I meant the time it takes for servers to answer your requests.

Yeah, the more people you follow, the faster server responses become. We find it incentivises community interaction.

Kidding. But seriously, I'm on the dev team, and we know load times are a problem :(

The root problem is that Hubski is a fork of an old version of Hacker News. HN is Arc Lisp, and makes a lot of poor decisions, like iterating over potentially huge lists. Also, all data was serialized as lists directly to disk. We're moving to SQL. But SQL is slower than loading a dumb binary blob, so naïve data conversions make it even slower. The solution is to fix pages to only load the data they actually need, and use SQL to filter. Unfortunately, it's hard to do all at once, so as we move data into SQL it gets slower, and then as we fix code to only load what's necessary, and use SQL to filter, it gets faster again. Eventually, when all the code is converted to Racket, and all the data to SQL, it'll be fast. But until then, we're stuck with what we have.

Another problem is that we all have day jobs. I feel like I could fix, like, all the problems, if I just had a month or two to work on Hubski 40 hours/week. Such is life.