🚀 The column: Notes on email newsletters (obviously)

Email is a lot older than the web, but the explosion of the web swallowed pretty much every other content form, and then in the early 2000s publishing tools like Typepad and especially Wordpress opened this up, and for a brief moment it felt like everyone online had a ‘blog’.

But what do you write, and how much, and how do you share it and see it and talk about it? RSS never went mainstream, and pretty soon Tumblr, and then Instagram, Twitter and FB, shifted the thresholds and friction for creating content and came with new organising layers on top for seeing and sharing and following. Meanwhile, a lot of blog (‘web log’) posts were just a link, a paragraph or a photo, not long-form writing. And so social, and mobile, mostly swallowed blogs.

But the average FB user feed has 1,500+ items a day - once you’ve followed everyone interesting you’ll never see what they post, and you’ve mixed your friends and your interests, and the algorithm hides what it will. I have close to 300k Twitter followers, but my average tweet is only seen by 10-20% of that. You get lost in the firehose of posts. And at the same time, if you do still want to write that long-form stuff, where do you do it? Medium tried and mostly failed to answer that, and LinkedIn is trying, but now email has come roaring back (or at least it seems like it) - indeed, Substack is disrupting Medium from below.

So now newsletters are a rediscovered form. Some of this is a reinvention of long-form blogging - people actually writing, but using email as a primary distribution model with a URL coming second: I keep seeing links to ‘a great blog post’ that’s actually a Substack issue. Some of course are actual newsletters - curation of what you missed in the firehose, which is nicely meta (message is the medium). Partly, the internet is now old enough for the pendulum to swing back and forth: people try to solve the same problem every ten years (I keep waiting for someone to recreate Delicious), and then that solution, ten years later again, has just recreated the problem in a new way. Pretty soon Google will make an app for reading newsletters.

However, part of that cycle is ‘no-one goes there anymore - it’s too crowded’. New internet distribution models work like slash-and-burn agriculture: OK for year or two and then it's time to move on. Once everyone has a newsletter, your inbox will fill up, and anyone who isn’t firmly entrenched will live or die by the Gmail ‘importance’ algorithm (60% of my subscribers use Gmail, and OH YES, there is an algorithm).

At that point in the cycle we start thinking about the network: that layer on top that social added, and that can maybe pull stuff out of the noise, or at least manage it. Chris Dixon describes some companies as 'come for the tools, stay for the network'. Wordpress was a tool, but Instagram or Tumblr brought both a tool and a network, that connected the creator with the audience, worked out an interest graph and drove discovery and suggestion, and that in turn tied the creators to the tool and made the whole thing self-reinforcing. Medium has tried to do this suggestion model with blogging, and I'd expect Substack to try too: 'see other newsletters you might like!' But then, what if your content isn't suggested? Again, here comes the algorithm - some people will get buried. (People are trying to do all this in podcasting too, another rediscovered form with some shared characteristics to newsletters - I don't have space to dig into that here). 

But. The new-and-interesting part to newsletters today is payment. Marc Andreessen is fond of saying that the web has an unused placeholder for payment (402), which is where Bitcoin is supposed to come in. But a paid newsletter uses very old tech: the change is the psychology, and the perception of value. An email comes to you, and it’s a tangible, almost physical thing that you have and that you can keep (rather like a screenshot), whereas a subscription to a website is nebulous, and you have to remember it, and you realise you never go there. And if you're paying, maybe you make sure it doesn't lost in the firehose. But how many things will you pay for individually? Maybe there should be an aggregator, that bundles multiple subscriptions into one and allocates payment based on attention. And it could do recommendations, and... oops. 

This can go to different places. Imran Amed turned a blog and a newsletter into the very paid-for Business of Fashion, bringing modern online coverage to fashion. Other people, of course, use this as lead-gen for other businesses, or of course this can just be a old-fashioned distribution model for old-fashioned subscription research businesses. Back in the 1990s Fred Hickey used to do a fax newsletter, as did a young analyst called Bill Gurley, who seems to have done OK for himself. I took this newsletter to a16z, and now I’ve taken it home. So let’s see. 

🧠 My posts

I published an updated version of my annual trends deck, thinking about what COVID means for tech this summer. Link

🗞 News

Facebook content moderation #∞. FB's own internal 'civil rights audit'  board says FB isn't doing enough to take down hate speech of all kinds (though it has made some progress). There are real philosophical problems here - how does a tool used by over a billion people decide what they can say to each other, and how does that spread and amplified by automated suggestion tools? No-one doubts it should do something, and very few people would like it to (say) monitor all of our private WhatsApp chats, but where are the lines in the middle? I have some opinions on this and there are many others, but what's not in dispute is that Facebook has totally failed to articulate a coherent position of its own. Part of the problem is that a corporation lacks legitimacy to make such decisions, and FB has reacted by creating external advisory groups, like this one, but if you do that, then don't you have to do what they say? Link (PDF)

More US tech antitrust: the US legislature's antitrust committee is having a hearing later this month, with the CEOs of Google, Apple, Facebook and Amazon all summoned. Expect a lot of canned talking points (down a shot of espresso every time someone says 'innovation'), but this reflects the ongoing impetus for... something. It is not at all clear why breaking up these companies would solve any of the problems people care about, but one should also ask if such action is plausible, given on one hand the state of current US politics and on the other the fact that these are national champions in the face of a growing perception of a strategic challenge from China. But that doesn't mean that, say, Google's ad-tech stack or Amazon marketplace won't get kneecapped at some point. Link

HK security law versus global tech. The new Hong Kong security laws effectively give Chinese state security agencies unlimited powers in HK, and one aspect of that is that they can demand access to user data from any tech platform operating in HK, covering both users in HK and users anywhere else. This is backed by both fines and the scope to imprison local staff. So what does Google do if Chinese police arrest their staff in HK and demand Gmail data for activists in Germany? One answer, of course, is end-to-end encryption, but that doesn't work for everything. So, companies are wondering if they have to leave - which China of course would welcome. Link

US banning TikTok? The US secretary of state says he's considering banning TikTok and a bunch of other Chinese social apps, following India last week. Not clear if he has the legal authority to do this, and I rather doubt TikTok is really the main vector for stealing US secrets (remember the OPM 'hack'? The US handed over admin access to some of its most sensitive data to... Chinese IT contractors). But the macro theme isn't going away. On one hand, the view that the Chinese state is now a generational strategic adversary is rapidly going from 'China hawks/cranks' to mainstream consensus on both the US and Europe, and on the other, the ongoing global diffusion of tech means that the next cool new social app isn't necessarily made in California anymore. And a deeper problem: define 'Chinese'? Tiktok's parent company, Bytedance, has tried to ring-fence it as an 'outside-China' subsidiary, separate from Douyin inside China, but it's clearly still seen as Chinese. Equally, Zoom is based in Silicon Valley, but the founder is a Chinese national, and most of the engineering is in China, subject to direct Chinese state coercion. So what's the threshold? And can you really just stop Americans (or Germans, or Koreans) liking and using something made in China? Link

Amazon bans/unbans TikTok. Amazon sent an email to all employees saying they had to remove the TikTok app from any phones with access to Amazon email for security reasons -  and then retracted it an hour later, saying this was a mistake. Um. OK.  Link 

Apple shifting supply away from China? Reuters has some smoke signals that Foxconn (Taiwanese, but makes iPhones for Apple on the Chinese mainland) is investing $1bn expanding iPhone capacity in India. Apple has already been building capacity in India due to local import taxes (and has tiny share there) but must also be thinking about geopolitical exposure to Chinese manufacturing, though for obvious reasons it would have to do this under the radar. For context, it's worth understanding that Apple's final assembly and manufacturing excellence mostly happens in contract manufacturing in the Shenzhen ecosystem, but most of the high-value components (screens, chips, image sensors etc) are made outside China. Link

Palantir has filed for IPO. Palantir was founded way back in 2003 to help governments, the military and law enforcement use modern tools to analyse their data: the client base and the fact that Peter 'Voldemort' Thiel was one of the founders make it rather controversial politically, and since it's remained private until now (at a valuation of as much as $20bn, apparently) no-one can look under the hood to see what's really going on. The other reason to look under the hood is that a lot of people in tech tend to presume that Palantir is really just a big outsourcing shop, charging out 'data scientists' by the hour rather than being a 'real' software company (remember the joke that a data scientist is just a statistician who lives in Silicon Valley), and that the magical seeing eye is just a bunch of Excel files and IF statements. 🤷🏻‍♂️ Link

🔮 Reading

Useful profile of ASML and the very very expensive, high-tech machines involved in the chip industry, which China does not and cannot build, but badly wants to. Link

Study of images and misinformation in WhatsApp groups in India. Link

Grocery stores are accelerating automation and self-checkout as a result of Covid. Link

A group of right-wing US media outlets, as well as some more serious titles, were tricked into running comment pieces written by people who didn't exist, and were in fact invented by a middle-eastern propaganda campaign. Link

People are watching Quibi, but in tech and media the main entertainment value is the behind-the-scenes drama. Link

Ten years of the Times digital paywall. A big gamble, many people said it was insane, and it was a huge success. Link

The Harvard Business School 'online classroom' looks like a parody to me. Link

😮 Interesting things

Great footage of the Caproni Ca.60, a 9-wing (yes, really), 8-engine seaplane intended to fly 100 passengers across the Atlantic in 1921. You could think of metaphors, or just admire the insanity. Link

How to use a rotary telephone dial as a USB keyboard. Link

Statistical analysis: football commentary uses different terms to praise players from different ethic backgrounds. Link

Six months ago today: the first BBC report on a 'new virus' in central China. All under control, claim authorities. Link

📊 Stats

The SCMP has a new report on the Chinese internet (free and longer paid versions available). Link