But the tech giants run a single system: Amazon in the delivery, FB in the social graph (DB), Google in the knowledge graph. Those would be incredibly hard technically to break up. The less technically complicated split would mean that one new company has all the revenue and the other has all the costs.

I haven’t heard any credible proposal on how this “breakup” would actually physically work. If anyone has links to good sources I’d be interested.

Fair point - there are certain cases where software platforms are too valuable and complex to break up, and should be regulated like utilities with rules for fair play. Amazon shouldn't be able to use it's marketplace data to monopolize entire categories, Google Search needs to be a public platform, with regulated rates for API access (which Google itself will have to abide by).

You don't have to break up many software systems to prevent the worst abuses. Facebook can keep the knowledge graph, but they can't keep WhatsApp or Instagram. Amazon can keep their store, but they can't own a FedEx competitor or AWS - those need to be broken off. Google can keep Search, but Ads, Doubleclick, Analytics, Waze, GCP, Google Home - all of that needs to be broken up.

Some of the resulting businesses may have to update their model or may not remain profitable, but historically break-ups have resulted in an increase in value for the new businesses.

>Some of the resulting businesses may have to update their model or may not remain profitable, but historically break-ups have resulted in an increase in value for the new businesses.

All of this teaches one lesson: be more like Apple. Treat your customers poorly and massively overcharge them. Make sure you are never the biggest in your field. Essentially, don't offer products that are too good for the price.

I'd also like to point out that if those companies have to be spun off on their own, then the only way for some of them to make money is to sell your data to third parties. Right now Google doesn't seem to do that, but how else would Analytics monetize itself, if it cannot offer ads?

Charge money for the product. Plenty of other companies in the space do so successfully. There's ample business value provided by good analytics.

Google Analytics was originally the product of Google's acquisition of Urchin Software. At the time, Urchin's self-hosted version cost $895 with additional optional "modules" that cost up to $3,995 extra. Their "on demand" version cost $199 per month.

You’re wrong on both counts: Apple has its own share of potentially monopolistic behavior, and there are billions of people worldwide who disagree with you on how the company treats them.

But I rarely see Apple being included in these calls for breaking up the tech companies. Also, I'd be surprised if Apple even has a billion customers, let alone a billion who are happy with Apple.

https://en.wikipedia.org/wiki/Small_but_significant_and_non-...

Does it also have a monopoly on apps that can be run on the “HomePods”? Does every smart TV manufacturer with their own OS also have a “monopoly” on their ecosystem?

Let’s go further down the rabbit hole. Does Tesla have a “monopoly” on software that can run on its cars?

If non insignificant switching costs defines a “monopoly” every software as a service app would have a “monopoly”.

Unfortunately, neither the EU or the US has ever defined “monopoly” like HN posters...

I just explained one possible scenario how this could be defined as a relevant market and you're now throwing a bunch of random assumptions on my comment. I would read the link thoroughly (it explicitly mentions DoJ?) and study more on the history of the antitrust law and its applications before doing such.

If that were the case, every single console maker since the mid 80s would be declared a monopoly. Why hasn’t that happened?

I already told you that the market defining process is a very case-specific one and big techs now are unprecedented. Why are you trying to find applicable prior arts on such cases? You asked how Apple can be monopoly on a market and I gave you one possibility which may or may not materialize due to its uncertain and complex nature. It's pretty hard to understand why you're being so defensive on this issue?

> If that were the case, every single console maker since the mid 80s would be declared a monopoly. Why hasn’t that happened?

There has been multiple antitrust lawsuits and investigations on Sony, Nintendo and Microsoft for their gaming consoles while their exercise of monopolistic power was nowhere comparable to Apple's nowadays. Why don't you google just 2~3 words before making such a false claim?

Maybe because I don’t believe that people can randomly make up definitions instead of citing precedent?

There has been multiple antitrust lawsuits and investigations on Sony, Nintendo and Microsoft for their gaming consoles while their exercise of monopolistic power was nowhere comparable to Apple's nowadays. Why don't you google just 2~3 words before making such a false claim?

Well seeing that Nintendo specifically has been very strict about what was allowed on its platform, forced third parties to use its manufacturing facilities since the 80s, forced all software whether distributed physically or virtually to be licensed and to pay a fee and had a much larger marketshare, where was the government intervention? Where were the consent decrees? Lawsuits?

Please provide one citation where any of the console manufacturers were ever forced to change their business practices?

If my claim is “false”, you should easily be able to find a citation.

While you're trying to frame my argument as "make up definitions", the reality is not; this is a standard practice since 1982. Spend your time on searching and studying the topic, not mine. This is an area of vast complexities and I don't think it's effective to spend my time to enlighten you.

https://www.justice.gov/atr/operationalizing-hypothetical-mo...

> where was the government intervention? Where were the consent decrees? Lawsuits?

Your ignorant in the topic doesn't necessarily mean an actual lack of a prior. In Nintendo v. Atari case, there was not much arguments on the market definition, but its practice was illegal or not. Yes, I've been talking only about the market definition and you're intentionally conflating the concept of the relevant market definition and antitrust violation. Don't do that. And please don't even try to say "come up with evidence". You can spend your time on studying this.

And your citation were about hypothetical arguments, and went to show more of my point. All throughout the article it speaks about the government’s arguments being “defective” and still doesn’t show a single example where a vertically integrated minority player was called a “monopolist” nor where government imposed remedies or sanctions were implied.

In fact, Atari vs. Nintendo affirmed that Atari did in fact violate Nintendo’s copyright when it tried to circumvent Nintendo’s control over its platform.

So in fact, you still haven’t come up with a single precedent where Apple could be considered a “monopolist” on its own platform or where they are violating “antitrust”.

Or just don't abuse your customers by hiding the ball on what you do with their data. Don't buy up all your competitors, so you can squeeze more and more money out of small businesses who have to buy impressions because the organic ways (which the same companies own) don't work any more. The products aren't too good for the price - we don't even fully know what the price is.

> I'd also like to point out that if those companies have to be spun off on their own, then the only way for some of them to make money is to sell your data to third parties. Right now Google doesn't seem to do that, but how else would Analytics monetize itself, if it cannot offer ads?

The same way Mixpanel, KissMetrics, Adobe and dozens of other web analytics vendors monetize (including Google's own 360 Suite). Or make it clear, up front, how the data will be used on free accounts, and give the user a way to monitor and revoke that agreement at any time. As an independent business, they will be free to find the best path.

If you really want to change the landscape and make it actually feasible, force companies to expose APIs or impose standards on them. Both of these are very hard things to do in practice but at least they give a hope of a better future - increasing competition.

Breaking up big tech, depending on the specifics, will cause either worse productivity or (in the best case) change nothing.

The facebook conglomerate could similarly be broken into Instagram, Facebook, etc. Each would get a portion of their current ad business unit.

In terms of software, each baby Facebook could get a flexible royalty-free license to all the software currently owned by Facebook, and they would be free to derive from it to differentiate over time.

If there is a will, there is a way.

Turn the relevant parts of it (the ones which are affected by network effects, or other barriers to entry) into an open platform that can be accessed on an equal basis. That's quite a bit easier than breaking up a mining company. For example, Android is already "broken up" from this POV, since the likes of Amazon can use AOSP to create their own Google-free platform, and applications written to run on AOSP will work on either version.

Is it the index itself? There's already a fairly decent open-source equivalent to that in the form of Common Crawl, which has been out since 2011. People (including me) have tried to build search engines off of it, but it never seems to work quite as well as Google.

Is it the computing infrastructure? That's already been commoditized and offered as a service by multiple providers - AWS, Azure, GCP, SoftLayer, etc.

Is it the serving infrastructure? Commoditized by ElasticSearch, which uses many of the same techniques as Mustang and in many ways does it better.

Is it the ranking algorithm? It used to be that Google would agree with you. However, the ranking algorithm changes basically continuously - the one in use now is very different from the one that was used when I left in 2014, which was different from the one where I learned how it worked in 2011. I've heard the new one is heavily machine-learning based: given a suitable training set and some learning-to-rank papers you could construct something similar.

Is it the log & clickthrough data? Imagine the privacy advocate conniptions if that were open-sourced. AOL got in huge trouble when they open-sourced their click-logs circa 2002.

Is it the evaluation system? Mechanical Turk exists, and Google's rater guidelines are public.

I'd argue that the real competitive advantage of Google now is the brand and associated consumer habits, and it's really hard to break up a brand. Same reason Coca-Cola remains dominant 150 years after they started selling cocaine-laced sugar water. This is a recurrent problem in the economy today - brands fuel not just Google, but also other giant monopolies like Coke, Nike, J&J, P&G, DeBeers, McDonalds, Wells Fargo, and so on, and in many cases the companies that own them get to practice some exceptionally bad behavior. But short of reaching into each consumer's head and getting them to consider each purchase on purely rational factors, I don't see how to fix this.

It's a huge advantage and critical ingredient for ranking in all major websearch engines as far as I know. Google gets billions of queries and clicks every day. Your startup? Pretty much none of that. How are you going to train your ranking algorithm with no data?

> Imagine the privacy advocate conniptions if that were open-sourced. AOL got in huge trouble when they open-sourced their click-logs circa 2002.

Right, logs with such level of detail as individual user sessions would never get public for this reason, too much legal risk. But even simple aggregated datasets of the form "how many clicks did this query-url pair get" would be very useful to bootstrap a competitor and can be anonymized much more effectively.

Doesn't this run counter to private property though? Or what would you have happen when Google says "no"?

Counterargument: you can't copy the people who know how it works.

Google has:

- Android OS

- Nest

- Search

- Advertising

- Chromebooks

- Pixel

- YouTube

- Google Cloud

There are clear lines where they could be broken up.

I’m not saying they should be broken up.

>My biggest problem is: How do you break up a software system?

There are people who have created entire careers about exactly that: antitrust remedies. I defer to experts where I can.

We entrepreneurial wannabes on HN are only going to provide unrealistic answers that become battles of will for the rest of this Sunday before all is forgotten overnight.

All of the virtuals you mentioned have different stories. Cloud is profitable, and is not the top company anyhow. News is not a product it's a grouping of news stories. Youtube makes money through ads and better access and could be considered a loss-leader but shutting it down will not make the field more competitive. Android is open source.. and perhaps could be seen as dumping to prevent others. Gmail is a mail service, others exist.. and starting a new company will not cost you billions unless you plan on serving billions of people. Drive is one of many companies that didn't cost a billion to start but might be worth it now.. try dropbox or box.com or rapidgator.

A democratic country has an interest in ensuring that our labor and goods markets also remain open, competitive, and democratic in nature. We've made this choice as a country repeatedly throughout our history, and it's time to do it again.

“Shady data practices” is hand-wavey and non-specific. What do you mean? Many companies today (again, across All industries) have had issues with keeping user data secure, selling it to untrustworthy third parties, not giving users transparency or controls in what information is shared, etc. Google’s track record in these areas is far better than most. How does this necessitate a breakup?

Concentrating industry through M&A is a legitimate issue, in my opinion, but it’s also probably the easiest to regulate.

Our tech companies are the most competitive in the world, bar none. I’m not sure that cutting them off at the knees will help with that. ”We did it before,” isn’t a convincing argument that we should do it now, under much different circumstances.

I used a cover-all term, because there are too many to really name. Facebook is a co-conspirator in defrauding American elections, Amazon uses their data to kill small businesses, Google has been repeatedly fined in the EU for using search anti-competitively. Those are just 3 tech companies, and doesn't get into how data is used abusively by credit agencies, banks, and other financial institutions.

Regulating industry is a fools errand, and part of how we wound up here. Breakups are the only self-executing, corruption-resistant solution.

Are our tech companies the most competitive at in the world? The largest companies have been cutting off our startups at the knees for a decade, so we have no idea how competitive we could actually be. It doesn't really seem like our big tech companies have to compete much at all these days, actually.

So, suppose Google Search is broken away from alphabet. What changed?

I can understand that argument about Youtube (which is losing money and would probably fail), but google search will keep being dominant and their ad revenue won't change significantly.