligence (AGI) will actually arrive.
It’s Zuckerberg’s blueprint for what he believes the world could look like once intelligence becomes abundant.
And a lot of it should sound familiar to Daily Disruptor readers.
Invention, Not Automation
One of the biggest questions surrounding AI right now is what happens to our jobs.
Zuckerberg thinks that’s the wrong question.
Because he believes the real economic impact of AI won’t come from machines replacing people. It’ll come from giving people capabilities they’ve never had before.
As he puts it: “Invention, not automation, will be the greatest contribution of superintelligence.”
In other words, Zuckerberg believes the next generation will help us create things that didn’t exist before. That could mean discovering new drugs, inventing new products or finding entirely new ways to run a business.
And eventually, Zuckerberg believes almost everyone will have an AI working alongside them.
He writes that: “Everyone will have an exceptionally capable personal agent that understands you, your goals, and everything you care about.”
That agent could work around the clock helping you manage your finances, career, health, relationships, hobbies and even your home.
Zuckerberg says his personal AI agent already flags information he might find interesting, helps him prototype ideas and monitors his sleep and workouts.
Obviously, this could have enormous economic implications.
Zuckerberg believes AI could dramatically reduce how many people you need to turn an idea into a business, saying: “People are starting to be able to manifest ideas themselves without having to raise money or build large teams.”
That’s something I’ve been writing about for quite a while.
We all saw how the internet dramatically lowered the cost of starting a company. With AI, Zuckerberg predicts company sizes could also shrink as it becomes increasingly capable.
Instead of a few enormous companies employing huge numbers of people, we could see far more businesses operated by smaller teams.
Eventually, he expects “small numbers of people with personal superintelligence agents” to run companies at significant scale.
Of course, all those digital workers will need compute. And Zuckerberg spends a surprising amount of his manifesto talking about it.
He says Meta plans to offer free versions of its superintelligence tools to billions of people.
But compute will never truly be unlimited.
So Zuckerberg envisions a system where people who need more computing power can pay for it through what he calls a “dynamic auction mechanism.” Meaning, everyone gets access, but the more intelligence you consume, the more computing resources you need to buy.
Supplying all that compute will also require an extraordinary amount of physical infrastructure. Which brings us to another theme we’ve been following closely.
Data centers.
Image: Meta
Meta is already building enormous AI facilities across the country. And Zuckerberg acknowledges that communities aren’t going to welcome these projects if they raise electricity prices, consume enormous amounts of water and leave local residents with nothing in return.
So Meta says it plans to build its own energy-generating infrastructure wherever it invests.
Zuckerberg believes these AI projects could eventually play the same role that railroads, highways, electrification and broadband played in previous economic booms.
Communities connected to those networks became centers of commerce and innovation.
Now he thinks AI infrastructure could do the same thing.
But the most interesting part of Zuckerberg’s manifesto comes near the end, when he tackles something we discussed recently in the Daily Disruptor.
That’s the point where an AI becomes capable of improving the systems that make AI itself more powerful.
Zuckerberg thinks this creates an enormous dilemma.
“There is a dilemma that once AI systems can autonomously improve themselves, any lab that doesn’t let their AI system direct a substantial amount of compute capacity towards recursive self-improvement will inherently fall behind.”
He believes this could create an AI arms race where eventually nobody can afford to stop moving forward. And he gives an extraordinary example of how quickly that could change the balance of power.
He says a self-improving AI could theoretically discover ways to extract “100x or more intelligence out of each gigawatt.”
If that actually happened, an AI running on only a fraction of the world’s computing infrastructure could suddenly possess more effective intelligence than everyone else combined.
That’s why Zuckerberg believes the ultimate challenge isn’t simply building superintelligence.
It’s making sure humans remain in control of it.
Here’s My Take
I don’t agree with everything in Zuckerberg’s manifesto.
And obviously, Meta has an enormous financial incentive to convince the world that billions of people should have personal AI agents.
But his larger argument is worth considering.
Most of today’s AI debate focuses on what AI will replace. But Zuckerberg is asking us to think about what it will allow us to create.
I believe that’s where AI’s greatest promise lies.
Regards,
Ian King
Chief Strategist, Banyan Hill Publishing
Editor’s Note: We’d love to hear from you!
If you want to share your thoughts or suggestions about the Daily Disruptor, or if there are any specific topics you’d like us to cover, just send an email to dailydisruptor@banyanhill.com.
Don’t worry, we won’t reveal your full name in the event we publish a response. So feel free to comment away!
