• Balls Into Bins In Distributed Systems: Balls and bins is a well known problem that is apparent in hash tables and load balancing. (Other places too I’m sure) This gets into the theoretical side quite a bit which I liked. :)
  • Against Incident Severities and in Favor of Incident Types: Severities can create behaviours that make it harder to dig into something weird happening in production. eg Somebody doesn’t want to call a sev1 until they’re absolutely sure something really bad is happening ultimately delaying resolution. Coming up with new buckets addressed this somewhat and also helped them decide who should be poked initially to get involved.
  • AWS Lambda turns 10: A rare look at the doc that started it: A look back at how lambda got its start in Amazon’s product writing process that was the initial pitch. (Amazon will write a press release and FAQ for a service that doesn’t exist as a way to convince people of value.)
  • “Founder Mode” and the Art of Mythmaking: This idea of “Founder mode” has been floating around lately that started with a talk. Charity has some issues with the person who gave it but also acknowledges there are reasonable ideas in here. Not enough humility or ownership of mistakes made at the ceo level certainly.
  • Homelab K3s: Kubernetes in small places: K3s sounds like something fun to play with in a raspberry pi cluster!
  • Benchmarking RSA Key Generation: Interesting read. This level of crypto theory is out of my wheelhouse but fun to learn about all the same.
  • Things we learned about LLMs in 2024: Developments in llm tools in 2024. DeepSeekv3 was a breakthrough model in 2024. On $6mil to train … ! :scream: