Author
Name
Thomas Ptacek
@tqbf
@tqbf
Image by Annie Ruygt
A boldfaced lede like this was a sure sign you were reading a carefully choreographed EffortPost from our team at Fly.io. We’re going to do less of those. Or the same amount but more of a different kind of post. Either way: launch an app on Fly.io today!
Over the last 5 years, we’ve done pretty well for ourselves writing content for Hacker News. And that’s mostly been good for us. We don’t do conventional marketing, we don’t have a sales team, the rest of social media is atomized over 5 different sites. Writing pieces that HN takes seriously has been our primary outreach tool.
There’s a recipe (probably several, but I know this one works) for charting a post on HN:
- Write an EffortPost, which is to say a dense technical piece over 2000 words long; within that rubric there’s a bunch of things that are catnip to HN, including runnable code, research surveys, and explainers. (There are also cat-repellants you learn to steer clear of.)
- Assiduously avoid promotion. You have to write for the audience. We get away with sporadically including a call-to-action block in our posts, but otherwise: the post should make sense even if an unrelated company posted it after you went out of business.
- Pick topics HN is interested in (it helps if all your topics are au courant for HN, and we’ve been very lucky in that regard).
- Like 5-6 more style-guide things that help incrementally. Probably 3 different teams writing for HN will have 3 different style guides with only like ½ overlap. Ours, for instances, instructs writers to swear.
I like this kind of writing. It’s not even a chore. But it’s become an impediment for us, for a couple reasons: the team serializes behind an “editorial” function here, which keeps us from publishing everything we want; worse, caring so much about our track record leaves us noodling on posts interminably (the poor Tigrises have been waiting for months for me to publish the piece I wrote about them and FoundationDB; take heart, this post today means that one is coming soon).
But worst of all, I worried incessantly about us wearing out our welcome. To my mind, we’d have 1, maybe 2 bites at the HN apple in a given month, and we needed to make them count.
That was dumb. I am dumb about a lot of things! I came around to understanding this after Kurt demanded I publish my blog post about BFAAS (Bash Functions As A Service), 500 lines of Go code that had generated 4500 words in my draft. It was only after I made the decision to stop gatekeeping this blog that I realized Simon Willison has been disproving my “wearing out the welcome” theory, day in and day out, for years. He just writes stuff about LLMs when it interests him. I mean, it helps that he’s a better writer than we are. But he’s not wasting time choreographing things.
Back in like 2009, we had a blog at another company I was at. That blog drove a lot of business for us (and, on three occasions, almost killed me). It was not in the least bit optimized for HN. I like pretending to be a magazine feature writer, but I miss writing dashed-off pieces every day and clearing space for other people on the team to write as well.
So this is all just a heads up: we’re trying something new. This is a very long and self-indulgent way to say “we’re going to write a normal blog like it’s 2008”, but that’s how broken my brain is after years of having my primary dopaminergic rewards come from how long Fly.io blog posts stay on the front page: I have to disclaim blogging before we start doing it, lest I fail to meet expectations.
Like I said. I’m real dumb. But: looking forward to getting a lot more stuff out on the web for people to read this year!
Next post ↑
The Exit Interview: JP Phillips
Previous post ↓
Did Semgrep Just Get A Lot More Interesting?
Instagram Engineering
Shopify Engineering