totally makes sense! that is a really awesome idea

ahhh that's really neat, so it operates on a smaller local window! I'm assuming that does hit issues with sparse query results or a specific ordering, or are there cool things you do for that too?

I do love getting rid of the run button - but we've found it's a bit harder on shared clusters and expensive/large queries. Is there any magic there or just "we have enough compute to do it" kind of thing?

I stumbled upon the first post a while ago and was equally puzzled by AWS's lackluster disk performance. I hope the whales in the AWS ecosystem push them to take it seriously as the current performance is atrocious. We already have open-source DBs & use cases that can easily saturate their NVMes.

I've been playing around with the idea of adding density/utility to a chart on hover (as the user is likely trying to drill deep into the data) - but rendering a "cleaner" state otherwise. We shipped some of this recently in our v2 and we'll see how it's received.

A common path to #2 is just push all your work onto S3 and just do some caching on top!