birdhalfbaked.com
Based in Stockholm, Sweden. Data engineer that works at Exabyte scale. Loves solving problems. I also have a kickass sailboat that is a testbed for many projects. Also have background in bioinformatics and some machine learning.
https://birdhalfbaked.com
150 posts
22 followers
17 following
Regular Contributor
Active Commenter
comment in response to
post
The smear is excellent
comment in response to
post
Bonus: if you disrupted the whole day like happened, it cost the US the worth of 2 billion. One email for 2 billion. Elon maybe didn't intend it but wasted about that much of tax payer money by not looking before leaping. They will do the exercise again supposedly. What a waste of money.
comment in response to
post
Thank you! ♥️
comment in response to
post
I am always fascinated by printers and scanners being the epitome of lack of documentation and driver support. I feel your pain.
comment in response to
post
on second read almost sounds like I did not enjoy the family visits. Quite the opposite though and already missing them and mexican food 🤤
comment in response to
post
Here's the spec of why this works. The content length is 3000. github.com/bluesky-soci...
comment in response to
post
The lexicon means you can separate supported implementations from the PDS and App. So an app can specify also what operations it expects a PDS to implement. Lexicons are like the contracts so in that regard some implementation is indeed tied to apps but data always is separated
comment in response to
post
you can as the data is separate from the application. If you have your own PDS, and bluesky shuts down the app your data is still accessible by you and other apps that hook into ATproto. It's the same way Mastodon works. AP is analogous to ATproto and Bluesky is analogous to Mastadon the app
comment in response to
post
You get sun till 4:30? Paradise!!!
comment in response to
post
Got a great clarification on this using an older thread from some days ago. Thanks to @fei in the discord for the call-out to this thread bsky.app/profile/duck...
Clarifies a lot
comment in response to
post
I think this is the thing people get caught up in a lot. Bluesky the app does have centralized features but at the end of the day the fact i can just spin up my own server and keep chugging is the big win. And to be fair mastodon did it well too. Both players are doing great
comment in response to
post
The points were interesting though i don't know if the core assumption holds still. There are many people that have indeed migrated to private servers already where their private server is interacting with the bluesky relay which is similar to the federation provided by Mastodon via AP.
comment in response to
post
Wow embarrassing. Autocorrect does NOT like the name Cramling and I can't edit :( oh well I'll leave my shame
comment in response to
post
Thanks! Will try to dig in when I'm back home but I've joined for now and am doing quick catch up on recent discussion! Thanks a bunch
comment in response to
post
Got a link? I mean I never knew there's a discord and I've been diving alone for the most part and making a lot of headway here
comment in response to
post
Let's make one! Beauty of open source protocols, we don't need to wait ;)
comment in response to
post
That said users can leak data of course, but that's not possible to sure up 100% even if you encrypt it. Nothing stopping a bad actor from just getting true blue access then leaking it.
comment in response to
post
You can also implement a custom resource authorization that would apply to followers only. Remember, Blueskys implementation is only a reference and the ATProto does not dictate a lot for PDS implementation outside of how content sharing happens. Well have a write up on this soon
comment in response to
post
Honestly the best if it's for estimation would be if you controlled the schema. You can generate your own samples, get a rough mean of the sizes and use that as a baseline size per row.
comment in response to
post
If you do need accuracy well unfortunately you're trading off error for speed. It's gonna come down to row sampling of some sort. Not enough info on what it is you're actually wanting to do (show a user some estimation? Apply the estimation to bandwidth managers? Something else?). Good luck!
comment in response to
post
There are ways to have private content separated. The reference implementation doesn't do it ofc, but fact PDSs can even manage multiple repositories per handle (and one can be private). was diving into it this weekend though it isn't trivial. No reference implementation for that yet though
comment in response to
post
This. I need this.