mbrt blog

I wrestled current1 LLMs into behaving like my childhood AI hero: here’s what worked, what didn’t, and why. Uno is my favorite character from my favorite childhood comic PKNA. He’s the friendly and sarcastic AI helping Donald Duck in adventures spanning 56 issues in the mid ’90s. My first attempt at re-creating him was a complicated Excel spreadsheet in my early teens with lots of nested IF functions. This is my second attempt. ...

I processed a large fraction of the Internet on a single machine, trying to find some valuable but forgotten domain names to buy (probably 25 years too late). Here’s the story of what I found and how I did it, and not the story of why I decided to try, because I don’t know myself. Spoiler alert: If you just want to know whether there is indeed something, and you should scramble to scoop it up to become a domain-flipper billionaire, hold your horses: I believe there isn’t much. The following sections are more focused on the engineering side, so if you only care about the conclusion, skip to Results. ...

I was frustrated by the gap between stateless and stateful applications in the cloud. While I could easily spin up a stateless application as a “serverless” function in any major cloud provider and pretty much forget about it, persisting data between requests was a game of pick two among three: cheap, strongly consistent, portable. Could I solve portability and lack of transactions myself with a single client-side solution? I thought it would be possible through object storage (e.g. AWS S3), which is strongly consistent, ubiquitous and cheap. ...

What is a design document? When is it useful to have one? How to make it useful? These are the questions I’m going to address in this post. The goal is to provide some bite sized, easy to remember guidelines. My promise to you is that after this post, you’ll be faster and more effective in making and reviewing designs. This is a written version of a talk I gave at my current employer. Since it was well received and helped a few folks, I decided to make it more available. ...

Running a Site Reliability Engineering (SRE) organization correctly is difficult and expensive. Spare the frustration, perhaps what you need is Sysadmins. Since the term was coined by Ben Treynor in 2003 at Google, lots of ink was spent on praising SRE practices. Not enough on when it is appropriate to have SREs. This post is a take on that angle. Disclaimer: I was an SRE at Google and this piece represents only my own views. ...