I have mixed feelings about blogging.

This tweet by Patrick McKenzie is relevant:

This kicked off a bunch of replies. Continuing the conversation, he said to the author of Stratechery:

Just as I have mixed feelings about blogging, I have mixed feelings about these tweets as well. I had a blog which was, for a long time, a very powerful and positive influence on my overall career. But I shut it down recently, and I’m ambivalent about this new blog (the one you’re reading right now).

It comes down to two problems: strategy and voice.


My blog had a wild and sometimes combative voice. I’ve changed somewhat as a person, and my old blog was always at least half about marketing. Between these two factors, I don’t feel my old blog represents me well. In fact, I’m sufficiently eager to distance myself from my old blog that I’ve considered blogging under my middle name as Dave Bowkett instead, and claiming to be Giles Bowkett’s evil twin. This could have a nice side effect: I might have to explain my name’s correct pronunciation less often.

It would also be fun to claim that a mad scientist created me, and my evil twin, in a lab. (It would almost be true, with the caveat that all three characters in the story were the same person.) More soberly, as an experiment, I created a small blog with a very high signal/noise ratio about only one topic of personal interest — algorithmic music in a very obscure programming language — and deliberately kept it under the radar.

Part of my old blog’s voice came from embracing criticism and playing with it. I had a weird experience over the years as a blogger where so many people accused me of deliberately trolling them that I started to do it, partly to understand their reaction, partly to see if they were right, and partly because it turned out to be fun and profitable. But I ended up disliking that approach quite a bit, and for the record, I initially started out with zero intentions of trolling anyone.

I still don’t 100% understand where the boundary is between genuinely provocative thinking and deliberate trolling. I think that boundary is important, but difficult to pin down. But I won’t deny that I crossed it from time to time.


The blog that I recently shut down began in 2006. It ran on Blogspot, it looked terrible on mobile devices, and it had deeper strategic problems.

I built it around Google. I assumed that I should write about absolutely everything, and allow users and their search results to shape what succeeded or failed. This was also part of what made the voice wild; since I bought into the idea that you could throw absolutely anything out there and trust the Internet to sort it by usefulness, I would do exactly that. I did posts that were entirely images, numerous diatribes of varying coherence, and random hack after random hack. So many random hacks.

Here’s me controlling a CSS3 animation with a hardware synthesizer. This ran on web sockets and MIDI, and was very unnecessary.

Here’s me building a Tachikoma in Three.js. I also made the music in this video.

(I think that the music I make these days is better, but you’ll have to take my word for it. In keeping with the general theme in this blog of retreating from publication, none of my recent music is available online as of this writing.)

Anyway, it wasn’t all just random hacks. For SEO, I frequently would just use an error message as the title for a blog post about the solution for that error message, which was a much more useful tactic before Stack Overflow came into being.

Modern blogs use a different set of strategies. Google is less important, although still in the mix; Stack Overflow has at least reduced the value of the error message approach, and maybe eliminated it entirely; and social media plays an important role in the overall strategy of a modern blog. My blog started before Twitter even existed, and although I adapted to accomodate social media, my adaptations were modifications, not strategic overhauls.

Another major change is that modern blogs are often just funnels into newsletters, whether free or paid. Because my strategy centered around Google, I didn’t have a newsletter, and I didn’t see any way to add one without completely reconceptualizing the whole project. So I waited until I was ready to do that. And that time is either now, or soon.

I think the reconceptualization will incorporate these ideas:

  • Good writing is inherently valuable.
  • Certain topics are also inherently valuable.
  • Writing on the Internet means both constructing a web of content, and partitioning that web into multiple streams.
  • Public webs are the default; private webs are rare for individuals but common for companies.
  • Private webs are probably not as common for companies as they should be.
  • Public and private streams are, in each case, common and useful.
  • Streams of content can produce streams of income.
  • Partitioning streams is an art, not a science.
  • Chronological streams are the default, but topical streams deliver the most value.
  • Email sales sequences, intro sequences, and mini-courses are popular and effective streams.
  • Personal branding is another default, but a risky one.
  • Even a fully sincere attempt to present yourself accurately may result in playing a character on the Internet, both due to subjectivity and because your content will move through mechanisms you do not control.
  • Systems such as Google, Twitter, Facebook, and others will partition your content into sub-webs and sub-streams.
  • This has upsides and downsides. Twitter decontextualizes, Google increases reach, etc.

These aspects of my new perspective are clear to me. Beyond this, I have some less precise observations. I love Dan Abramov’s deep dives into React topics. I like Stratechery, Above Avalon, and Trapital. I like the 30x500 system from Amy Hoy and Alex Hillman. I find that technical topics desperately need timestamps, but it’s self-defeating, and a disservice, to timestamp essays on strategy. (Although that’s truer for some strategies than others, as this discussion demonstrates.)

What’s Next?

There’s enough uncertainty remaining that I’m still a little paralyzed by it, but that paralysis applies only to publication. When I stopped blogging, I turned my attention to other forms of writing. I have some detailed outlines for fantasy novels, for example, along with the characters and speech patterns to match, and I’m still in the process of deciding what to do with them. If nothing else, they were great exercises.

As far as blogging goes, I still have a ton of open questions. Here’s a list of what I’m thinking about, when I have the time. I think they’re a decent approximation of The Big Questions™ of modern blogging.

  • Balancing public vs private, and web vs stream
  • Partitioning streams
  • Timely content vs timeless (“evergreen”) content
  • Supporting online content with public speaking, but not to an exhausting degree
  • When somebody finds me on a suboptimal stream and/or web — for instance, a web site or an app that somebody else controls, and which consequently makes me less money than I’d get from a similar system which I controlled — how to transition them to a stream and/or web that I want them to consume instead

A note about that last point: this obviously covers newsletter signups, but I want a strategy that is general enough to withstand new developments more gracefully than my old strategy did. Also, I’ve seen plenty of other content creators grapple with this, so I think it’s an essential question for content online. For example, there’s a D&D “show” called Dimension 20 which puts some content on YouTube, but seeks to pull its YouTube audience away to a subscription site called Dropout.tv instead. That’s another instance of this general question.

I’m also wondering about a couple more personal concerns:

  • What to do with the old blog content I still like
  • How to filter the massive number of posts from my old blog for stuff I would want to put online again (all of it is offline now)

That last question also applies to my Twitter account. As a stopgap solution, I’ve protected the whole account, but that’s a blunt instrument, and it has obvious downsides.

There’s a lot of questions remaining for me, but I’ve always found it useful to start before I was ready, and that’s what’s going on with this post.