tychoish (sam)

I'm here less these days, but you can hit me up on telegram t.me/tychoish I've also made this telegram channel where I've been putting the things that I'd otherwise put here: t.me/tychoish_pub Blog announcements will continue.

@Zephraph Yeah, happy to chat. DM me (or same handle on telegram)

Do I have any mutuals who are tech contractors/freelancers in the state of NY? I'd love to chat about your experience / process / setup / etc.

@GMMeyer event ant? ;-)

my most iconoclastic take is that i hate makefiles. I think they are more difficult to read, understand and edit than basically any alternative, using them to do anything more complex than a simple build winds up in a mess of the worst kind

I'd like someone to make a rust library (Go would also be interesting) that would make it easy a binary shebangable.

@mim_djo @GlareDB anyway, it was really fun to work on this feature. Down to answer questions or help if you or folks need/want.

@mim_djo @GlareDB it's kind of wild that write support is so absent: particularly when analytics workflows are so often centered on materialization. it's such a little thing with such big (potential) impacts.

this is kind of big deal !!! as far as i am aware @GlareDB is the first of the "new wave" open source OLAP engine that support write ( insert and delete ) to delta table

@arthurdenture This but really just mmap() on your parquet files: blazingly fast analytics database.

the working definition of "how big is big data" has to be whatever the maximum EBS volume size is.

so 64 TiB right now...

@criccomini I'm interested in how you define fun in other areas of your life. 😂

🤔 single process, non-distributed Kafka (with pluggable storage?).. would be a fun Rust/Go project…

@hdevalence hero

> wow, look at this super tricky problem, we need advanced formal methods to have a hope at solving it ok. but what if we tried using basic software engineering first?

@jackzampolin same! :(

@tychoish And they've got none in my size 😭😭😭

Looking for a good pair of khaki shorts (ideally sold in a few colors) that are good for hiking/summer activities. Does anyone have recommendations?

Want to build a @getdbt pipeline that can seamlessly connect across different data sources? You can do that with GlareDB! Check out the video for a preview, this repo for an example (twtr.to/Ma2xu), & this post for a step-by-step walkthrough (twtr.to/NqZkQ).

@cwgoes I can ignore many more people than meet this qualification😂

If someone uses the word "decentralize" in their sentence, but on request cannot replace it with a more specific description of an actual mechanism involving actual (they can be example) participants, they don't understand what they're talking about and you can ignore them.

@comradesanchez and crown heights? ;)

as seen in Williamsburg, but Instagram saw it first. will delete later ...

@BillyBamf @Noahpinion The "society abuses smart young people who aren't responsible if they commit genocide"

@Noahpinion The whole conceit of the end is that he doesn’t understand what he’s doing.

If you like Ender's Game, I hope you realized that Ender is the villain

@monkchips general purpose database people often think that all data tooling is a feature of a general purpose database.

Larry Ellison thinks vectors are a feature of a general purpose database rather than a product in their own right? weird.

@criccomini @warpstream_labs also capacity planning in highly optimized systems is super hard.

Pretty brutal takedown of tiered storage from the @warpstream_labs folks.

@mark0baricevic The infrastructure is never the product driver.

First it was rollups and app chains. Then it was all app chains. Next it will be L2s aren’t real and everything is an L1. Bamboozled

@editingemily WHY ARE WE ALWAYS THE SAME

Them: What’s your love language? Me: Aaron Sorkin dialogue.