Glauber Costa on Technical Blogging
Writing helps you reach the people who want your solution – and that's often harder than solving the technical challenge
Welcome to our latest attempt to (not-so) gently nudge you to write more! Following up on writethat.blog and Writing for Developers: Blogs That Get Read, we’re sharing the perspectives of expert tech bloggers: why they write, how they tackle writing challenges, and their lessons learned. This time, let’s hear from Glauber Costa.
Glauber is the founder and CEO of Turso, which is rewriting SQLite in Rust. A recovered Linux kernel hacker, he also spent quite a few years at ScyllaDB – the birthplace of his technical blogging duty under capital punishment opportunity.
Over to Glauber…
Why did you start blogging – and why do you continue?
I started blogging years ago at ScyllaDB. I was initially forced to do it, but I ended up really enjoying it. [This is strikingly similar to Sarna’s “Stockholm Syndrome” story in Chapter 1 of the book].
I’ve always liked teaching people and I saw that technical blogging was a way to do that…at scale. As I was learning new things, often working with previously unexplored technologies and challenges, blogging gave me this opportunity to teach a large audience of people about what I discovered.
I kept doing it because it actually works. It really does reach a lot of people. And it's very rewarding when you find that your blog is getting people to think differently, maybe even do something differently.
What has been the most surprising impact of blogging?
It’s been quite surprising to see how well a blog post can reach people who might be interested in using your product. Don’t assume that you need a fancy marketing website. Especially if your audience is developers, what people really want is the pure technical opinions and information. Just capturing that in blogs generates a very large and healthy “inbound funnel” of interest for your product.
Another surprising thing that’s related to blogging: some of the most passionate commentary on the orange site comes from people who don’t even read the article. I know how Hacker News is, but this still never ceases to amaze me. I mean, if you’re going to take the time to share some heated opinions, I would at least expect that you would first read beyond the title of the article that you’re criticizing.
What blog post are you most proud of and why?
The one on Deterministic Simulation Testing [A deep look into our new massive multitenant architecture]. I wrote that blog specifically with the hope that ThePrimeagen would read it on his livestream. Just like some applications are built to be cloud native, that blog was designed to be “Primeagen native.” He did end up reading it, and it worked fantastically in that context.
I took some really big risks by targeting that one for ThePrimeagen’s stream. For example, I inserted some intentionally outrageous phrases (like “a circle jerk of bug-free glory”). That could have been a disaster if he hadn’t read it. But since he did feature it – and that enabled the message to reach lots of people – the risks were worth it.
What post was the most difficult to write and how did you tackle it?
I’d say the most challenging blogs to write were the benchmark blogs at ScyllaDB. Blogging can be challenging and benchmarking is always challenging – so you’re taking on two challenging things at once. Plus, you know that readers will be highly suspicious of any benchmarks (as mentioned in the benchmarking chapter of Writing for Developers, benchmark authors are “guilty until proven innocent”).
Doing a good benchmark requires a lot of knowledge, a lot of preparation. Sometimes it would take a whole month to write a good blog sharing the results of a benchmark. Even when you’re being very careful not to deceive people, you know that people will still assume that you’re lying or rigging the results somehow – so the amount of preparation and review that you have to do is tremendous.
Any lessons learned that you want to share with the community
Even to this day, I am 100% unable to predict which blogs will do well and which blogs will not. Don’t spend too much time trying to figure out whether something is going to work. You will never know. If you have an idea, just write it!
For instance, take the post I wrote on io_uring in 2020. I was actually very reluctant to write that one. ScyllaDB’s CEO [Dor Laor] really pressured me to get some results and share them in a blog, but my initial response was “Why? Nobody really cares about this.” It turns out, that was my best performing article ever. It was a Hacker News hit and it still gets a steady stream of traffic.
Honestly, I see no correlation between what I think will do well and what actually does well – and vice versa. So just write… and some of them will eventually do well.
Your advice for people just getting started with blogging?
Number one, just do it! But also, keep it short and to the point, focusing on just one thing.
Sometimes your team might try to pressure you into including all sorts of interesting details that are related to that topic. I never liked this approach, and my blogs became a lot more successful when I stopped trying to squeeze in details that weren’t critical to what I was trying to convey.
It’s natural for you to want to tell the whole story, but having a clear focus is important for keeping the reader interested. If you dump everything into the blog post, readers won’t discover the highlights. Nobody cares about you. Really! People have very little time. Just cover the key points, and know that people will seek you out – maybe engage on social media – if they want more details.
So, keep it short, keep it simple, keep it to the point. Find one thing that’s going to be memorable, and focus the article on that. The rest doesn't matter. Resist the temptation to write everything that's on your mind.
A few blogs that you particularly enjoy?
Anything from TigerBeetle.
Anything else you want to add?
It doesn't matter how much you write – you could always write more. And you should write more. It truly helps, especially when you're starting a company. Finding the people who want your solution is a much bigger problem than the technical problem. Writing has a disproportionate impact on your ability to reach those people.
***
Stay tuned for more tech blogger spotlights mixed in with the monthly writethat.blog updates.