Alex Burdiyan
@burdiyan.com
Software engineer. Co-founder at Mintter.
Tell us more about your setup :)
September 9, 2025 at 1:58 PM
Tell us more about your setup :)
I thought Java was a verbose language until I saw Swift. Besides being terribly and annoyingly verbose, the most atrocious part of Swift for me is the idea of function argument labels.
docs.swift.org/swift-book/d...
docs.swift.org/swift-book/d...
August 31, 2025 at 5:58 PM
I thought Java was a verbose language until I saw Swift. Besides being terribly and annoyingly verbose, the most atrocious part of Swift for me is the idea of function argument labels.
docs.swift.org/swift-book/d...
docs.swift.org/swift-book/d...
Same with dishwasher detergent for me 😅. But now I try to buy an extra package as I’m half-way through the current one — this way I have at least a few days of leeway.
August 31, 2025 at 5:50 PM
Same with dishwasher detergent for me 😅. But now I try to buy an extra package as I’m half-way through the current one — this way I have at least a few days of leeway.
What are your thoughts on libp2p? I remember you've said some good things about it in the past.
July 10, 2025 at 9:42 AM
What are your thoughts on libp2p? I remember you've said some good things about it in the past.
You can actually do `str.join(" ", ["a", "b", "c"])`, but I agree — it's absurd that the form you mentioned is the common one :)
May 25, 2025 at 1:13 PM
You can actually do `str.join(" ", ["a", "b", "c"])`, but I agree — it's absurd that the form you mentioned is the common one :)
I'm not particularly happy about this pattern though. It forces callers to `break` from the iteration instead of just returning, to make sure the cleanup error is actually checked.
Granted, in this particular case the cleanup error could probably be just ignored :)
github.com/seed-hyperme...
Granted, in this particular case the cleanup error could probably be just ignored :)
github.com/seed-hyperme...
seed/backend/util/sqlite/sqlitex/query_iter.go at adc48d1cbd7147a89fc893e6a3d6ae8e06a2f22e · seed-hypermedia/seed
Seed Hypermedia Desktop app. Contribute to seed-hypermedia/seed development by creating an account on GitHub.
github.com
May 16, 2025 at 9:20 PM
I'm not particularly happy about this pattern though. It forces callers to `break` from the iteration instead of just returning, to make sure the cleanup error is actually checked.
Granted, in this particular case the cleanup error could probably be just ignored :)
github.com/seed-hyperme...
Granted, in this particular case the cleanup error could probably be just ignored :)
github.com/seed-hyperme...
I've been trying to use iterators for SQLite results, and too missed some established patterns for error handling and cleanups. I ended up returning a `check func() error` along with the `iter.Seq`, so the caller would have to call `check()` after iterating over the rows to check any cleanup errors.
May 16, 2025 at 9:20 PM
I've been trying to use iterators for SQLite results, and too missed some established patterns for error handling and cleanups. I ended up returning a `check func() error` along with the `iter.Seq`, so the caller would have to call `check()` after iterating over the rows to check any cleanup errors.
Riiiight! And the street parking system, because it’s mostly solar-powered (at least in Spain). The whole country without power and I went down to pay for my parking with a credit card like a dumbass — and it worked 😁
May 8, 2025 at 8:10 AM
Riiiight! And the street parking system, because it’s mostly solar-powered (at least in Spain). The whole country without power and I went down to pay for my parking with a credit card like a dumbass — and it worked 😁
I've been following Pants 2 for a while before, but last time I checked it still wasn't supporting Windows in any way. Is it still the case?
May 7, 2025 at 8:20 PM
I've been following Pants 2 for a while before, but last time I checked it still wasn't supporting Windows in any way. Is it still the case?
At first it wasn’t clear tbh, because you just mentioned that the experience is not comparable and drastically different. But now after re-reading the post a few times, I guess you mean Cursor is worse, which I inferred from the fact you downloaded it again, so you gave it another try.
May 7, 2025 at 11:09 AM
At first it wasn’t clear tbh, because you just mentioned that the experience is not comparable and drastically different. But now after re-reading the post a few times, I guess you mean Cursor is worse, which I inferred from the fact you downloaded it again, so you gave it another try.
So, which one is better? :)
May 7, 2025 at 9:24 AM
So, which one is better? :)
> Kinda disappointing that current LLM tooling can't automatically fetch Go docs or interface with the LSP yet.
Couldn't this be solved by MCP?
Couldn't this be solved by MCP?
April 10, 2025 at 9:36 PM
> Kinda disappointing that current LLM tooling can't automatically fetch Go docs or interface with the LSP yet.
Couldn't this be solved by MCP?
Couldn't this be solved by MCP?
So I guess this would bring us back to the lang of some blockchain-based consensus, or some kind of anarchy where each user decides for themselves which fork to trust.
April 10, 2025 at 9:33 PM
So I guess this would bring us back to the lang of some blockchain-based consensus, or some kind of anarchy where each user decides for themselves which fork to trust.
Another reason why federating PLC could be hard is this 72-hour window during key rotation. If the chain of PLC operations could get forked, every replica would somehow need to resolve this conflict.
April 10, 2025 at 9:33 PM
Another reason why federating PLC could be hard is this 72-hour window during key rotation. If the chain of PLC operations could get forked, every replica would somehow need to resolve this conflict.
That could probably have a workaround — by making the managed PDS sign a DID operation to add a custom highest-priority rotation key which would be stored by the user and could be used for this "recovery" scenario, but I'm pretty sure it's not a very well supported use case *yet*.
April 10, 2025 at 9:33 PM
That could probably have a workaround — by making the managed PDS sign a DID operation to add a custom highest-priority rotation key which would be stored by the user and could be used for this "recovery" scenario, but I'm pretty sure it's not a very well supported use case *yet*.
So if PLC goes bad, presumably Bluesky goes bad too, hence your "managed" PDS too, which means that you can't update your DID document even if you replicate it, because the keys are on the PDS.
April 10, 2025 at 9:33 PM
So if PLC goes bad, presumably Bluesky goes bad too, hence your "managed" PDS too, which means that you can't update your DID document even if you replicate it, because the keys are on the PDS.
I've been digging into the self-sovereignty of the PLC Directory identity the other day too. But I was mostly wondering whether I could own my signing keys, and seems like this is where the whole thing kinda falls apart, unless you fully self-host the PDS (which is much harder than storing keys).
April 10, 2025 at 9:33 PM
I've been digging into the self-sovereignty of the PLC Directory identity the other day too. But I was mostly wondering whether I could own my signing keys, and seems like this is where the whole thing kinda falls apart, unless you fully self-host the PDS (which is much harder than storing keys).
If my passkeys are never sent to any servers and never exposed anywhere at all—everything is client-side, what if I derive an encryption key from some combination of credential ID, RP ID, and user ID (which would be some random 128/256 bit salt generated when passkey was created)?
March 14, 2025 at 12:00 AM
If my passkeys are never sent to any servers and never exposed anywhere at all—everything is client-side, what if I derive an encryption key from some combination of credential ID, RP ID, and user ID (which would be some random 128/256 bit salt generated when passkey was created)?
Given that PRF extension is very recent and not widely available, I was thinking about workarounds. I wonder if what I came up with is insane or fine :)
…
…
March 14, 2025 at 12:00 AM
Given that PRF extension is very recent and not widely available, I was thinking about workarounds. I wonder if what I came up with is insane or fine :)
…
…
Hahaha, yeah, it’s just the other side of the politics here.
March 5, 2025 at 11:31 AM
Hahaha, yeah, it’s just the other side of the politics here.