Sign in

Luka

@lukajcb.bsky.social
82 followers 95 following 47 posts

github.com/LukaJCB

PostsRepliesMedia
Luka @lukajcb.bsky.social · 10/10/2026
It was a hellish game for everyone 😔
000
Luka @lukajcb.bsky.social · 08/10/2026
this would be insane if true
010
Luka @lukajcb.bsky.social · 08/10/2026
Why rip?
100
Reposted by Luka
Gabriel Volpe @gvolpe.com · 25/08/2026
This was very hard to write, but couldn't be silent any longer. gvolpe.com/blog/tax-pay...
gvolpe.com
A tax-paying prisoner in Poland
Today is my birthday, but I’m not in the mood to celebrate. Today marks 131 days since my legal status in Poland has basically become tax-paying prisoner, …
142
Luka @lukajcb.bsky.social · 30/07/2026
Huh idk the bot is still wrong all the time for me
010
Luka @lukajcb.bsky.social · 25/07/2026
this looks a lot like wandrer.earth!
010
Luka @lukajcb.bsky.social · 03/07/2026
It's accurate that it's a skill issue, or what Cory originally stated is accurate?
2130
Luka @lukajcb.bsky.social · 27/05/2026
What do you use to do away with it? I have a pretty bespoke eslint config + plugins + madge + knip, and yet I still get bit sometimes by comparing objects or Arrays with `===`
000
Luka @lukajcb.bsky.social · 19/05/2026
ah gotcha, I thought it was this: rust-unofficial.github.io/too-many-lis... The smart pointers are a bit clunky to work with IMO, but coming from C++'s shared_ptr it's pretty similar.
rust-unofficial.github.io
Introduction - Learning Rust With Entirely Too Many Linked Lists
Learning Rust With Entirely Too Many Linked Lists
010
Luka @lukajcb.bsky.social · 19/05/2026
Is this from the learning Rust with Linked Lists? lol
120
Luka @lukajcb.bsky.social · 15/05/2026
Will they be encrypted?
110
Luka @lukajcb.bsky.social · 15/05/2026
yeah, that's fair. I don't know much how bun is being developed at all. But if any of the OSS projects I've ever worked on pulled something like this I would be livid.
010
Luka @lukajcb.bsky.social · 14/05/2026
To me the concern isn't that you changed your mind or became convinced one way or the other, but the fact that you committed to a million line rewrite in one week. I'm curious about the blog post, but there are sure to be a lot of consequences from a decision like this.
140
Luka @lukajcb.bsky.social · 12/05/2026
monads being an antimeme is itself a meme.
000
Luka @lukajcb.bsky.social · 06/05/2026
Go really has made the worst decision for error handling, instead of return a result OR an error, they make you return a result AND an error
000
Luka @lukajcb.bsky.social · 17/04/2026
Similar number of rides but much more time
2278 rides 3652 miles 540 hrs
000
Luka @lukajcb.bsky.social · 23/03/2026
I do have questions about the write notifications though, a while ago you mentioned that it shouldn't be possible for private activity to be publically viewable. bsky.app/profile/dhol... If I understand correctly it's PDS -> Space Host -> AppViews. So there's some leakage there, but not as wide?
110
Luka @lukajcb.bsky.social · 23/03/2026
The ECMH stuff is really cool, I'm cautiously optimistic this is going to be great!
110
Luka @lukajcb.bsky.social · 13/03/2026
How about Pod and Cell? I.e. a Pod with 10 members each having their own Cell. Or perhaps Enclave and Chamber?
000
Luka @lukajcb.bsky.social · 13/03/2026
Yeah that's sort of what we were discussing in the replies of a previous post of daniel's here: bsky.app/profile/luka...
001
Luka @lukajcb.bsky.social · 13/03/2026
Eager to hear more about how the sync happens!
111
Luka @lukajcb.bsky.social · 08/03/2026
I think this is kinda real though, people that don't rely on turn by turn navigation tend to have a much better sense of direction IME
020
Luka @lukajcb.bsky.social · 03/03/2026
What a comeback!
000
Luka @lukajcb.bsky.social · 26/02/2026
MLS makes rotation fairly cheap, i.e. removing someone from a group is O(log n) complexity.
010
Luka @lukajcb.bsky.social · 26/02/2026
Interesting, will have to think about that more!
010
Luka @lukajcb.bsky.social · 26/02/2026
HMAC would require everyone in the bucket to agree to the same key, so you would likely need a GKA protocol like MLS, right?
230
Luka @lukajcb.bsky.social · 26/02/2026
I like this quite a bit, the buckets idea maps somewhat cleanly to MLS groups, though there is of course still a lot of unanswered questions.
010
Luka @lukajcb.bsky.social · 19/02/2026
Right the message in this case is only the metadata that we want to protect
010
Luka @lukajcb.bsky.social · 19/02/2026
Yeah and that's a very hard problem space from what I can gather. Essentially how can we send a private message from A to B, C, and D across servers? All without the servers trusting each other and without leaking metadata.
120
Luka @lukajcb.bsky.social · 19/02/2026
I like this a lot, keeps a lot of the existing infrastructure. I could see something like this working well with E2EE data as well. The metadata leakage is somewhat unfortunate, but unless we add some sort of authenticated relay, I'm not sure there is a better way.
010
Luka @lukajcb.bsky.social · 19/02/2026
I think that narrows down the solutions space quite a bit then, no? If that kind of metadata shouldn't be public then that also eliminates publishing anything to the firehose for that content and we would need some kind of separate notification system so that an AppView can know of the new content.
130
Luka @lukajcb.bsky.social · 13/02/2026
I think this is good! I've been thinking about this space for a while and I think having permissioned data like this could be used for both e2ee and non-e2ee systems. Though, crucially I think how that data goes from the PDSs to the AppViews should also take into account e2ee and non-e2ee systems.
020
Luka @lukajcb.bsky.social · 08/02/2026
Do you have some links to where the prototyping is happening? I would be interested to collaborate!
110
Luka @lukajcb.bsky.social · 23/01/2026
In any sane system you would not want to perform expensive queries for analytics purposes on the real-time parts of your storage layer
000
Luka @lukajcb.bsky.social · 01/01/2026
Amazing! I've been trying to decide between Ente and Immich as well, might have to try this out!
110
Reposted by Luka
Soatok, the Furbidden One @soatok.bsky.social · 14/10/2025
This is all old news. What's new is this: There is now a first draft reference implementation of the Public Key Directory specification, written in Go. github.com/fedi-e2ee/pk...
github.com
GitHub - fedi-e2ee/pkd-server-go: Reference implementation - server-side software for the Fediverse Public Key Directory Specification
Reference implementation - server-side software for the Fediverse Public Key Directory Specification - fedi-e2ee/pkd-server-go
13611
Luka @lukajcb.bsky.social · 14/10/2025
Very exciting!
000
Luka @lukajcb.bsky.social · 05/09/2025
Side note I wrote about something pretty similar many years ago here: typelevel.org/blog/2018/11...
typelevel.org
Typelevel | A tale on Semirings
Typelevel is an ecosystem of projects and a community of people united to foster an inclusive, welcoming, and safe environment around functional programming in Scala.
010
Luka @lukajcb.bsky.social · 05/09/2025
And more general the cardinality of a function A => B is B to the power of A. In this example: 1^X = 1 X^1 = X
120
Luka @lukajcb.bsky.social · 05/09/2025
To add to this you can think of cardinalities as how many different implementations can there be for this signature. So Boolean -> Unit has one possible implementation, whereas Unit -> Boolean has two
120
Reposted by Luka
Soatok, the Furbidden One @soatok.bsky.social · 25/08/2025
Barking Up The Ratchet Tree – MLS Is Neither Royal Nor Nude One of the first rules you learn about technical writing is, "Know your audience." But often, this sort of advice is given without sufficient weight or practical examples. Instead, you're ushered quickly onto the actual tactile aspects of…
soatok.blog
Barking Up The Ratchet Tree – MLS Is Neither Royal Nor Nude
One of the first rules you learn about technical writing is, "Know your audience." But often, this sort of advice is given without sufficient weight or practical examples. Instead, you're ushered quickly onto the actual tactile aspects of writing--with the hope that some seed was planted that will sprout later in your education. Science communication is famously a hard problem.
0306
Luka @lukajcb.bsky.social · 30/07/2025
It did lead me to find about CHURP though and that to me looks like a great solution for backup: github.com/CHURPTeam/CH...
github.com
GitHub - CHURPTeam/CHURP: Decentralize your secrets!
Decentralize your secrets! Contribute to CHURPTeam/CHURP development by creating an account on GitHub.
010
Luka @lukajcb.bsky.social · 30/07/2025
Funnily enough I came across this yesterday as well. It looks great although to me the fact that you still need to remember a password is a bit of a downside. I think id rather just use a PBKDF at that point
110
Luka @lukajcb.bsky.social · 05/07/2025
Just published 1.0.0 of ts-mls, a library implementation of MLS written in typescript, put a lot of work into this and really with how it turned out! github.com/LukaJCB/ts-mls
github.com
GitHub - LukaJCB/ts-mls: A Messaging Layer Security (RFC 9420, MLS) implementation in TypeScript
A Messaging Layer Security (RFC 9420, MLS) implementation in TypeScript - GitHub - LukaJCB/ts-mls: A Messaging Layer Security (RFC 9420, MLS) implementation in TypeScript
040
Luka @lukajcb.bsky.social · 27/05/2025
The Web Crypto API for browsers and Node does this by basically treating HMAC and signatures the same way with a sign and verify function for each
010
Luka @lukajcb.bsky.social · 18/05/2025
No worries, thank you for fixing it!
000
Luka @lukajcb.bsky.social · 18/05/2025
I tried to create an account for atproto.wiki, but I never got the verification email so I can't complete the sign up
atproto.wiki
AT Protocol Community Wiki
Homepage of the AT Protocol Community Wiki
120
Luka @lukajcb.bsky.social · 16/05/2025
Would love to collaborate on this, I'll have to give it a full read soon. I'm currently also working on creating a TypeScript MLS library that is almost complete that should make prototyping a lot easier!
020
Luka @lukajcb.bsky.social · 16/05/2025
I'm really glad someone is thinking about this! I wrote a prototype of using MLS to create E2EE content in ATProto recently that I've described in a GH issue here: github.com/bluesky-soci...
github.com
Private, non-shared data in repo? · bluesky-social atproto · Discussion #3363
I have a use-case for keeping private data in a repo, i.e. data that does not leave the repo via the standard sharing mechanisms, and before I leap into modifying the code to support it, I'd like t...
150
Luka @lukajcb.bsky.social · 12/02/2025
Hello world!
120