My essays

I wrote about the regulation of technology. We regulate lots of industries, from food to banking to airlines, and now, increasingly, we’re going to regulate tech. But that means global platforms collide with local regulators, with complexity, trade-offs and mutually incompatible demands. Tech is going to become a regulated industry, and this will be expensive. Link

News

Mega Ant IPO. Ant Financial runs Alipay, the universal and ubiquitous mobile payment system that most of China runs on, with over 800m users. China leapfrogged conventional payment cards and went straight to this. It's a semi-spinout of Alibaba, and now it's filed for IPO in its own right, with a valuation of perhaps $150bn. Of course, $150bn isn't very big in an age of $1tr tech companies, but there will be a lot of interesting data on just how this is used, and another wave of attempts to copy it in the west. Link

Slack versus Microsoft. Slack filed an EU competition complaint against Microsoft, saying MSFT shouldn't be allowed to bundle Teams for free with Office 365 subscriptions. History repeating itself? Well, yes and no. 20 years ago Microsoft was accused of blocking OEMs from including Netscape at all, and forcing them to preload Internet Explorer on new Windows PCs. There's no compulsion here: this is bundling, not tying. But should Microsoft be allowed to make Teams free as part of a bigger bundle at all? Well, it's pretty obvious that that communication and sharing are part of work and productivity, unless you think that feature should just be integrated into Windows (and I doubt Slack would find that an appealing argument). And adding features that already exist as separate companies is pretty much what software does - this is my favourite example. This is hard to build a headline case around. Link (Note: my former employer a16z is an investor in Slack)

Intel falls behind. Intel lost mobile totally, but it was always in the lead on raw semiconductor tech. The key benchmark for that is how fast it moves its fabrication processes to steadily small sizes: the current generational change is the move down to 7 nanometers, and Intel just announced it will now be at least a year late, getting to the market with 7nm in 2023 at the point that Taiwan's TSMC plans to have moved on to 3nm. (Intel both designs and makes chips, whereas TSMA mostly makes other people's designs for them - hence it makes Apple's custom ARM-based CPUs for iPhones and now Macs, for example). This is good news for Apple, since Macs with its new custom chips will compete with slower PCs, it's annoying for PC makers, and it probably doesn't affect you directly (unless you care a lot about high-end PCs). But this does feel like the end of an era, especially in the context of China's big and much-debated strategic push to gain its own semiconductor capabilities, and with Taiwan now in a very clear lead in global chip-making, this has all sorts of interesting geopolitical implications. Link

Drip-drip of the Facebook ad boycott. Disney is the latest big brand to jump, but remember that something like three quarters of FB advertising is SMEs. Link

Google's use of Android telemetry. Apparently, Google analyses which Android apps people use, with a view to analysing competition from, say, Instagram or Tiktok. The privileges of owning the platform. Link ($)

Amazon's corporate venture fund poaching ideas? The WSJ has a story that Amazon has a pattern of meeting companies as a potential investor and then launching competing products. I'm not entirely sure they make it stick. The devil is in the detail here: the opening example in the article talks about a product launch that was four years (!) after the meeting, which is hard to see as bad faith. But some of these anecdotes do look bad - confidentiality is the foundational ethic of venture capital, but does the same apply to corporate VC? Link ($)

Amazon is American's biggest advertiser. Ad Age says Amazon spent $7bn in 2019, up 37% year on year. (It also generated over $10bn in ad sales, selling placement in search results, which is a neat arbitrage - buy traffic cheap and sell it high). Bezos once said advertising was a tax on bad products, but maybe it's become another aspect of his famous virtuous cycle diagram, and another way to leverage Amazon's negative cash conversion cycle. Link (PDF)

Garmin taken down by ransomware? Garmin (basic smartwatches, satellite navigation) has had its platforms taken down for several days by a security breach - apparently, someone is demanding a ransom to let them get up and running. Link

Can bankers make Tiktok less Chinese? Apparently, Tiktok's Chinese parent company Bytedance, as well as some of its foreign shareholders, are considering some kind of unbundling transaction that allow them to claim that it's no longer a 'Chinese' company, to ward off threats of a ban from the USA, India's current ban and potentially others. I'm sure if you get enough M&A bankers to pull enough all-nighters they can come up with a structure that looks good on paper, but can you really move the product and engineering, and even if you can would anyone believe you? And are the calls for bans motivated by legal structure anyway? They have to try, but I'm skeptical this is a solution. Link

More on Twitter's security SNAFU. Apparently, the tools that hackers used last week to compromise a bunch of high-profile accounts are accessible to over a thousand people inside Twitter. I am not a security analyst, but if true, that would be a very long way from best practice. But then, nothing would be surprising about Twitter. Either way, make a note never to say anything on DM that you want to keep private. Link

Apple commissioned another lobbying report on app stores. This one points out that Apple's 30% commission is pretty standard, which is mostly true, but a little circular given it played a role in setting 30% as the de facto standard in the first place. This is of course aimed at anti-trust regulators, both in the USA and the EU, but it has little relevance to anything they'll consider. 'Other platform monopolists do this too' isn't a winning argument. Link

Twitter has banned QAnon, a bizarre online conspiracy theory group. Unfortunately, it's not really clear what that means, how it's defined and how Twitter will avoid playing whack-a-mole with sort-of-similar posts and accounts. It's easy to say 'get rid of the bad stuff' and much harder to do it. Link. 

Reading

There's a steady flow of current and former FB employees thinking and worrying about the mechanics of the site. We connected everyone, and that includes all the bad people, but more importantly all of our own worst instincts, and how does that work? Link

Fascinating story about online retail in Bangladesh, which runs mostly in private Facebook posts (I remember seeing Instagram camel sales in the Gulf a few years ago). Link

A useful Stanford white paper on China's 'information operations' (= propaganda). China does this deliberately, over decades, at scale, and we should be aware of that. Link

Interesting video discussion on the state of of media and media businesses, with the outgoing CEO of the NY Times, Mark Thompson, and the CEO of Axel Springer, Mathias Döpfner. Link (Youtube, one hour)

Western luxury goods firms are starting to try live streaming, a hot trend in China a few years ago. Link

Interesting WSJ story on developers looking at building a 'virtual' Hajj experience. I always thought the point of a pilgrimage was the journey, and the shared experience, but not this year. Link

This list of important bitcoin essays is also a useful index of the hype cycle, if you look at the dates and names. Who sees things first, and when does the thinking proliferate? Link

Interesting things

If you do video calls on a Mac, check out Camo - it lets you use the great camera in your iPhone in any video call app instead of the not-great camera in the Mac. Link

General Magic is a great documentary about the company that invented (or tried to invent) smartphones in the early 1990s. So many lessons, and lots of significant figures with embarrassing haircuts. Link

Stats

The BBC now has weekly reach of 470m people. (Most of them aren't paying a licence fee, though). Link

The column: learning from the Chinese internet

Twenty years ago, at the top of the internet bubble and the mobile bubble, people were very excited about the idea of the ‘mobile internet’. We only had a few prototype WAP handsets, but in Japan the mobile internet was really, really working. By the end of 2000 over 25m people had subscribed to mobile web services provided by the cellular networks, of which DoCoMo’s i-mode was the best known. And the phones had packet-switched data, built-in frictionless payments, colour screens and often even a camera. Japanese phones were amazing. I-mode was amazing. This was clearly The Future. 

The problem was, you couldn’t read the language, so you couldn’t read any of the press coverage or even the product marketing; it was a 12 hour flight away; and as a foreigner you couldn’t really try the products even if you did go there. So there was a huge amount of garbled information floating around and a big market for 100 page PDF presentations explaining it all. And yet in the end, it didn’t matter. The future wasn’t a cut-down mobile web on proprietary feature-phones controlled by mobile operators. I-mode told us that data services on phones could work, but not how - it was the Minitel of Japan. 

Every time I look at the Chinese internet, or hear about WeChat, I remember i-mode. Here, again, is a market where all this amazing stuff is happening, and where they sometimes seem to be living in The Future, but most of us can’t try the products or read the language (and Google Translate isn’t a substitute) and again there’s lots of garbled information floating around. So, is this telling us the future? Is the amazing Chinese company, product, business model or consumer behaviour you're just heard about third hand about to come and eat Facebook's lunch?

Well, perhaps. Some of what happens in China just reflects a different market structure and history. WeChat took dominance in China at the right time and the right place with the right model, and none of that was true when it tried to go to India and failed totally. A model doesn’t necessarily work because it’s the ‘right’ model, but because of how it got there and what else was going on. Indeed, sometimes looking at China tells us what the 'obviously doomed to failure' Yahoo or Myspace could have been with different founders and different starting positions. 

Equally, the barriers to entering China work both ways - it can be just as hard to leave. The market is so big and so specific that you can over-fit, and struggle to succeed outside in a quite different ecosystem. Bytedance did succeed, partly by making TikTok a different thing, but WeChat failed totally, partly because India (etc) wasn’t China but also partly because the home market always took priority. 

However - China has leapfrogged a lot of developed-world infrastructure, going direct to mobile and digital, most obviously for finance (see Ant Financial below) and retail without any legacy assumptions. And, it has 8-900m consumers with smartphones and a vast frenzy of creation, copying and innovation on top of that. So, it's not that this necessarily shows us the future, or that this or that big Chinese consumer internet company will go global, but that we should expect the flow of ideas to go both ways. WeChat may yet be the i-mode of China, but China is another huge pool of creation and innovation.