Tech Companies and Competition

(Disclosure: my current employer is a subsidiary of Microsoft. I’m not a journalist or anything, and I don’t think that really influenced my writing, but it would have felt dishonest not to mention it.)

Daring Fireball, weighing in on a recent IBM executive shakeup:

I don’t offer this observation as an argument against any and all regulation and antitrust investigations of big tech companies. I’m simply arguing that regulation and antitrust lawsuits should be wielded with surgical precision, not broad strokes. Competition and progress work.

If you listen to his podcast Dithering you quickly pick up on the fact that John generally is in favor of letting the market do its thing, and his point is well taken that IBM spent decades being dominant only to have a much more diminished role in tech today. But the dominance of today’s tech companies feels substantially different from the dominance IBM had.

IBM Was Big When Tech Was Small

IBM was dominant in the world of tech when computers were something that only businesses owned and used. IBM did make some consumer tech in the 80s, 90s and early 00s, but their ability to be big was mostly due to the fact that they were big fish in a relatively small pond. That’s not true for the biggest tech companies today.

Entrenchment is Deeper

Social networks had relatively short lifespans before Facebook came along. They’d often fizzle out after a couple of years. Facebook has managed to reach an escape velocity that’s kept them from having that same issue. They were wildly popular in the aughts, and they’re even bigger today. They measure their users in billions, and they’ve been known to buy up companies achieving dominance in areas and markets they didn’t (see: Instagram and WhatsApp).

In a similar vein, there was a cottage industry of search engines before Google came along. Whereas Facebook got more powerful as a social network because social networks get more useful the more people are on them, Google largely earned their dominance by making a great search engine. But in doing this, they also became deeply entrenched in their position at the top and it’s structurally really difficult to unseat them.

Big Tech is Big With a Capital B

At press time, there are five companies with market caps above $1 trillion, and four of them are tech companies (Apple, Microsoft, Amazon and Alphabet; Saudi Aramco is squeezed in there at #3). Facebook is worth nearly a trillion.

Tech is big business now, and not only are tech companies now our biggest companies, they wield a massive amount of control over our daily lives in a way that IBM never came close to achieving.

If you participate in the modern world, you almost certainly are being served directly or indirectly by servers in a cloud run by one of these big tech companies. If you use a computing device, you almost certainly are using an OS made by one of those trillion dollar companies.

IBM’s unseating

Gruber’s right that IBM lost dominance, but let’s remember how that happened.

IBM’s downfall was in not anticipating the PC revolution. They falsely assumed that mainframes were where the big money was at, and when they made a PC, instead of buying exclusive rights to DOS from Microsoft, they licensed it. Dozens of other companies started making IBM-compatible PCs, and PCs sold like gangbusters, easily eclipsing IBM’s business selling mainframes to big companies.

But that only was possible in an age where there weren’t many computers. Today, most every person, home, and business has a computer, and often several. There’s not really any new category of device that can come along and eclipse the computers and mobile phones that Apple, Microsoft and Google power, or the servers that Amazon, Microsoft and Google power.

A side note: IBM didn’t really suffer a downfall so much as it ceded growth to other companies. Even in their diminished role, IBM continued growing throughout the 90s and aughts, with their market cap peaking in 2017 at just under $175 billion.

IBM missed out on a fundamental growth wave, but what fundamental growth wave is there for today’s tech companies to miss? Billions of people own a smartphone now, and Apple and Google have a technical head start of well over a decade of building mobile OSes. Even if you can imagine a newcomer coming along and taking over that market, there’s no longer much room for that market to grow a lot higher.

Assuming no major new policies that might shake things up, the big tech companies we know today are almost assuredly going to be the big tech companies we know in a decade or two. The last 20 years of tech have seen advancements comparable to the last 50 years of the 20th century. And not only that, the biggest tech companies are now in far more businesses than IBM ever was.

Competition is a powerful thing, but once your market cap is measured in trillions and you have almost no peers, it’s a lot less meaningful.


The “Big” in Big Tech Breeds Extremism

Tech companies hit a turning point this week and have actually started wielding power against far right groups aiming to incite violence. Most dramatically, Facebook and Twitter both permanently banned Donald Trump, and Parler, a “free speech” social network that minimizes moderation and has become a favorite among white supremacists, has been removed from Google’s and Apple’s app stores, and in an unprecedented move, Amazon will no longer let Parler run on Amazon’s AWS infrastructure in the cloud, forcing the entire service offline (as far as I know, Amazon has never banned a company of this size because of the company’s own lack of handling its users’ misconduct).

This has inevitably led to conversations about whether Big Tech has gone too far, and conversations about the role big tech companies should play in moderating what’s on their platforms.

I don’t want to get too in the weeds about whether tech companies made the right move; my longstanding opinion has been that these are their platforms and it’s up to them what they decide to let people do on them, and that I wish they prioritized the safety of their most vulnerable users. I will also acknowledge that as much as Parler had it coming, when you see the entire business get pulled offline in just a couple of days because of a few external companies, it’s only natural to have the panicked realization that these big companies could just as easily do this to your business if they really wanted to. As much as conservatives have been pretending to be oppressed by tech companies (and that’s a whole post unto itself, but I can guarantee you that “anti-conservative bias” is 100% bullshit), it really is true that for most of the tech companies you interact with, they could pull the plug on your account if they wanted, and you would have little to no recourse.

I do agree that it’s complicated for owners of big platforms hosting user-generated content to effectively come up with a set of policies that govern what people can talk about on these sites. These are hard questions because a huge chunk of the world’s communication happens on networks like Facebook and Twitter. If Facebook and Twitter get really strict about what you can say and do on these networks, they could be stifling much of the communication that occurs on the internet in general.

These will never stop being hard questions, but there is one straightforward way we can make these questions less necessary to even ask: take the “Big” out of “Big Tech”.

It’s the Scale, Stupid

Put simply: our society and institutions are not equipped to correctly handle the existence of massive tech companies that have literally billions of users.

Social networks aren’t a super new development; they’ve existed in one form or another since before the internet, and while there has always been some concern for fringe groups online, none of these groups or networks had previously posed an existential threat to democracy in the US.

But that changed when companies that run social networks hit real scale. And when I say “real scale,” I’m talking tens of millions, hundreds of millions, and even billions of users.

Facebook isn’t just a community; it’s a community of communities that is centrally managed by Facebook itself, where Facebook has an incredible amount of data about you. In the US there was a national debate for years about creating a new national ID card to replace state-issued IDs but there was this concern about centralizing control with the federal government. And yet, here Facebook stands as a central entity that tracks the identities of more people than the most populous countries on earth.

A network as large as Facebook is essentially a government. Facebook’s CEO has said as much on the record. But unlike most democratic governments around the world, Facebook’s policies don’t get determined by its users or people its users choose to represent them; Facebook is free to make these decisions unilaterally, and its users don’t have much recourse. “If you don’t like it, leave” is a tough sell when the network has pretty much every online user in the world on it.

Automatic radicalization at scale

On massive social sites like Facebook, you don’t need to find new communities; new communities will find you. Facebook will analyze your profile and activity and recommend new groups to join. On the surface, that sounds perfectly innocent; Facebook helps you find new groups you might like. But in a world where Facebook is home to all of these non-publicly visible extremist groups, that’s super dangerous, because now Facebook is doing the heavy lifting of recruitment for these groups. Facebook knows what those group members are like, so it can identify other people that might be sympathetic to these extremist causes and just casually recommend the hidden group to them. And just like that, Facebook just unwittingly became a tool to radicalize people at scale. Whoops.

It’s not a problem unique to Facebook; I just keep referring to them because it’s easier to point to a concrete example. It also isn’t just a problem unique to social networks. YouTube’s got a similar issue where if you follow recommendations on certain innocuous videos for long enough you go down a rabbit hole that often leads to increasingly extremist videos.

Critically, these tools to radicalize people through recommendations are only possible because companies like Facebook and YouTube have absolutely enormous numbers of users.

Facebook and YouTube are so big, when they make tweaks to their algorithms that promote content, it can destroy businesses and livelihoods. And individual bans can be useful (Twitter banned Milo Yiannopoulos a few years ago and he and his toxic fan base haven’t resurfaced meaningfully since), but individually banning people is at best a band-aid when your network overall is continuously producing new extremists.

Scale is hard

For years, there has been sizable public pressure on these companies to do more to moderate. Initially companies seemed to hope they could handle this kind of moderation automatically, but in practice that works poorly; algorithms are bad at understanding the full context behind the content of everything posted and can’t make accurate determinations.

So now Facebook’s moderation is powered by an army of humans that must toil away around the clock, slogging through deeply disturbing content and trying to make human decisions at the pace of a machine. It’s a mentally taxing job.

In reality, the viability of tech companies running networks of user-generated content at scale is a myth. Moderation is a nightmare, and these companies are barely even trying to pretend they can keep up with it, and they really only tread water with moderation by subjecting a team of people to terrible working conditions.

But big tech companies want you to ignore that and just continue to let them exist because they don’t want you to even fathom that it’s possible for the world to exist without companies with billions of users.

But that scale, and that scale alone, is the very core of this problem. If we stop having companies with billions of users, we suddenly stop worrying that there are companies that function as pseudo-governments. If we stop having companies with billions of users that are breathlessly handing out recommendations to get those users to join new communities that are trying to overthrow democracy, democracy can be safer.

The question, then, is how we might do that, and what a world that rejects scale might look like.