The long-shot plan to save TikTok from a US ban

So far, ByteDance has shown zero willingness to spin off TikTok in the US. The Chinese parent company seems to be banking on the Supreme Court or President-elect Donald Trump rescuing the app before it’s banned next month.

The obvious names that would would buy TikTok if they could — Amazon, Google, Meta, Microsoft, Oracle, etc. — are sitting on the sidelines and waiting to see what happens in the coming weeks. The clock is ticking. Congress just sent letters to Sundar Pichai and Tim Cook reminding them that they will be legally liable for continuing to host TikTok in their app stores after January 19th. 

Then there’s Frank McCourt, the real estate billionaire and former owner of the Los Angeles Dodgers. For months, McCourt has been very public about his desire to buy TikTok. He has ramped up his drumbeat since ByteDance recently lost its legal fight on appeal. This week, he pitched more investors on his Project Liberty plan to buy the app’s US operations. 

When I spoke with McCourt over Zoom in between those investor meetings, he told me he currently has roughly $20 billion behind him for a bid. He has asked Kevin Mayer, who was briefly TikTok’s CEO the last time it was almost banned from the US, to be involved, though Mayer hasn’t signed on. McCourt told me his team has talked to “most” of ByteDance’s biggest American investors and that they’re “very interested” in his plan. (Spokespeople for these firms either declined to comment or didn’t respond to my pings, and Mayer didn’t have a comment.)

There are several reasons McCourt’s attempt to buy TikTok is a long-shot, the biggest being that, even if ByteDance wanted to sell, the Chinese government may not let it. Then there are the technical details of his proposal, which would see TikTok put on a decentralized protocol that is funded by McCourt and untested with a platform of TikTok’s size. For me, the biggest red flag of all is that there’s a cryptocurrency called Frequency tied to the project.

McCourt has high-minded ambitions for how the internet should work that are in line with the rise of federated platforms like ActivityPub and Bluesky. He imagines TikTok offering a marketplace of user-created algorithms, much like Bluesky does today, and users having the ability to own their profiles. Keep reading for more from our conversation this week…

The following Q&A has been edited for length and clarity.

My understanding is there are multiple pieces of Project Liberty. You’ve got the Decentralized Social Networking Protocol (DSNP) and then the TikTok bid, and they’re connected. It all seems a little bit afield from what you have historically done.

Having seen the harms of social media and where the internet was going as it became highly centralized, I dedicated some resources to start a public policy school at my alma mater, Georgetown. I was perhaps naive in retrospect, but I thought that we could get the policymaking apparatus out in front of the problem and steer things in a better direction out of that school. 

Then I realized that the public policy making apparatus is no match for Big Tech, so I began to go back to my roots. My family has been building infrastructure for 131 years, so this actually isn’t very far afield from our core competency when you think about this not as software at the app layer, but as infrastructure at the base layer. I talked to a few brilliant computer scientists that we have in the company and put the task to them of solving this from an engineering perspective if you had no limitations. The answer came back that you would go ahead and create another protocol that would connect us, just like TCP/IP connects devices and HTTP connects data.

Our goal here is bigger than buying TikTok. It’s reimagining how the internet works. Purchasing TikTok and moving 170 million people to a reimagined, upgraded internet would catalyze that alternative and compress time.

Is this a financial endeavor or is this philanthropy for you? Do you see a return on investment here? 

I see it as both and we’ve been very careful to separate the two. Project Liberty has an institute, a 501(c)(3), that’s purely not-for-profit. DSNP has been gifted to the world. That is being supported by the institute. 

The layer-one blockchain, Frequency, is tokenized. The community will own the majority of the tokens but that will be a commercial endeavor. TikTok, when we buy it, will be a commercial endeavor.

If you were given the opportunity to buy it but a condition was that you couldn’t do it on this protocol you built, would you still do it? 

Probably not because I don’t know how you would achieve the objectives we want to achieve, which is to be able to run the platform without the algorithm and give people ownership and control of their relationships. You need an implementation device to use DSNP. 

I would be totally open, however, if someone else built an alternative to Frequency. We’re looking to decentralize, not further centralize. 

Have you studied the fediverse: Mastodon, ActivityPub, and then also what Bluesky is doing?

I admire all the people who are doing that because they are trying to improve the internet and respect individuals. A federated approach is very different than having a universal social graph that’s globally accessible. 

We’re saying that something has to change fundamentally with the internet. What Jay Graber is doing with Bluesky is great but you still have a Bluesky identity. You’re still on Bluesky and your relationships are on Bluesky. Let’s stipulate that it’s better but you’re still there, right? At some point maybe Jay is not there.

Give me an update on the status of your investor pitch and how much money you have committed. 

People are super excited about this and capital will not be an issue at all. The issue is going to be what ByteDance does. We’ve been saying for over six months that we felt the government was going to win the case.

I don’t think ByteDance’s appeal is going to be successful. This is going to be a shut-it-down or sell it scenario. Like President-elect Trump, I’d like to see it not banned. We’re not asking China for the algorithm. We’re not an antitrust threat. We’ll pass CFIUS vetting. We don’t need or want the algorithm. We have a clean stack where the user base can migrate.

We don’t know what ByteDance is going to decide. We certainly hope they decide to sell and preserve some value for their shareholders. We’ll do it on terms that we think can be a win-win. China keeps the algorithm. US citizens are protected. The app stays alive.

Is $20 billion the amount that you still have committed?

That’s what we’ve got circled right now and that’s an order of magnitude what we think this is worth. Now, I say that with a huge caveat: we don’t know what ByteDance is selling. We think we have a very good idea of what the current numbers are, but it’s not like there’s a data room that ByteDance has set up and we’re inside of it. I think we have a good sense of what this would be worth if ByteDance keeps the algorithm and sells the user base and the content and the brand.

There are three categories of investors. One group is the existing American companies that have invested in ByteDance. We’ve talked to most of them. They’re very interested, assuming ByteDance agrees to sell. They would either put more capital in or keep their capital in. 

A second category are people that are familiar with the asset because they’ve looked at it in the past. Some are looking at it for the first time. These are the large balance sheets that deploy big amounts of capital and they’re interested in what we’re doing. The third category is the bucket of cultural capital, people with influence who are interested in being a part of this and bringing their communities with them. 

When will you share who your backers are?

We’ll be able to share once we know what ByteDance is doing. It makes no sense to be barking up the wrong tree here. 

You’ve had no conversations with ByteDance leadership, right?

That’s correct. I reached out before the decision. They were not interested in speaking at that time. We’ll try it again. They know at this stage of the game that we exist, we’re interested, and we hope that there’s incoming at some point as well. But we respect their decision-making process. They’re going to decide what they think is best for their interests.

Have you spoken to Trump or his team?

We’re in the process of reaching out. I’ve heard President-elect Trump say he doesn’t want to see the app banned. I’m very interested in having that conversation once this gets sorted on the China side and they decide what they’re going to do. 

Are you concerned about other large would-be acquirers, like an Amazon, also circling TikTok?

I don’t see the big incumbents being bidders here for a couple of reasons. One is antitrust, obviously. Secondly, imagine you were a tobacco company and you built a big business and then people started to get worried because people were getting sick and dying because of the addiction of cigarettes and carcinogens. If you bought a tobacco company after the surgeon general’s warning, then you were putting a target on your back because you were buying it with the awareness that your product kills people.

The growth of these platforms happened really fast. Then we discovered that there are harms. Anybody buying TikTok and replicating the current design, which is to scrape peoples’ data, apply algorithms, and manipulate people, they’re buying this with their eyes wide open. So I don’t see a lot of bidders because people realize they’re just putting a target on their back. 

Let’s take this problem and actually use it to catalyze an upgraded internet. Then, a lot of the problems go away. A lot of the lawsuits go away. A lot of the harms go away. That’s Project Liberty. 

Hundreds of the world’s top researchers gathered this week in Vancouver, Canada for NeuralPS, one of the world’s top academic gatherings in the field of AI. I asked The Verge’s Kylie Robison (who just published a fascinating deep dive on the history and future of ChatGPT) what it was like being there on the ground:

Are there any recurring themes you’ve been hearing in conversations?

The “Is AI hitting a wall?” conversation came up sometimes, but this is not the place where people actually believe that. AI safety was something I wound up talking about a lot. A research topic that also came up a lot was what Fei-Fei Li focuses on, which is helping AI understand the world like we do in terms of physics and objects. That touches a bunch of cool things — robotics, agents, advanced reasoning.

Which lab does everyone wish they worked at?

This conference was filled with students from Waterloo and University of Toronto who basically want any AI job. (It was quite sweet sometimes watching those students corner an AI exec at a party.) If parties are any indicator, everyone wants to get as close to OpenAI as possible. This is a highly academic conference with lots of people just wanting to do good, deep research.

What was the hottest party?

The Midjourney party was 4x oversubscribed and researchers kept sending me the RSVP link. Everyone was asking about the secret OpenAI parties, though.

There are plenty of interesting takeaways from Mustafa Suleyman’s recent Decoder interview with Nilay Patel. The Microsoft AI CEO doesn’t seem to agree with Sam Altman on the timing of AGI’s arrival, and he revealed that Microsoft won’t train its own frontier models when OpenAI is doing it for them anyway. He also takes some shots at Google’s management culture versus Microsoft’s, which I’m sure will cause some eyes to roll in Mountain View, given the way Suleyman excited.

As a follow-up to this conversation, I recommend listening to Sayta Nadella’s interview this week with Brad Gurley and Brad Gerstner. Reading between the lines, it’s clear that there is ongoing tension between the level of compute Altman wants and what Nadella feels comfortable spending. For those closely following Nvidia’s stock price, Nadella also mentions that Microsoft no longer feels GPU constrained but rather constrained by the power requirements for its data centers. 

• Zuckerberg, Bezos, and Altman buy their inauguration tickets. There have been lots of headlines this week about these three tech CEOs donating $1 million to Trump’s inauguration fund. That number just so happens to be the minimum requirement to receive “as many as a half-dozen tickets to eight inaugural events,” including access to an “intimate dinner” with the Trumps. (And you thought Taylor Swift tickets were expensive!) I’m not surprised by Bezos or Altman making donations. This is a 180-turn for Zuckerberg, however, who told me just a couple of months ago that he was going to stay out of this political cycle.
• Google’s big week of demos. Aside from a smaller version of its Gemini 2.0 model that appears to be quite good, Google’s blitz of product announcements this week — a quantum chip breakthrough, Android XR platform for headsets, a handful of AI agents — were all about hyping stuff that’s coming some day or not publicly accessible. I completely understand the desire to want to upstage OpenAI’s 12 days of “shipmas” but I am asking for a New Year’s resolution from tech companies: please stop announcing so many things without a ship date.

A few notable job moves this week:

• Rohan Anl, a DeepMind distinguished engineer and head of Gemini pretraining, joined Meta to help lead training for Llama.
• Zheng Gao, Tesla’s former head of hardware design for Autopilot, joined Zoox as director of hardware engineering.
• Google’s chief lobbyist, Mark Isakowitz, left to be chief of staff for incoming Pennsylvania Republican Senator-elect Dave McCormick.

• Juicy new emails and texts from Elon Musk’s lawsuit against OpenAI.
• We put Sora to the test and it was… fine?
• Cruise was reportedly planning “to launch a driverless service in Houston in 2025” before GM pulled the plug this week.
• Amazon is opening a new AI research lab in San Francisco.
• A deep, technical dive from SemiAnalysis on AI scaling laws.
• YouTube is the top social media platform for US teenagers followed by TikTok, Instagram, and Snapchat.
• The global ad market is expected to surpass $1 trillion next year, with more than half going to Google, Meta, Amazon, ByteDance, and Alibaba.
• Ev Williams made a social network for people who travel a lot.
• Pat Gelsinger sends his prayers.
• One of Mark Zuckerberg’s gold (plated!) chains sold at auction for over $40,000.

If you aren’t already getting new issues of Command Line, don’t forget to subscribe to The Verge, which includes unlimited access to all of our stories and an improved ad experience on the web. You’ll also get full access to my archive, featuring scoops about companies like Google, Meta, OpenAI, and more.

As always, I want to hear from you, especially if you have a tip or feedback. Respond here, and I’ll get back to you, or ping me securely on Signal. I’m happy to keep you anonymous.