META: Did Zuck Blow it On AI?
Posted April 10, 2024
Ari Goldschmidt
From even the earliest days of Facebook, it’s been clear that Mark Zuckerberg has had big aspirations.
Just two years after launching Facebook, Yahoo offered Mark $1 billion to buy the company.
At the time, Zuckerberg owned about half of Facebook's stock, meaning he could’ve walked away with $500 million from two years of work.
It’s the kind of offer that just about anyone else would have accepted.
Zuckerberg said “No thanks.”
Instead, the Facebook CEO would go on to turn his company into a global behemoth, propelling his net worth to $167 billion, or 330x the massive payday Yahoo offered him.
He’s done this, in part, by taking an aggressive approach to innovation.
Hackathons, events where people collaborate to build software projects often lasting just a weekend or less, became an incubator for new ideas. Early Facebook features like photos, groups, chat, and more were first developed at hackathons.
When a photo-based social network called Instagram started to gain popularity, Facebook went out and bought it for a billion dollars. Instagram has gradually become more popular than the original Facebook platform.
Zuckerberg seemingly struck gold again in 2014 when it acquired WhatsApp for $19 billion. Ten years later, the messaging app is used by nearly one in three people worldwide.
Of course, not all of Zuckerberg’s gambles have paid off.
The company famously failed in its attempts to launch a cryptocurrency payment system.
And of course, there are Facebook's (now called Meta) seemingly bottomless investments in the metaverse that have yet to pay off.
Despite these two setbacks, Zuckerberg continues to plow forward with its aggressive approach to innovation.
The company’s latest projects are arguably its biggest gamble yet.
Giving Away The Milk
When it comes to AI, Meta is no slouch.
The company has been investing in deep learning technologies for more than a decade.
Famously, Meta was one of the first tech companies to incorporate facial recognition technology in its platform.
Although the company has pivoted away from facial recognition due to privacy concerns, AI sits at the core of almost all of Meta’s different platforms.
Friend recommendations, newsfeed content, targeted advertising, and spam and safety monitoring are just a few of the many features the company has built on top of AI.
In addition to building AI for specific applications, since 2013, Meta has also invested in an AI research lab tasked with building next-generation technology.
This research lab, known as Facebook Artificial Intelligence Research (FAIR), sits at the heart of Meta’s AI investments.
More importantly, FAIR is taking a radically different approach to commercializing its technology.
As I described in last month’s Altucher’s Investment Network issue:
“While companies like Google and OpenAI are jealously guarding their AI code and research secrets, Zuckerberg has taken an entirely different approach… He’s giving it all away.
In February of last year, Meta released Llama, an AI application that lets anyone build their own chatbot nearly as good as ChatGPT.
Although Llama was originally only available to researchers, over time, Meta adjusted its policy to allow any business – other than a handful of competitors like Microsoft, Google, Amazon, Apple, Alibaba, etc. – to use Llama completely for free.
By doing this, Zuckerberg is turning up his nose at the long-standing belief that software is the most valuable component of technology products.
Zuckerberg’s logic is that by releasing Llama and other AI tools for free, they can benefit from the free labor of the open-source community.
Although it sounds peculiar, part of the open-source philosophy is that any developer can take open-source code, improve it, and then share it openly for the community to use.
By providing their AI tools for free, Zuckerberg is gambling that some smart developers – who aren’t necessarily motivated by money – will put their minds to work for Meta.
Another benefit is that open-sourcing its tools allows Meta to assemble an army of developers who are already familiar with its software.
Training new software developers is an expensive process that can easily take 6 months to a year, and sometimes longer, depending on the complexity of the code.
By making its software openly available, Meta can skip this expense and easily add developers who are ready and able to contribute from their first day on the job.”
Zuckerberg’s gamble is so large and controversial that the Altucher’s Investment Network team decided it was worth another look.
Meta’s Llama language model, for example, is in large part powering a tremendous amount of innovation in the AI space.
Countless startups have taken Llama, modified it for their own use cases, and built rapidly growing companies on the back of Meta’s innovation.
Meta’s contribution to the field of AI cannot be understated.
In a keynote event earlier this month, NVIDIA CEO Jensen Huang even went as far as saying that the release of the second version of Llama, Llama 2, was “the single most important event that took place last year” for AI development.
In my article last month, I suggested that with Llama and Llama 2, Zuckerberg was attempting to prevent Google and Microsoft from gaining an insurmountable lead in AI development.
By offering AI tools for free, Zuckerberg and FAIR are throwing a bone to startup founders and researchers around the world.
As I described in that article, this is obviously good news for firms that produce AI chips like NVIDIA.
However, the benefit to Meta is less clear.
As our team has researched the topic further, we’ve become even more bullish on how Meta will turn their open source investments into long-term opportunities.
A Massive Data Moat
One thing that Meta has that no startup could ever hope to replicate is a deep repository of data of every single one of its users.
This includes everything including images, interests, location history, friends, browsing history, and much more.
The amount of data is simply staggering.
This gives Meta a massive advantage when it comes to putting the latest AI innovations to use. After all, the accuracy of AI depends on having reliable data as a starting point.
By open sourcing its technology, Meta benefits from the free labor of other software developers in the open source community.
Meta is banking that researchers and open source developers will give it the keys to the castle to more effectively use its enormous trove of data.
The stakes are absolutely massive.
With improved AI, Meta will not only gain deeper insights into its users but be able to more effectively influence their decision-making process.
While Meta has given a tremendous amount to the open-source community, they don’t have an obligation to give it everything.
FAIR’s leadership has suggested that they don’t intend to open source anything that uses Facebook user data.
However, Meta’s data advantage is just one part of Meta’s overall AI strategy.
A Massive Distribution Network
Another asset working in Meta’s favor is its large distribution network.
Late last year, Meta flicked the switch on Meta AI on WhatsApp.
Meta AI on WhatsApp – which can be accessed in WhatsApp by creating a new chat and selecting “New AI Chat” – will gradually be rolled out to WhatsApp’s 2.7 billion users worldwide.
The feature release makes Meta’s AI tool one of the most easily accessible AI chatbots available on mobile devices today.
This represents only the first of the many new AI features we can expect Meta to release in the coming year.
I would not be surprised to see Meta release new AI tools for editing videos, writing comments and captions, or helping advertisers create ads.
Meta’s ability to use its existing platforms to get these tools into the hands of users will ultimately give it a big leg up in the AI race.
Meta Stock At Record Highs
We first recommended buying Meta back in January of 2017 for the Altucher’s Investment Network portfolio when the stock was trading at $116 per share.
As of earlier this month, Meta’s stock now trades at $468, a 300% gain.
Our return on Meta runs circles around top investors like Warren Buffet, who only produced a 147% return in the same time frame.
This is the power of our tech investment research.
Our team believes that these returns could look like pocket change compared with the gains to be made from shrewd AI investments.
So what does this mean for Meta stock?
Long term, I’m optimistic.
Shorter term, I’m less optimistic.
Meta stock has a history of crashing 30%+ from all-time highs.
Mark Zuckerberg has a history of taking smart risks that pay off in a big way. However, the market doesn’t always appreciate his long-term vision.
I would not be surprised to see this play out yet again as it will take time for Meta to see the benefits of its open-source investments.
Conclusion
Mark Zuckerberg has a history of making big, aggressive bets on innovation that have paid off handsomely for Facebook/Meta over the years.
His latest gamble involves giving away Meta's cutting-edge AI research and tools like Llama for free via open source, banking on the free labor of developers and researchers to improve the technology which Meta can then leverage with its enormous data advantage and massive distribution platforms like WhatsApp.
While this open-source AI strategy is controversial…
It has already spurred significant innovation and Meta's early contributions are being hailed as among the most important for advancing AI development.
Long-term, Meta is well-positioned to be a leader in AI, but the stock may face volatility as the market assesses the payoff from Zuckerberg's bold open-source AI investments over time.
The Altucher’s Investment Network team is on the lookout for new, high-flying opportunities in the AI space.