alt Hacker NewsI hear you. However, I have one rule of writing: assume the reader is lazy. It is not that they are, but the assumption goes a long way in making content digestible.
Also, I think knowing the combined effect is super interesting. For example, micro-benchmarks are fun to use and see improvements, but I also want to know the effect on the whole program.
I do wonder though if assuming the reader is lazy is the best. Especially in technical posts. I think there is a difficulty in balancing forcing the person to digest what you say and making it approachable (especially when you consider a noisy audience). It is a natural filter, is that good or bad? Guess depends.
Agreed about the microbenchmarks and scale. Things don't always scale as expected. But I think there are a lot of variables here so it might be difficult to portray an accurate expected result. Though I can see this being worthwhile for anyone wanting to build RAGs or process lots of text into some embeddings. Also looks like the project is still under active development and started 6 months ago (single dev?) so I'm not sure we should expect to see too big of scale: https://github.com/bosun-ai/swiftide
So idk, that seems like exactly the kinda thing HN should be well suited for: new projects where people are hacking together useful frameworks. But idk, I guess if YC is funding companies who's business model is to fork an OSS then the bar might be lower than I think. But I thought we were supposed to be hackers (not necessarily crackers) ¯\_(ツ)_/¯