FTC Chair Lina Khan at Y Combinator

Thanks everybody for coming to White Combinator today. We're so excited to host Cherina Khan of the Federal Trade Commission. You know, I’m Luther LOM, the new head of public policy at White Combinator, and this is the first event of its kind, I think, at least since I've been here. I’m on week five, but I hope that we have a lot more events like this.

I think one of the reasons I’m feeling optimistic about kind of this bridge building that’s happening between the technology industry and Washington is because of the leadership of Gary Tan. You know Gary's got this great line about that you may not care about politics, but politics cares about you. So, I just want to make sure folks know that you know I’m a resource at Luther White Combinator. Please don’t ever hesitate to reach out. We want to be the voice of Founders in Washington.

I’m really excited for Bruce, Lena, and Gary to come up on the stage, but let me just say a few things about Cher Khan. In Washington, or in our industry, it's not uncommon to have just a massive impact when you're really young, but it's far less common in Washington. Cher Khan, before she'd even turned 30, wrote this seminal article on Amazon that was published in the Yale University Law Journal, and it went viral. Law Journal articles don't typically go viral, but it really helped reshape thinking around large technology companies.

So, the fact that she came into the scene, and when President Biden was elected, she was appointed to the Federal Trade Commission, and he named her as the chair is just so significant. At the age of 32, she is the youngest ever to hold that position and is helping shape our economy in that way. It's just very exciting. Gary and Lena are going to have a conversation about our industry, and there'll be plenty of time for Q&A, so please give them a warm welcome.

Thanks so much, Luther. So, I'm going to make Luther embarrassed in return, which is I received some messages from various people in tech, and they all sort of resembled like, "Oh, bringing Luther on at YC was like signing Michael Jordan in his prime for what he does!" So, I agree and welcome Luther. Cher Khan, thank you so much for joining us. It's just such a pleasure to see you here in San Francisco and pretty close to Silicon Valley as we call it.

I guess one of the funniest things about prepping for this, you take a look at Cher Khan's bio on the FTC homepage, and you know, it has her bio, it's very illustrious, obviously. Then, a few lines down, the font gets bigger and it says, “A huge disclaimer about CLA scams.” The font is 3x bigger. So, it's really important.

I guess, why is that there? What does it actually say?

Yeah, it's “Why do you have to have this big disclaimer about that right now?” Well, first of all, let me just say it’s so great to be here. Thanks so much to YC for hosting this chat. We’ve really had a great time here in SF for a couple of days, having the chance to meet with the business community, hear from Founders, and hear from investors. It’s just really a big part of what we’re looking to do in DC is make sure that we’re not just in a silo out there, and we’re in fact hearing from actual people in the business community.

So, I think you’re referencing probably some disclaimer that says FTC officials will never ask you for money, never demand money, or make threats? More specifically, that’s right. We see a lot of scams and frauds in the marketplace.

Sometimes scammers are pretending to be the government; sometimes they’re pretending to be some type of legitimate business. Some of our work at the FTC is not just in the antitrust space, but also in consumer protection. We enforce the laws that prohibit deception and that prohibit unfairness. I think there's been an epidemic at various points of people pretending to be FTC officials and demanding money, people pretending to be IRS officials making threatening calls and saying, “Hey, you owe us $10,000 or we're going to seize your home.”

So, this stuff is scary for people, right? Sometimes they target elderly Americans, sometimes they target people for whom English is the second language. And so part of the FTC's work is to make sure that that type of fraud and deception and predation is not happening in the market.

One thing, so Cher Khan has been sort of on a whirlwind tour of Silicon Valley. One thing I really liked about your Stanford talk you mentioned is that there’s sort of a history of your antitrust and anti-monopoly regulation actually leading to a lot more innovation. The most noteworthy of which maybe is the AT&T consent decree. I believe that was in, I mean, when was that? And how did that sort of lead to Fairchild Semiconductor?

Yeah, so the talk I gave was really about how we should understand the relationship between Silicon Valley and DC, or the relationship between entrepreneurs and innovators and what the government is doing. I think sometimes there can be somewhat of a view that like the only thing that the government should do is just totally get out of the way.

What I was trying to share is that from the perspective of antitrust enforcement, enforcing the laws of fair competition has actually been really critical to ensure that the market is open and competitive and that the next tier of innovators are actually able to come onto the market and scale, becoming tomorrow's giants rather than getting killed in the crib by the existing incumbents.

There’s been a history of that. I mentioned the AT&T consent decree in 1956, where the Justice Department required AT&T to basically open up its patent vault. You had Bell Labs that had done a lot of phenomenal research and discovery, but all of that discovery had been locked up. So, DOJ said, “Hey, you have to make these patents available to the rest of the business community on a royalty-free basis.” That ended up then spurring a whole set of additional innovation, including the example that you mentioned.

I also noted that the Microsoft case, that the DOJ brought in the late 90s, was in part a response to what the government had been hearing from startups, be it the Netscape of the world or others that were basically introducing this middleware tier and the browsers that risked disintermediating Microsoft’s operating system monopoly.

Microsoft felt threatened by that, and so they engaged in a set of tactics to try to kill off those companies. The DOJ coming in and bringing the antitrust case ended up re-oxygenating that market and I think really providing runway to firms like Google, right, and the next tier of innovators to come into the market and not get preemptively killed.

But, you know, it seems like I’m sure you can’t confirm or deny this, but from our view or from many of our friends' views, they have pulled up the ladder behind them a little bit for now. So, there are a whole bunch of lawsuits underway right now and you know, these lawsuits differ in important ways. But I think it’s fair to say, if you zoom out, that each of these lawsuits against some of the existing dominant platforms is more or less about that story— that you had these firms that in the early years were really competing for users, really doing what was best for their customers, through that they were able to gain scale.

Then at some point or another they really switched tactics to start focusing on building moats, protecting moats, and pulling up the ladder so the next tier of startups and entrepreneurs really don’t have the same opportunity in the marketplace.

Yeah, I guess one of the things that has been pretty fun after we announced that you were coming to YC is I received a number of DMs from friends in tech who say, “Well, you know, don’t you know that she sues tech? You know she sues Microsoft and Facebook and all these folks; isn’t she an enemy of tech?” What do you say to those critics?

So, look, we are a friend to entrepreneurs and Founders and open markets that are necessary to make sure that if you have a good idea and you're a smart business person, you have an opportunity to succeed. Ensuring that people's success is premised on the quality of what they're doing in the marketplace, right? Are you making something that customers want, rather than success depending on exploiting some existing privilege or having to bow down to one of the existing incumbents?

So that’s really what the goal of the antitrust laws is, to make sure we have these open, fair competitive markets so that people can get a fair shake.

Absolutely, well, I’m super aligned with you on that. Obviously at Y Combinator, we’re here to try to make—I mean, as we mentioned downstairs earlier—we had lunch, and the funniest thing was we’re realizing that our motto of “make something people want” actually very much comes into conflict with the goals and ideals of many of the tech giants because they, you know, sometimes make things people want, but sometimes they make things that make themselves a lot of money.

And you know, that’s one of the great pieces of the FTC and the U.S. government that, you know, we want—I don’t think we need special treatment. I don’t think the startup world or like little tech is asking for anything more than a fair playing field and that’s something that I feel really aligned with you on.

Yeah, for sure. I mean that really goes to the heart of why the antitrust and anti-monopoly laws were created, to make sure that you have that level playing field and that, you know, the big firms are not able to use their muscle to unfairly squeeze out the competitors that, you know, but for that anti-competitive conduct would have a real shot.

So, I am curious though. Some folks in the audience bugged me about this too. You know, one of the things under your purview is commenting or regulating acquisitions. Of course, you know, a great many folks in this room may want to retain that ability to have that be a viable exit for their business; they would like to be able to have that as an option.

Things like Adobe, you know, and some of these disclaimers, like this specific action—like Adobe versus DOJ, Adobe and Figma versus DOJ—that’s not your purview. But I’m sort of curious how you respond to that, you know around the acquisition side. Is there some aspect of blocking acquisitions that you know would actually help little tech?

It's a good question. So, both the FTC and the antitrust division, you know, review mergers and acquisitions. If you're making a deal that's over like $11 million, you have to notify the government and then we have 30 days to figure out, “Hey, do we think this is a deal that might violate the antitrust laws, that may substantially lessen competition?”

In any given year, two percent of all deals get a second look by the government, and an even smaller percentage ultimately we end up suing on. So the vast majority of deal-making—98%—goes through without the government even asking a second question.

I think we are at a moment where there's a recognition that the kind of more hands-off approach that the government had over the last couple of decades really led to consolidation and allowed some of the dominant firms today to kind of buy up firms that they were threatened by.

The FTC, before I arrived at the agency, filed a lawsuit against Facebook, noting that when the market was shifting from desktop to mobile, Facebook realized that that could be a threat and that they needed to survive that transition to mobile. They tried to do so organically but couldn’t really cut it.

So that’s when they made a series of defensive acquisitions, buying up Instagram and WhatsApp to kind of take them out of the market as independent players. The Justice Department brought a lawsuit earlier this year against Google, given its role in the ad tech space, and they similarly in that lawsuit identify acquisitions that Google made that they allege were anti-competitive and designed to really roll up the market in unlawful ways.

So those are concerns that we think about. We’re very sensitive to the idea that for startups and Founders, you know, you want to have the exit options. What we also hear though is that if your exit option is just one company, that’s going to depress your valuation; it’s going to give you less leverage in the types of deal terms you’re able to get.

So even if you want to exit through acquisition, having more competition among acquirers is ultimately going to be better for the startup in terms of having more leverage in that negotiation, being able to get a better valuation at the end of the day.

Yeah, I’m going to go off script a little bit, and I guess I just spent the week in Singapore. One of the craziest stories I heard about Singapore when Uber went to Singapore: a lot of jurisdictions threw up roadblocks, didn’t want them. I met the former Transportation Minister in Singapore, and they literally said, “Okay, here’s how you do it, and oh, you’re going to need a union. We know the unions; let's make a union for you.” And they sped up the process for Uber to enter.

And then it just came up because, is it possible for these acquisitions to get some sort of private approval through the government? I really feel for my friend Dylan Field right now just because, you know, he has a management team, he has a business that he’s trying to run, and he’s trying to run it basically with total uncertainty about whether the Adobe deal is actually going to happen or not.

I think it’s been months; it might be going on a year at this point. I’m sort of curious, like, are those things that, I mean, Singapore is very different from, you know, how giant the United States government is. But, you know, what do you say to things like that? I mean, is it possible to collaborate?

Maybe not specifically with like private approval, but you know, how open is the government to that sort of thing? For many of us, this is actually sort of the first time that we might have spent time with someone of your stature from the U.S. government.

So, we certainly can do like information exchange with other enforcers around the world. But ultimately, each country has its own laws and regulations, and so we have to look at, you know, and deals can affect different countries differently, right? So we need to look at the facts specific to the U.S. and figure out how do we apply U.S. laws here.

Sometimes different jurisdictions come out different ways because, you know, the market effect is different or their laws allow them to do different things. We are thinking about what else we can do to provide more clarity and certainty in the market. So, this past summer, we along with DOJ released draft merger guidelines that lay out clearly, you know, what are some of the analytical tools and frameworks we’ll be using to review deals.

We’re also updating what’s known as our HSR that’s going to, in some instances, increase the information we’re collecting on the front end, which we’re also hoping will be able to expedite our review.

So let's move on to the next topic, which I think a lot of people are thinking about and worried about, which is AI. There is an executive order that the Biden Administration just released on AI which dropped Monday. There has been some backlash from folks from our community around especially smaller AI startups.

You know, one of the sections that was mocked widely—though before we mock it, I think there are very important parts around it, certainly around camp, you know, preserving privacy, competition, and then, in particular, supporting skilled immigration—those are all incredible pieces of the order.

But one part that I wanted to ask about, you know, if there’s a model running more than 10 to the 26 floating-point operations, for instance, you know, sort of like banning math, which seems nearly un-American to me. I’m sorry if I’m editorializing, but there are call-outs to encouraging competition.

Can you talk about how you view competition in the AI space? What are the bottlenecks, and what role do you see for open source in this discussion?

Yeah, so there was a lot in the AI executive order, and a lot of it affects other agencies and other parts of the government. The FTC’s role, as you noted, is really about competition and privacy.

I mean, I'll just say up front, you know, this could be a revolutionary technology, and I think historically we’ve seen that these types of technological inflection points can be super important in terms of opening up the market and allowing more competition to enter.

So, you know, I think the way in which the antitrust lawsuit against Microsoft made it sure that, as we saw the expansion of the internet age, this next set of companies were able to enter and inject more competition.

I think there are very reasonable concerns right now that the critical inputs into this technology may already be kind of controlled by a handful of companies and that we may already be seeing bottlenecks in ways that could impede innovation and impede competition.

What we're doing at the FTC right now is really just making sure we understand the full anatomy, right? What are all of the layers and sublayers, and where do we see some of those bottlenecks?

So, be it in GPUs and compute more generally in Cloud, how are we thinking about the data, the models, the model access, and where do we see bottlenecks? In what instances are those bottlenecks due to supply constraints that could be addressed through investing more in production?

Some of what we’re seeing DC do in terms of the Chips Act and really trying to incentivize more production here in the U.S. could help relieve some of those bottlenecks. Some of those bottlenecks may be because these layers are highly capex intensive, and you're going to see huge economies of scale and data feedback loops in ways that is going to be more difficult to inject competition.

Some layers, you may in fact be able to envision more competition, such as in the apps. What you want to make sure is that the firms are not using the bottleneck to squash competition in those layers. I think especially when you have layers that, you know, there are a lot of economies of scale, open source can be a really key vector of opening up the market and having more competition.

We definitely see that as an important tool. That makes a ton of sense. Well, with that, we did want this to be really a conversation with our overall community, and it's such a rare moment to be able to get your time here, just a meeting of the best minds in the world here.

So, with that, I'd like to open it up for questions. Who here is a brave first volunteer to come up here and ask a question, please? I’ll just say as we’re going to the question and answer, it’s really helpful to hear that, you know, the executive order landed here in a particular way and folks had certain negative reactions. I’ll just say, you know, that’s partly why I’m here.

I think it’s really important for DC to be engaging directly with founders, with startups, and getting a better sense of, you know, what are the risks that you see? What do you think the executive order gets right? What do you think it should be doing better?

Especially given our mandate from a competition point of view, like what are the things we need to be looking out for? Please, introduce yourself.

Yeah, hi Cher Khan, I'm Alex. I'm from a company called Atmo. I really admire the work you're doing. I think you have a very bold vision, and I think you’ve taken really bold action. The question I wanted to ask was about the FTC and venture capitalists.

So, a lot of us here raise a lot of money from VCs, and a question that is often put to us when we raise money is, how are you going to build moats, and how are you going to become a monopoly? People even use these words very literally, very directly. They’ll say, “What are your moats? How are you going to be a monopoly?”

And they say in very explicit terms, “How will you do that task of pulling up the ladder behind you after we fund you, and you become really successful?” So, I’m curious about your thoughts on, you know, what it means when VCs ask that, and what you think the nexus between the venture capital world and the FTC should be.

Peter, we actually have a viral video on our YouTube channel about “Competition is for losers,” right? Awkward laughter. I mean, I’ll say, first of all, we have a complaint database. We collect information and are always eager for information.

But no, look, it’s a good question. It's natural for when, you know, companies are looking to enter and scale, like it’s not illegal to want to be successful, right? I think under the antitrust laws, there are certain ways of competing or undermining competition that are illegal, right?

Once you do become dominant, engaging in anti-competitive practices that are pulling up the ladder, like that is problematic under the law. I’ll be curious to hear from you all: is this universal across VCs? You know, do you see some variety in VCs?

We’ve also heard some concerns that, you know, the VC ecosystem itself could probably do with a little bit more competition. Those are things that we’re always interested in hearing about as well because, you know, when you see homogeneity in any layer, including the VC layer, that also affects, you know, what is and isn’t getting funded.

You know, one concern that we’ve heard is that if VCs are dead set on exit by acquisition by one of the big guys, what does that do in terms of potentially distorting or short-circuiting the innovation trajectory that we should be seeing, right?

I mean, historically, we’ve seen that the breakthrough innovations have come from outsiders, right? The startups that are able to see something that the existing giants are not seeing. So if the only things that are getting funded are what people are imagining the existing incumbents are going to want to buy, I worry about what that does to innovation.

Does that really short-circuit the more breakthrough innovation that we should be seeing?

I have a concrete example of this that I sort of wonder about on a daily basis. You know, I have an Amazon Echo, and you know, large language models have been out for years at this point. Like why is it that we can't speak to—why isn't there an open platform that allows, you know, OpenAI or an open-source model to come in and replace Siri, or replace, you know, that should happen.

But, you know, a lot of these conversations sort of boil back down to, well, it’s someone’s moat and they bought that. They want control of that so that they can make sure that you have enough toilet paper in your house, right? Like that’s the definition of anti-consumer.

Do you think, like, even philosophically that’s where the role of the FTC or the role of the government comes into play? You know, there’s fair play and then there’s unfair play. Obviously, it’s a very hard job to figure that out, but is that one of the things that you think through?

Yeah, absolutely. I think what we’ve seen through some of the lawsuits that have come out, again, some of the dominant platforms is that, you know, we see a life cycle, right? In early stages, firms can be very pro-consumer; they can be engaging in practices that help them build up their user base.

But once they switch to the model of successfully pulling up the ladder, they can really start exploiting their monopoly power with impunity, right? You can kind of become too big to care and start screwing over your customers without real repercussions in the marketplace that you would otherwise face if there was competition.

So I think that’s kind of, in some of these lawsuits, the story that’s being told. We’re at the extraction phase of the monopoly life cycle, and everybody’s losing out. The startups that can’t enter the market are losing out, but the existing customer bases are losing out too.

The FTC's lawsuit against Amazon yesterday—a whole bunch of additional stuff became public, including the fact that we allege Amazon was kind of polluting its search results page with ads even when they knew those ads were irrelevant and actually steering people to more expensive products.

Similarly, in some of the Google litigation, there are claims of practices that seem flagrantly anti-consumer. I think having more competition would make sure that that is getting addressed too.

Pep, over here had a question.

Thank you so much, Cher Khan, for your time. So, a lot of us have benefited significantly from open source over the years, and we’ve also been contributors to the space as well.

The question I have, it’s not really clear reading the executive order exactly how some of that applies to open source systems. Now, as it currently stands, there are foundation models that are being released with open weights, with a lot less visibility on how these things were trained. But then other people are taking these foundation models and forking them, fine-tuning them specifically to make new ones.

What is the FTC’s perspective on the applied use of those fine-tuned models? And two, if a company decides, let’s say one of us here, to start training a new model, at what point does it actually become subject to government review?

Yeah, it’s a good question. The executive order was crafted with a lot of other agencies and a lot of input and engagement. The FTC’s involvement is really just on the competition and privacy side.

But I think those are really good questions, especially on the open-source side. I will say, you know, one thing I’m really curious to hear from you all about is, in the past, especially in digital markets, we’ve seen what’s been called the “open first, close later” strategy.

Firms will start off with more open ecosystems in ways that benefit, you know, businesses that are able to build on it. Also help the platforms in terms of, and that lets them scale more quickly. But then at some later point, they switch from open to closed, and that can really have devastating consequences as well.

Are those similar risks here at all, as well? And how would you see those playing out, the value that you’ve helped create, suddenly you’re being locked out from?

Yeah, I mean, it cuts both ways, right? So if you're a brand new company and you’re starting something and you’re training a brand new model, which, quite frankly, not many of us can do because we’re capex constrained, in the scenario, we do that, eventually, there may be a situation where what you’re describing comes true, right?

But right now, the foundation models are kind of trained and controlled for the most part by pretty big companies. So they could very easily stop doing that.

Let’s take OpenAI for example. OpenAI has a consumer product; it’s actually something that’s powering their API, it’s powering my product. But the moment they stop permitting my use case, essentially I’ll have to switch over to something new, and that can happen anytime, right?

So is it a “available API forever” perspective, or is it, “Hey, you really cannot do these XYZ things?” I mean, what are the constraints on actually training and using an open-source model beyond the 10 billion parameters limit? Is it actually possible for all of us here, for example, to contribute engineering time to train an actual open-source model that is in direct competition with the big ones?

There’s no clarity on that, so this is a question that I’ve definitely been asked a few times, and I’m like, well, I don’t have any information. I can definitely follow up posit it, and an opportunity arise.

Hi, I just wanted to challenge your point on innovation and big players. Some of the most innovative papers that we’ve seen have come out of monopoly companies' research labs, right?

There is this concern that if these proposed breakups do happen, we lose the ability to fund these, you know, relatively expensive research projects that might yield something like, you know, “Attention is All You Need.”

So what are sort of the FTC and government's thoughts on that? There is a real risk that if they lose that status, they stop being able to fund research that could be extremely valuable to both open-source communities and Founders. So just curious to hear your thoughts on that.

Yeah, it's a really good question. It kind of goes back to this age-old question of what are the market conditions that best favor innovation, right? Is it monopoly or is it competition?

There has long been an argument by the school that it’s actually monopoly that best favors innovation because you can sink your monopoly profits into all these lines of research. Then the Kenneth Arrow school said, “No, in fact, it’s competition.”

You really need to drill down to the question: what kind of innovation are you talking about, right? The empirical research suggests that monopolies may be better at what is called incremental innovation—marginal improvements on the kinds of technologies that already exist out there.

But historically, the kind of breakthrough, paradigm-shifting innovations have actually come from outsiders. So having an ecosystem that’s open to startups and people who are looking at things differently and are going to come up with ideas that the existing incumbents are not thinking of—that’s actually been a really critical way to have more paradigm-shifting innovations.

More generally, I mean, you know, we’re a law enforcer, and so we have to enforce the antitrust laws, and those reflect a choice by Congress to say, as a general matter, we’re going to favor competition over a monopoly.

That’s why the laws are designed to protect competition. If Congress says, “Hey, there are some areas where we think monopoly is the way to go,” they can do that separately, but it's not really for us to say, “Hey, we think a monopoly is doing good here, so let’s allow them to exist.” Our job is pretty limited; it’s just, you know, enforce the competition laws.

Take one over here.

Oh yes, please, hi Cher Khan! Thanks for coming to San Francisco. I’m George Abosa, CNBC. You’ve been on our air before. I have two questions for you.

My first one is, you’ve been here a few days now. What have you learned? What’s been the most interesting thing you’ve learned? Will it change what you say or what you do when you go back to DC?

The second question is, one of the biggest criticisms of regulators is that they’re fighting yesterday’s battle. When you’re looking at Amazon and e-commerce or the DOJ looking at Google and search, they’ve moved on to the next big thing.

How do you make sure—and there’s a lot of anxiety here—that the big players are going to dominate the next generative AI shift? How are you looking at that and making sure that that doesn’t happen?

Yeah, good questions. I mean that’s partly why I’m here—to be hearing from and meeting with Founders, startups, and little tech. Because, you know, in DC, more established companies have very established ways to be sharing their thoughts, be it with Congress, be it with regulators, and it’s really the more kind of little tech and founders who understandably don’t have the resources to hire expensive lobbying shops.

Those are the voices that we don’t always hear from, and I think you’re absolutely right that we’re at a really important inflection point.

As you all know, there are a lot of conversations in DC happening right now about AI, about what the government should be doing, and I think it’s really important that we get this right. In order to get it right, we need to be hearing from a broad set of people.

That all includes you all. Already in the conversations, we met in a smaller setting earlier today with some startups and folks who are going to be launching new services, and it was just really instructive to get a sense of what they’re already seeing in the market in terms of some of the practices by these incumbents that could be risking solidifying their dominance and locking out some of the new competitors.

So, I will definitely take that all back to DC, and especially as we want to make sure we're staying vigilant in the AI space, those conversations are definitely going to inform where we’re looking.

Right here.

Oh, hi Gabriel from Coverage Cat here! My question was about how the FTC and maybe the government enforcement agencies writ large think about more pervasive price coordination through software that we see happening across different types of markets.

Not just through software, but also things like consulting firms. Whether or not you felt like the government had the adequate tools to deal with these types of anti-competitive behavior, or if Congress needs to address these matters and pass additional legislation.

Yeah, it’s a very timely question, and as enforcers, we’re definitely tracking this. There has been a whole bunch of reporting, including in the real estate space, noting how there’s software that may be allowing that coordination.

The Justice Department also brought a lawsuit recently in the agriculture industry, where they noted that this company called Agrat was basically collecting pricing data from all of these entities and kind of allowing them to coordinate and set prices in a way. So that lawsuit’s really important.

I think one of the interesting questions that we face is, you know, what price fixing looks like changes based on, you know, the technology that’s being used. Whereas, like a hundred years ago, it may have been a handshake in a smoky room, now it can look different, right?

So, what it looks like for algorithms to be enabling that collusion, I think is going to run into some questions for the courts about, you know, one is kind of rapid price matching, just competitive price matching versus colluding.

I think some of the lawsuits that we’ve seen of algorithms that are enabling kind of market-wide collusion may be an easier case. So I think it’s a very live issue.

At the FTC, we’ve been hiring up with technologists because we understand that when we’re looking under the hood, we need people and specialists and experts to actually understand what’s going on there.

So already hiring technologists has made it so that, you know, when we’re bringing lawsuits, there are certain algorithms that we can call out that without those technologists, we wouldn’t have even understood what those were doing.

So, you know, we’re still in the early stages in terms of fully being able to scale up and hire in the ways that we need, but it’s forward movement.

So speaking on the demonopolization, I think I’m curious if you’ve thought about how the DOJ looks at universities in terms of having nonprofit governance.

Pretty similar to how OpenAI is doing right now. I spent a decade trying to build a for-profit university and kind of ran into some of this.

But I think in shy of breaking apart companies are there other sort of structural or governance-based regulations that you’re considering to ensure there’s less extraction and more value creation for consumers?

It’s an interesting question. I hadn’t thought about the kind of nonprofit designation as a key tool. But historically, there have been kind of a set of regulatory paradigms that have had a seat at the table.

So on the one hand, you have antitrust and competition tools, which are primarily about decentralization. But you’ve also had what have kind of been known as the public utility tools, which have been applied in contexts where you realize that competition really wasn’t going to happen.

So be it in railroads or telecom, there are such strong economies of scale and strong network effects that you’re going to have a relatively small number of players.

In that context, we’ve applied tools like non-discrimination, kind of similar to net neutrality in other contexts, like you can’t pick and choose winners and losers; you have to offer equal pricing, equal service to everybody, with the recognition being that if you have parts of sectors that are so critical for commerce, you don’t want the gatekeepers really discriminating about who has access to the market and who doesn’t.

There have been other tools like mandatory interoperability. To go back to the question of, you know, when you have concentration and lack of competition, what are the vectors that can open that up?

Open source may be one of them. Interoperability might be another one, right? If you're a multi-homing entity, that can be a key tool to inject some competition.

There have also been tools like structural separations to say, “You know, if you own a railroad, you can’t also own a coal mine because otherwise you’re going to allow your coal on the railroad and kick off the coal of the independent producers.”

So again, a recognition that when you have concentration, allowing vertical integration can create conflicts of interest that can have real distortionary effects on the market.

So that kind of public utility toolkit is one of them. Another one has been public options, so the government actually standing up its own service and injecting competition in the market that way.

A bunch of municipalities over the last decade have actually tried to have their own municipal broadband to inject competition against some of the private entities. Those are just some of the other kind of regulatory tools that traditionally we’ve looked to when competition or antitrust weren’t enough.

I guess one of the things I’m noticing in your responses is that there is sort of this desire for a human flourishing of a sort, sort of backing how you’re approaching this, right? Like, what is the thing that might actually help people get better outcomes?

And usually, it comes back to the things that you’ve been saying, like competition, no self-preferencing, a level playing field, that sort of thing. So that just—I couldn’t help myself to editorialize.

So, yeah, I mean, I think the other fact is, like, you know, there's always uncertainty. The government's not going to be in the best position to necessarily be predicting the future; that’s not our job, right? That’s not our expertise.

So what are the best market structures that we can help create and protect to make sure that the best ideas are rising to the top? Traditionally, you know, competition and fair competition and competing on the merits has been the way to do that.

Thanks.

Maybe one over here.

Matt.

Oh yeah, thanks! Matt Tabergalis. I have a startup; I’ve also spent 20 years in Democratic politics, and I’m very excited to have you here and what you’re doing.

My question is on alliances and even investments. I’m curious what your perspective is— you know, for example, is a hyper-cloud investing in an independent AI company? Is that really about competition, or is that about capturing spend back to the hypercloud?

There are so many examples, I think, in this industry where the nature of the integrations and the structure of the investments might be relevant. I’m just curious what your point of view is on that and what your involvement might be.

Yeah, it’s a really good question and something that we’re thinking hard about. In the past, we’ve seen sometimes companies resort to what are known as kind of avoidance devices.

They’ll structure a deal or structure an investment to try to sidestep antitrust review. We haven’t looked closely enough at all of the investments happening right now to figure out is that what they’re designed to do, but it’s possible that there may be certain loopholes that are, you know, investments being done for the purpose of achieving control for the purpose of exerting influence in ways that may be undermining competition.

But it may be done in a way that’s, you know, not mapping on 100% to some of the existing laws, so it’s something that we’re thinking about and I think is very timely right now.

Maybe we do two or two more, please.

Hi Cher Khan, I’m Saba. I had a question around how you’re thinking about consumer privacy protections, especially as there are massive foundation models that have been trained on swaths of, you know, consumer data and copyrighted data as well, allegedly.

I was just wondering how the FTC is thinking about regulating that or thinking about it.

So we are, you know, the de facto privacy enforcer in the United States in the absence of some type of federal privacy law. So, we’ve been very active here, and we’re very mindful of the fact that, in the same way that behavioral ad-based business models that were monetizing people’s data created a race to collect as much personal data as possible, that the shift to large language models and AI could create just another set of incentives to be endlessly vacuuming up people’s data.

We’re aware that some of those business models could create incentives that are in tension with protecting people’s privacy, and I think that creates some challenges for us.

You know, we’ve been clear that there’s no AI exemption from the laws on the books, and so we’re going to kind of continue enforcing the laws. I think we do face some challenges.

You know, just the other week, we held a round table with creators, so we heard from artists, authors, fashion models, just people in the creative business who shared that, you know, a lot of them have woken up one day and realized that their life's work has suddenly been fed into these machines and it’s kind of spitting out results that are now harmful to them or really diluting the value of their work.

It was just really eye-opening to hear how that’s already affecting people. And you know, they were saying things like, “We are fully appreciative of how these technologies could really help us. We’re not anti-progress, but we think the fact that there’s no consent model right now where we get a chance to decide whether we’re opting in or not could be really damaging and really change over the long term people’s incentives to create and produce.”

So I think there are a lot of really fast-moving developments, and we’re trying to track them as much as possible. Some of them are going to be in our wheelhouse; some of them are going to be in other government enforcements' wheelhouse.

But we’ve been fairly active in the consumer privacy space. We’ve been particularly focused on people’s sensitive data. So we brought a lawsuit relating to geolocation data against this company Coava that had been basically making people’s sensitive geolocation data available for third parties.

Third parties could access fairly easily, you know whether somebody was going to an addiction facility, addiction recovery facility, whether somebody was going to a reproductive health clinic, whether somebody was going to a religious place of worship—like fairly sensitive data about people that I don’t think people assume that just because you’re carrying a phone around with you, that data will be available on the open market.

So that’s a lawsuit that’s still pending. We’ve also been fairly active in the health data context, so people’s sensitive health data—making sure that if apps are collecting data for the purpose of providing you some type of health service, they’re not then turning around and sharing that sensitive health data with third parties.

So, we’ll continue to do that work. Last one, Luther, who do you have in the back yet?

So I’m going to go.

Okay, thank you.

Hi, this is Jason. I’m a Founder in the AI space, so all of that’s very relevant. I have previously worked at like Google and these giants, so kind of see what they’re doing as well.

I guess, what I’m interested in learning about is how the FTC works with other government institutions or even like other countries. Like the EU, for example, they have GDPR. I think their rules are a bit more aggressive compared to the states.

Particularly, I was kind of inspired by the case of the Blizzard acquisition, where Microsoft kind of bought out like the competition from Sony. I know the FTC was against that case, but what’s interesting to note was I looked into the story a little bit, where Microsoft just negotiated with all of their competitors, like with other studios, and then they licensed the rights to Call of Duty, which is like the biggest kind of thing in tension between the case.

So that—there are two parts of the question. The first part is how do you work with other kind of agencies? The second part was what’s your take on that case, and do you mind sharing a bit more?

Yeah, unfortunately, that case is still pending; it’s on appeal, so I can’t say too much about the specifics. But on the first question, you know, the FTC has a long bipartisan tradition of building up its international engagement program.

There are a lot of agreements in place with other enforcers to be able to share information, and so we do that. At the end of the day, we have limits on how much we can coordinate because every country has its own laws.

Every case affects different countries differently, and so oftentimes you’ll see countries come out with different outcomes. But behind the scenes, at the very least, we can generally talk to each other, share information, and make sure that we’re being good partners.

Cher Khan, thank you so much for being so generous with your time. Thanks so much. I’ll hang out a little bit, so if you have individual questions or private questions, you know.