Mark Zuckerberg won’t say these two words anymore, and it’s a clear sign of how much TikTok terrifies him

October 28, 2022, 11:30 AM UTC
Facebook founder Mark Zuckerberg
Facebook founder Mark Zuckerberg in earlier days.
Noah Berger—Bloomberg/Getty Images

Remember the social graph?

There was a time when that’s all Facebook founder Mark Zuckerberg would talk about. The interconnections between Facebook users and their friends, their friends of friends, and all their interests and activities, were the glue—or graph—that made Facebook so powerful.

And given Facebook’s massive number of users, the social graph is what made it impossible for challengers to catch up.

But Zuckerberg no longer seems interested in the social graph. In fact, you almost get the sense that he’s afraid to even utter the words.

Just look at Wednesday’s conference call in which Meta (as Facebook is now called) discussed its disastrous third quarter earnings. As Zuck fielded questions from Wall Street analysts about why the company’s profits were less than half of what they were the year before, the Meta CEO repeatedly pointed to progress it was making in its “A.I. discovery engine.”

The discovery engine Zuckerberg is so jazzed about is an artificial-intelligence–powered algorithm that chooses the videos, photos, and other content to serve Facebook and Instagram users. It doesn’t matter whether the video is from a close friend, a secondhand acquaintance, or a complete stranger—if the A.I. thinks you’ll dig it, it will show it to you.

Zuckerberg’s A.I. discovery engine is his attempt to replicate the secret sauce at TikTok, the Chinese-owned social media service that’s eating Meta’s lunch right now. The engine, in practice, would harness the same TikTok “for you page” powers and bring them to Meta. And Meta is spending big bucks in capital expenditures to beef up its infrastructure so that it can offer those A.I. capabilities.

What’s really remarkable though is how deeply Zuckerberg appears to have bought into the A.I. model pioneered by TikTok, and how little faith he seems to have left in what was once Meta’s secret sauce: his beloved social graph.

“Our A.I. discovery engine is playing an increasingly important role across our products — especially as advances enable us to recommend more interesting content from across our networks in feeds that used to be primarily driven just by the people and accounts you follow,” Zuckerberg said on Wednesday’s call.

It was one of a half-dozen times he referred to the discovery engine. The word A.I. was used roughly two dozen times by various other Meta executives on the call.

Number of social graph mentions: zero.

Zuckerberg hasn’t completely given up on the idea that interpersonal connections on the platform have value. Yet even when referring to the benefits of the social graph on Wednesday, Zuckerberg steered clear of the term, referring to content from “friends and family” rather than content from the social graph.

It might seem like semantics, until you go back to past earnings calls, when the graph was one of Zuckerberg’s key talking points, wielded like a cudgel to emphasize the platform’s unique power.

In a 2012 earnings call Zuckerberg mentioned the social graph and variations of it, like the “search graph,” 11 times. In the company’s pre-IPO S-1 filing, Facebook described the social graph as one of the “foundations” of a newly evolving internet.

But the web has continued to evolve in ways that Zuckerberg didn’t anticipate. The social graph model may have even helped push young users off Zuckerberg’s platforms—many of us are just not interested in our Great-Aunt Betty’s rants anymore. TikTok made incredible use of this changing dynamic of attention, which in turn made the platform the titan it is today.

It’s a fundamental change, and Zuckerberg’s reticence regarding his former buzzword underscores just how vulnerable his empire has become.

Do you have insight to share? Got a tip? Contact Kylie Robison at, through secure messaging app Signal at 415-735-6829 or Twitter DM at @kyliebytes.

Read More

Artificial IntelligenceCryptocurrencyMetaverseCybersecurityTech Forward