Get email delivery of the Cadence blog featured here
This is the second post about Benedict Evans' annual big presentation about the internet. Part 1 appeared yesterday in my post Benedict Evans 2020: Standing on the Shoulders of Giants.
We connected everyone... including the bad people. Of the 5.5B adults on earth, 5B have a phone, and 4B have a smartphone. By the way, as someone (I forget who) pointed out elsewhere, this is more people than have a toothbrush. The result: 7 out of the 10 largest global companies by market cap are tech:
There are lots of issues, as you can see in the bubbles above. So there must be a simple answer: break them up seems to be the war cry of the day. But you need to look carefully at what problem you are trying to fix:
A lot of antitrust issues center around market share. But what is that? Is Amazon's market share 5% (of share of all retail) or 35% (share of e-commerce). Is Apple's share 15% (global smartphone unit sales), or 75% (share of US teenagers) or 100% (share of iOS app store sales)?
Another issue is what could feasibly be split apart. You could make a good case that YouTube, Instagram, WhatsApp, and AWS would be viable as standalone businesses. Ben reckons that breakups would affect ad competition but not product competition. If you split out Instagram or YouTube, it is unlikely to create YouTube or Instagram competitors.
But actually, few of the issues that make people worry about "big tech" are related to anti-trust. Regulation is boring, complex, and somewhat effective. So Ben reckons tech will become a regulated industry like lots of other industries (cargo shipping, pharma, airlines, cars, utilities, and more). But probably we won't regulate "tech", we will have different regulations for different issues, just as we don't regulate banks, we regulate retail deposits, credit cards, mortgages, options...and many more.
Is this tech problem a new problem, or an old problem expressed in a new way?
A big part of the problem is that we don't know what we want and there are tradeoffs. One big one is whether we want the platforms to know and control what happens on their networks? Or not? Or both?
There's bad stuff on your platform—take it down.
There's bad stuff on your platform—take it down.
But that's easy to say, but social networks are a mix of many publishing forms with different speech and distribution models:
So in those domains, we go from totally public to totally private, and potentially items can go from innocuous, to rude, to offensive, to illegal. What is bad enough to "take down" and what does "take down" mean?
It is the end of the American internet: China and India (together) have five times more smartphones than the US. And about twice as many as the US and Western Europe put together. China and India use more data than the rest of the world combined. 40% of global e-commerce is in China (India is a big laggard here).
One message from TikTok is that in the future, every hot new thing will not be made in the US and their community standards might not be your community standards.
The global user bases for the big US tech companies are mostly not in the US anymore (red is RoW, grey is US). These are percentage shares, not user counts. As a result, there are lots of competing jurisdictions competing, such as US, EU, China, and subsets within such as Germany or California with their own additional rules. Increasingly, they are applying (or trying) to apply domestic laws to the whole global internet:
One big downside is that extra-territorial regulation means companies start complying with the harshest rule that applies, wherever it is. For example, many US newspapers' websites cannot be viewed in Europe because...why bother even understanding GDPR or CCPA.
This all ignores what happens when regulators conflict. It is not hard to imagine a situation in which one jurisdiction insists on strong privacy for some item of data, and another jurisdiction insists that the police must be given access. This makes the extra-territoriality even more extreme. What is bad content in one country may be innocuous in another, and illegal in a third.
But wait...it gets worse. Part of the point of crypto (blockchain) is to remove central control and sovereignty:
The only certainty about this is that regulation is a regressive tax. It is largely a fixed cost that affects new entrants disproportionately. That is why GDPR and CCPA are such own-goals in Europe (home of internet giants...err...right). Something similar happened with Sarbanes-Oxley, which significantly increased the size of a company for which it is feasible to have an IPO in the US.
Software is part of the world, so all the world's problems are now expressed in software (we connected all the people, including the bad people). Tech has become part of the world and so it gets regulated as part of the world. It is now a systemically important part of society. Every important change gets industry-specific regulation: railways, industrial food, ships, cars, databases...and now the internet.
But regulating tech is complex and full of tradeoffs. Law and policy remain the art of the possible. I mentioned Sarbanes-Oxley above. Here's another big financial regulatory reform. This chart shows the number of pages of rules related to the Dodd-Frank banking regulations. Welcome to a regulated industry.
Benedict Evans produces a weekly newsletter, which goes out to 130,000 subscribers (I'm jealous). I highly recommend it. You can sign up at his website. Or you can download the pdf of the presentation.
Sign up for Sunday Brunch, the weekly Breakfast Bytes email.