todd-davies

Correct, the click through from AI Overview is better than nothing, but it's not typically (anywhere near as) high as it would have been if the website was shown in the results without AI Overview at all.

Furthermore, opting out of AI Overview means that _other_ websites will be used for its grounding. Those competing websites will be cited and receive some citation traffic instead. So there's really very little incentive to opt-out for most websites.

Disclosure, I recently wrote a detailed write-up of these issues here: https://academic.oup.com/jeclap/advance-article/doi/10.1093/...

The argument of technical impossibility seems implausible, although it would no-doubt be difficult. Whether there can be enough engineering will mustered within the company to separate out the AdTech stack from the rest of Google's services is another question.

In terms of how the AdTech stack could look afterwards, here is a paper (I authored) advocating for the introduction of an interop layer such that users could choose which advertising network to use with Google's products and services: https://doi.org/10.36633/ulr.1113

I think the competition/antitrust law community is beginning to develop some effective antibodies against some of these. Bork's ideas are pretty well and truly debunked these days, and issues with regulatory capture, lobbying etc. are getting lots of attention (e.g. see this excellent paper: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4979205).

Absolutely not! But I've never been a materialistic person and lived below my means while I worked in tech, so my quality of life didn't really drop. Although academia isn't anywhere near as lucrative as tech was, I find my work really quite fulfilling now, which counts for a lot. I have no regrets!

For posterity, if you're reading this and are interested, feel free to drop me an email if you have questions.

I guess it depends on your circumstances. In Europe, for instance, the cost of a degree is sometimes quite low. My gateway from tech to law was a part-time masters degree in political science, and which cost around 200 euros a semester (in Germany). That degree gave me enough experience to then apply for a PhD in law.

Which brings me to the next point. Doing a law degree and passing the bar is perhaps the obvious path to doing policy things. It’s basically the only way that you can end up actively participating in courts, for example. But there are many other options! For myself, the plan is to stay in academia and not take any bar courses (then again, who knows what will happen!). Academics have lots of potential to shift policy, especially as neutral agents who aren’t paid by either side of particular debates. Our papers are read by policymakers and judges, who often don’t have the time or resources to think deeply about particularly gnarly topics. But there are lots of other options which could also work, and I guess finding a "niche" would depend on your specific circumstances, connections and skillset.

If you’re looking to spend more time thinking about policy issues, I’d start by simply sleuthing online. Bruce Schneier, for example, regularly writes excellent pieces at the intersection of technology and policy, which are very well hyperlinked to other high quality stuff. These kinds of blogs are a great way to get into the space, as well as to learn about opportunities which are coming up. Reading journal articles that sound interesting is a good option too (and US law journal articles are often quite accessible). There are also spaces offline, such as conferences which encourage both law and tech people (there’s one happening in Brussels soon [1]), or even institutions set up specifically to operate in this space and which have in-person events (Newspeak House comes to mind [2]).

[1] https://www.article19.org/digital-markets-act-enforcement/ [2] https://newspeak.house

Many of us in the antitrust/competition law community are trying. One issue, specific to digital markets, is that the field has very few people who are both legally and technically literate. If you're a technical person looking for a career shift, moving into legal policy/academia has the potential to be quite high impact for that reason.

Criminal sanctions in competition/antitrust law cases are an option in some jurisdictions, notably in the US and the UK (but not in the EU).

For an ageing but interesting case, see https://en.wikipedia.org/wiki/Lysine_price-fixing_conspiracy

There are a few ways to think about this. One is deterrence based on cost-benefit analysis, which is essentially a game theoretic way to think about firm behaviour. The logic here would be to fine the firm enough to deter anti-competitive behaviour, as has been mentioned.

Another way to think about it, is to say that we care about safeguarding the process of competition itself. That could include ensuring that competition is fair, ensuring that firms can enter markets, ensuring consumers get to choose which firms to consume from, etc. There's lots of precedent for that in EU competition law [1]. Taking that view, Apple was using its dominant position to restrict the economic freedom of Spotify (and others) and thereby harming the process of competition. Specifically, it limited rival firms from "fully informing iOS users about alternative and cheaper music subscription services" (as per the press release), thus harming competition.

All that to say, if we take the objective of EU competition law as being to prevent large firms from exercising power over smaller firms and to protect the process of competition in a general sense, then these big fines are easier to justify.

[1] https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3166005

(edit) TL;DR: see tivert's comment.

In most cases, the Commission sets a fine which is based on the harm caused by some anti-competitive conduct, with relatively small adjustments for extenuating or attenuating circumstances. In this instance it's the opposite; the economic harm was small but the adjustment was huge.

You're right that the logic - deterrence - is the same in both cases. But what's different (at least in my view), is the object of the deterrence.

Ordinary fines are designed to make anti-competitive behaviour unattractive in terms of the costs and benefits. Maybe some underhanded conduct generates €40m extra profit, but the risk of a €40m fine plus legal costs and adjustments makes it not worth it.

The trouble is that these fines might are essentially just rounding errors for large firms. In this case, a €40m fine would be tiny in relation to Apple's revenue stream (~€350bn euros a year), thus not an effective deterrent. That's for two reasons. First, the 40m figure is too low since a "significant part of the harm caused by the infringement consists of non-monetary harm, which cannot be properly accounted for under the revenue-based methodology as set out in the [Commission's guidelines]"[1]. Second, the fine is trying to to "deter [Apple and] other companies of a similar size and with similar resources from committing the same or a similar infringement"[1] even when they could absorb the ordinary (small) fine as essentially a rounding error on their cost of doing business. In that case, large conglomerates could basically just ignore competition law.

So here, the Commission is deterring all firms which have a "particularly large turnover" [2] (e.g. Big Tech firms) from using their power in one market to gain an advantage in another market, as in this case where Apple used its control over its App Store to gain an advantage in the music streaming market. The fining guidelines allow for fines to be much larger (~50x in this case) for tech giants, even if the actual infringement didn't cause that much quantifiable harm.

You're right, there's no restitution here. As you say, the fine is payable to the EU and would be paid into the EU budget.

[1] https://ec.europa.eu/commission/presscorner/detail/en/ip_24_... [2] Para 30 https://eur-lex.europa.eu/legal-content/EN/ALL/?uri=CELEX%3A...

Note that this fine is made up of 0.04bn of fine and 1.8bn of deterrent against future anti-competitive behaviour [1]. The the 2006 fine-setting guidelines allow the Commission to do that [2].

We should read the 1.8bn lump sum (roughly 0.5% of Apple's revenue) as partially being about music streaming and app stores, but mainly a warning to all large firms which are currently jockeying for a dominant position in emerging tech like generative AI and visual computing.

The warning: play fair and compete on the merits, or see you in court.

[1] "the Commission decided to add to the basic amount of the fine an additional lump sum of €1.8 billion to ensure that the overall fine imposed on Apple is sufficiently deterrent" https://ec.europa.eu/commission/presscorner/detail/en/ip_24_... [2] See paras 30 and 31. https://eur-lex.europa.eu/legal-content/EN/ALL/?uri=CELEX%3A...

The specific guidelines that the Commission is relying on were written in 2006 [1].

[1] https://eur-lex.europa.eu/legal-content/EN/ALL/?uri=CELEX%3A...

> In October, the Consumer Financial Protection Bureau (CFPB) proposed a set of rules that if implemented would transform how financial institutions handle personal data about their customers. The rules put control of that data back in the hands of ordinary Americans, while at the same time undermining the data broker economy and increasing customer choice and competition. Beyond these economic effects, the rules have important data security benefits.

No worries :) You're right that the law shouldn't be arbitrary. Lots of what the law is applied only when a cases passes legal tests to determine if some conduct violates the law. These tests are applied the same way to everybody and are thus impartial. Naturally, there's lots of debate and research into which legal tests we should use, and what their substance is.

Let's wait and see what the remedy will be in this case. For the past few decades, antitrust enforcement has mostly relied on fines as opposed to requiring firms do something differently. As a result, courts and agencies don't have that much practice figuring out what remedies to apply. This is an area where software engineers who understand what's technically possible can come in very handy.

The devil is in the detail, and there are a few ways that "[opening up] Google's devices", "allow[ing] other app stores on its Android phones and tablets" and allowing developers to charge customers "without having to fork over a cut to Google" could be implemented.

> It seems like the implicit assumption is that there must be a harm somewhere, we just haven't found it yet... isn't any business's primary purpose to compete with / harm competitors?

As a general rule, firms want to escape competition in order to make higher profits. There's nothing wrong with that! Indeed, the mechanism by which economic competition generates many of its benefits is that firms innovate in order to escape competition, and for those innovations to be useful for us all. So, where does competition/antitrust law come in? In part, it's about ensuring that firms escape competition in the way that we want. Innovation and competition on the merits is good, underhanded tactics to harm competitors is bad. All competitions need these kind of rules, regardless of whether they're economic, political, sporting, etc. When you have a large population of thousands of firms, you can be sure that some of them will be trying to compete unfairly, hence the assumption that there is some harm that we're yet to find.

> there isn't any consumer harm to be found, only harm to smaller competitors' businesses

We can distinguish between 'static' and 'dynamic' harms. Static harms are those which happen in the short run, such as a cartel agreeing to increase prices or not innovate. These harms are quite concrete and easy to define. Dynamic harms are those which affect the way a market might function in the future. For instance, a harm to innovation may result in people not having access to new products. It's hard to say for sure whether these harms will actually manifest, so we're usually talking about tendencies instead of certainties. It's perfectly reasonable to consider tendencies under the law though (e.g. we might prohibit drink-driving for the same reason). Dynamic harms usually have harm to consumers as a second order effect (e.g. reduced innovation or choice).

I can't think of a good example for a sass product. I'm sure it goes on though and I'm always interested in hearing about examples!

A similar strategy which seems to be quite common these days is to cross-subsidise, which is when a firm sells one product at an artificially low price by using profits it makes from selling another product. If we think about cross-subsidisation, then lots of multi-product sass offerings might fall under our scope. That said, cross-subsidisation has economic benefits, so it's not clear-cut.

As I said, to properly adjust to digital markets I think antitrust will have to identify new patterns of harm and invent new metrics to measure them. Predatory pricing (and similar offences) will always be useful, but they might just not fit well onto these kinds of markets.

Yes, it's a bit of a problem for the field! Like many aspects of antitrust, predatory pricing applies cleanly for an industrial-era economy but as you point out, it's less clear how to translate it into the context of 21st century informational capitalism. A significant amount of legal and economic research in the field is asking these kinds of questions, and the answers are still forthcoming.

Dropping prices below cost to wipe out competitors is predatory pricing which is prohibited under the antitrust laws. It's not always easy to prosecute, but it against the law nevertheless.

Great example. Another one is state governments joining a union or federal system, such as US states joining the Union or European states joining the EU, and becoming bound by federal/EU law.

For sure they are throwing lawyers at it. It's 13bn after all. But the underlying reason for the appeal is the same; the law is not yet clear as to what should happen.

What's the cause of all the flip flopping between Apple having to pay the bill or not? I think the main problem that this area of law is fundamentally uncertain and hard to apply. Legal mistakes get made, and judgements get appealed until people figure out what the law actually says.

A secondary issue is that judges in appeals courts are generalists, not tax or state aid experts. Despite this, they get landed with the hardest cases - situations that haven't been seen before which throw up new legal issues - and need to figure out what to do. Most of the time they get it right. Some of the time, they mis-interpret the law and get it wrong. That's why we have the appeals process.

So where are we up to in terms of this case, and how many more flip-flops we can expect? This article is citing an advisory opinion of one of the EU's Advocate Generals (themselves, very senior judges). These advisory opinions are issued prior to the CJEU (the highest court) making a final ruling. They are there to guide the CJEU so that the final ruling is as good as possible. Although this specific case should be finalised after the CJEU makes its ruling, I'm sure the case law will continue to evolve in the future as multinational companies push the boundaries of the EU's tax and state aid regimes.

Likewise, thanks for being polite.

I agree with you that regulation has costs. I never intended to communicate that it didn't. Yet, at risk of stating the obvious, the fact that something has costs doesn't make it not worth doing if its benefits are greater than its costs. The question then returns to the crystal ball; will society be better off if we allow or deny the merger? I get a sense that you are of the opinion that free markets generally lead to good outcomes. Me too! Central planning of economies is generally to be avoided. Yet at the same time, I view large concentrations of corporate power as an endogenous source of unfreedom in otherwise free markets. Big companies, if we're not careful, become mini central planners in their own right.

Are you suggesting that the only reason competition might not exist is because of government regulation? If so, that's not correct. Being free to enter a market doesn't mean that it's viable to do so. Barriers to entry are common in all kinds of markets so it's not as though people can always start competing against a weak vendor. Exclusive contracts, the threat of predatory pricing, increasing returns to scale, high switching costs, network effects, IP monopoly, geographic isolation, etc can mean that it's simply infeasible to enter a market and start competing against an underperforming incumbent. Local cable companies are a good example. If you want to start a cable company, you either have to lease cable from an existing owner or lay your own cable. The former is infeasible if your competitor owns the cable, and the latter may be prohibitively expensive. All that to say... merger control is an important tool which can help prevent markets from becoming more concentrated when barriers to entry are high [1].

Regarding WhatsApp & Instagram, we don't know what the counterfactuals are if Facebook didn't acquire them. It's not a dichotomy where either WhatsApp was either acquired or would cease to exist. Indeed, there could be another world where Instagram (a startup monetised via ads) merged with WhatsApp (another startup without a monetisation plan) to challenge Facebook. Perhaps that additional competition would have led to all sorts of innovations that we haven't thought of yet. Of course I can't prove that, but again, we're back to the crystal ball again ;)

Finally, it's important to acknowledge that antitrust is political. That's because public power (the state) is used to discipline/reshape private power (privately held firms). Furthermore, it's asking a very political question: who gets to coordinate economic activity? [2] Many other areas of law are similarly political (e.g. taxation law, electoral law), so antitrust isn't special. Yet it does explain to some extent why there are strong disagreements. That's okay :)

[1] https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1999829 [2] https://heinonline.org/HOL/Page?handle=hein.journals/uclalr6...

Power is notoriously hard to define. In antitrust "market power" is usually defined as either the ability to set prices above marginal cost, or an ability to act free from competitive constraint (i.e. if the firm takes some action, it doesn't worry about other firms responding and taking away some of its business).

On the point of consumers switching away from bad firms, you say that

> a corporation only exists if customers are happy with its services

I'm not so sure. There are many firms that exist despite consumers being unhappy with the products/services provided. The notion that consumers can take their business elsewhere is shouldn't be taken for granted. There are many times where that's not possible; maybe alternative firms don't exist, or they're not convenient, or they have products you like (even) less. Maybe you're "locked in" to the existing firm even though it's starting to provide a worse service, through a contract or because of some path dependency. Thus, we can't take it for granted that firms and consumers have equal bargaining power and that consumers can simply stop purchasing at any time.

Good point. I don't think the FTC is myopically focused on mergers, but on anti-competitive behaviour in general. That said, I don't think we have a good handle on the imitation problem, there's probably some great research questions to ask in that area.

The VC side isn't my area so I don't think I have very good thoughts on it. Perhaps somebody with more expertise can chime in. That said, there's lots of great work about financialization and antitrust coming out, and I should probably read up on it.

With that caveat, I think it's plausible that some startups are/were getting VC funding not because they were good businesses, but rather because they presented enough of a competitive threat to a big tech firm that there was a good chance of a large acquisition. The net effect of those acquisitions is probably to transfer some of the surplus from Big Tech companies into the VC/founder ecosystem. Is that good? Maybe for the founders and VCs. From a social perspective though, we want to ensure that startups can meaningfully contest incumbent firms' market positions, and that can't happen if all the startups get acquired before they get a chance to do so.

I think it's a false dichotomy to say that businesses either succeed by IPO/acquisition or fail. Can't a startup "just" be profitable within its niche, serve its customers well and make money while being privately held? It's not clear to me that we should live in a world where every middling startup is eventually acquired or has an IPO. Sure, some startups are destined for exponential growth and huge success, but not all. Likewise, if a business isn't profitable (and can't get funding to tide it over until it is profitable) then maybe it's not a good business. That's catastrophic for the firm, but on a social level, it essentially survival of the fittest.

With regards to personal freedom, we need to think more broadly than just founders. There's also the personal freedom of citizens to consider too; when powerful firms control large parts of the economy they essentially operate as private governments which can also impinge on people's personal freedom (see the below quote). Should a startup founder be allowed to sell a startup to monopolist/oligopolist if that contributes to more economic concentration? Maybe each individual merger isn't that harmful, but taken together their cumulative effect is. Besides, Khan isn't trying to block all mergers. She is focusing on mergers with already dominant firms; perhaps it's possible for the founder to sell a firm which isn't in a dominant position.

> ...power that controls the economy should be in the hands of elected representatives of the people, not in the hands of an industrial oligarchy. Industrial power should be decentralised. It should be scattered into many hands so that the fortunates of the people will not be dependent on the whim or caprice, the political prejudices, the emotional stability of a few self-appointed men. The fact that they are not vicious men but respectable and social minded is irrelevant. That is the philosophy and the command of the Sherman Act. It is founded on a theory of hostility to the concentration in private hands of power so great that only a government of the people should have it. [1]

[1] https://supreme.justia.com/cases/federal/us/334/495/

There are a sizeable bunch of antitrust scholars/economists/lawyers who are very much against stronger enforcement. For instance, see [1].

[1] https://laweconcenter.org/

I don't think it's correct to paint Khan as following a "big is bad" standard. For instance, in [1] she explicitly says the opposite:

> Antimonopoly does not mean ‘big is bad.' The New Brandeisians—like Justice Brandeis—recognise that certain industries tend naturally towards monopoly. This is especially true of networks. In such cases, the answer is not to break these firms up, but to design a system of public regulation that prevents the executives who manage this monopoly from exploiting their power. A second goal is to ensure that executives face the right incentives to provide the best service possible to everyone who relies on the monopoly to sell or to buy a particular product or service. In the past Americans have used both direct government regulation, and various forms of antimonopoly law and policy, to achieve these ends.

[1] https://academic.oup.com/jeclap/article/9/3/131/4915966

There are lots of different views about the goals of antitrust law and policy, but few (if any) are hostile to business. Khan is pro-business but also, broadly speaking, pro-dispersion of economic power within the economy. That means having fewer 'gatekeeper' firms which exercise power over key economic bottlenecks, and at the same time, a greater focus on helping smaller and medium sized businesses succeed.

For a concrete reference regarding small business, see this recent speech of hers [1,2].

[1] https://www.ftc.gov/system/files/ftc_gov/pdf/remarks-chair-k... [2] https://www.ftc.gov/news-events/news/speeches/remarks-chair-...

The "antitrust community" is people working in antitrust law/policy/academia who write, tweet, speak at conferences, etc. Two caveats are that a) this is my view based on personal observations, and b) that a consensus doesn't mean that everybody agrees. Undisputedly however, there is a lot of academic work these days which comes to the conclusion that merger review should be strengthened. For instance, see [1].

Khan's aim isn't necessarily to make mergers more difficult, or to prevent them per se. Rather, it's to make a more concerted effort to prevent mergers which on balance, appear not to be in the public interest. There are several reasons why a merger might be harmful.

First, some mergers are "killer acquisitions" whereby a large firm will acquire a "nascent" competitor and then discontinue its product (e.g. [2]). One danger here is that killer acquisitions nip start-up competition in the bud, such that new firms don't have a chance to grow and compete against incumbent firms. That's a problem because competition means that consumers, rather than incumbent firms, ultimately get to choose market outcomes. If there is no choice, then consumers can't choose. The ultimate failure mode here is some kind of command economy where monopoly firms get to make most of the decisions about how markets work and consumers get little say at all.

Second, mergers lead to market concentration because you're taking a market with n firms and moving to a market with n-1 firms. Market concentration isn't inherently bad, indeed, some markets are 'naturally' concentrated. That said unconcentrated markets are generally preferred to concentrated ones because concentrated markets can lead to things like tacit collusion (which has similar outcomes to a cartel) [3].

Third, if there is a real possibility for startups to grow organically and challenge large incumbent firms for the market, then theory goes that VC funding could be stimulated on the promise of a potentially huge return. For instance, a VC might be willing to fund a firm if there's a 5% chance that it will be the next Google, but not if there's a 10% chance that it gets acquired in a year or two.

You make a good point regarding acquisitions, IPOs and the personal incentives to work at a startup. I'm not assuming that IPO's would become more likely or lucrative; I don't have a view on that (except perhaps what I said in the previous paragraph). I guess if a firm was going to be successful, and a merger wasn't an option then it would eventually IPO instead, no? That might lead to delayed compensation, but it shouldn't affect the viability of a solid business. Fewer acquisitions might end up with some firms failing before IPO which would have otherwise been acquired. That's bad from the perspective of those at the firm, but isn't inherently bad from a social perspective. After all competition necessarily entails winners and losers. If all else fails, there is a 'failing firm defence' which would allow an acquisition if the only other option is the firm going out of business [4].

Finally, the lack of enforcement over the past few decades hasn't given regulators much opportunity to "learn" what is a good/bad merger. A more active merger review policy would entail regulators building up expertise and fine-tuning their approach. It's important to remember that merger control isn't necessarily adversarial. At its best, it's a positive-sum dialogue between firms who want to do business and regulators who are trying to provide public-minded oversight.

[1] Kwoka, John. "The structural presumption and the safe harbor in merger review: False positives or unwarranted concerns." Antitrust LJ 81 (2016): 837. https://heinonline.org/HOL/Page?handle=hein.journals/antil81... [2] https://news.ycombinator.com/item?id=38145568 [3] https://www.tutor2u.net/economics/reference/oligopoly-tacit-... [4] https://www.concurrences.com/en/dictionary/Failing-firm-defe...