So Confluent pulled a "ship of Theseus" on Confluence Cloud, it still looks like Kafka, but everything underneath has been replaced by a proprietary system.

Please, look into your use cases and keep asking "why", why do I need Kafka? Kafka is a beast of a tool. Example, I have seen people using Kafka after an MQTT broker, because... Kafka has an MQTT connector. Well, MQTT is a pub/sub architecture, so you are putting a pub/sub (Kafka) after another pub/sub MQTT, you need to be aware of that and need to think it thru if it makes sense for your use cases.

Given their description on separated scaling on storage and compute, I was wondering if Kora has an architecture more similar to Apache Pulsar than Apache Kafka.

AWS Managed Kafka is eating their business for lunch and this is the attempt of layering on the non-open source options into only their cloud.

I wonder which language they chose for this project? I couldn't find any reference to that in that blog post.

I wondered the same, with so much focus on performance across the blog post.