r1b

Neat, curious how “multitenancy” impacts these calculations - thinking about big DWH players like BigQuery, Redshift etc - or even stuff like motherduck.

in this situation, is it just like, you should front kinesis with a service that can apply appropriate quotas / limits?

re: the kinesis key - curious, what is the right way to configure log delivery for remotely deployed appliances?

Nodding along furiously. One area where this comes up is when you want to export some piece of logic that effectively “returns a table”.

For example, I work with a source of data that requires a particularly hairy JOIN strategy. The logic is the same for any table with the same “shape”, but no one can reuse it without copy / paste.

> Discoverability is especially hard because 99% of the music people create sucks

This - as a listener, quality is the hard problem. It is encouraging that the proposal affirms the value of curative functions (like labels).

As an artist, I actually don’t really care about music’s commercial problems - I’m more annoyed by the constraints on musical art objects inherent in all music platforms.

Like, experiencing art objects in a gallery hits different vs scrolling through bandcamp. The internet is, already, the gallery but it’s like we replaced all of the paintings with tiny prints, eclipsed by the placards.

The thing I would really love is a music platform that feels like a hosting platform, not a marketplace. Where a user can simultaneously act as a listener, an artist, a curator or a critic.

This was my first thought as well. I like the language, but I’m hesitant to adopt it without tooling that can extract an AST, ideally interoperable with the hard-won tooling we already have for SQL.

Improved UX at query development time is nice, but data teams need so much more than this. Stuff like lineage, linting, cross-dialect interop - these are no longer niche concerns, they are table stakes, and they come from having a shared syntax for data transformation.

It’s strange to me that the interop story here only considers compatibility with GoogleSQL.

These days, we have tooling that can generate an IR for SQL and transpile it to any dialect (see: sqlglot).

Doesn’t coupling SQL syntax to dialect-specific SQL semantics throw a wrench in this?

WFMU is indeed the one I’m referring to here :)

To get the freeform format with newer, “shinier”, and often more electronic sounding music, I also like NTS.

This mostly an assessment at scale. By volume, irrespective of genre, and even accounting for subjectivity, most of the material that gets published isn’t very good.

That being said, I generally don’t agree with conclusions like “culture is frozen”. You are correct that there is _more_ high quality material available than ever before. The challenge is just that it’s harder than ever to find it.

Outlier here (musician, spend hours per week trying to find new music) - some thoughts:

- The search space for music is really large and noisy. Most of the stuff out there isn’t very good, and the stuff that is good isn’t always discoverable with a single strategy - The best strategies almost always exploit human connections

Some strategies I use:

- Spatial locality, who is performing with or near artists that I like? - Publishing locality, who is on the same label as an artist that I like? - Artist locality, what other projects has an artist I like contributed to? - Fan locality, what other artists does a fan of an artist that I like enjoy?

——

Note that none of these strategies are as effective as “relinquish control”. For example, there is a freeform radio station near me that I listen to all day at work. I have a rule that I won’t turn the radio off in the middle of a DJs set, even if I don’t like a song. This has helped me “break through” to interesting artists I wouldn’t have discovered otherwise.

To the article’s question, I think the main factor here doesn’t have much to do with music. Cultural production has exploded, and it’s really hard to navigate any cultural space in a non-obsessive way.

I thought it was interesting that the effect of “generational preference for music released when teenaged” seemed to wane around Gen Z. I wonder if this is just exhaustion, perhaps with tendencies towards pastiche as a consequence.

> a single HIFI Enterprise account can easily have half a gigabyte of associated royalty data representing over 25 million rows of streaming and other transaction data

Curious about why they were struggling with this workload on BigQuery. It just doesn’t seem like very much data. Maybe they were cost-constrained and using a tiny slot reservation?