Grégoire Locqueville

Hi, I'm Grégoire, a person passionate about a bunch of things, including maths, software, music, voice, human and computer languages, and virtual places.
Feel free to get in touch!
Private account: Grégoire Locqueville (npub1pcc…cn6z)

Blog: https://glocq.com/en/blog/

Public Key

npub1pcclfh3rg0zpetm79x6ax68gjzgtep7xzamchl6ek5yydmygrces25cn6z

NIP-05 Address

[email protected]

Profile Code

nprofile1qqsquv05mc3583qu4alzndwndr5fpy9uslrpwautlavm2zzxajypuvcpz4mhxue69uhhyetvv9ujumn0wd68ytnwv46qz9rhwden5te0wfjkccte9ejxzmt4wvhxjmca88kag

Publishing to

relay.nostr.net

relay.damus.io

nos.lol

Author Public Key

npub1pcclfh3rg0zpetm79x6ax68gjzgtep7xzamchl6ek5yydmygrces25cn6z

Show more details

Last Notes

2026-06-20 16:56:30 UTC

Ever wanted to make a game that takes place in a procedurally generated world, but don't want to deal with the hassle of instantiating chunks as your player moves?
With my #fernweh library for #bevy, you declaratively define what chunks are made of, using the brand new Bevy scene system, and fernweh takes care of asynchronously computing and spawning/despawning them as needed.
Fernweh is *very* generic: it doesn't care if your game is 2D, 3D, or whatever else. *You* get to decide how to refer to chunks, *you* get to decide what chunks should be visible at what point, and chunks may be composed of *anything* that can be represented as a Bevy scene, that is, pretty much anything you could want (meshes, colliders, audio sources...).
https://crates.io/crates/fernweh
#rust #RustLang #BevyEngine #procgen #ProceduralGeneration #bsn

2026-05-09 11:58:53 UTC

- reply

Interesting, I've seen software people complain/warn about this problem's counterpart in software development: companies don't want to hire junior developers anymore, because they can just use LLMs for tasks that juniors used to do; but no junior developers today means no senior developers tomorrow.

2026-05-07 11:08:14 UTC

- reply

There is no recorded case of schizophrenia in anyone congenitally blind. No one knows why.

2026-04-10 23:10:44 UTC

- reply

Does it really qualify as microtonality when there's no precise target pitch, though?
@npub1nf4…nqe4

2026-03-25 17:16:21 UTC

- reply

> Physics is very different from math. It involves understanding at a gut level what particles and fields tend to do in different situations.
Right, but can't the language of math help those who understand it build an intuition for a physical theory? Not as a means to get a rigorous, complete picture, just as a means of communication.
@npub18a6…szuj

2026-01-25 18:23:47 UTC

It seems that installing audio plugins on #NixOS is its own can of worms 🙃

2026-01-20 07:51:28 UTC

- reply

What do you mean? Doesn't
```
someArray.join(", ", last: ", and ")
```
work?

2026-01-10 05:22:40 UTC

- reply

Thanks a lot!

2026-01-10 04:42:44 UTC

- reply

Now I'm curious what conceptual frameworks Poincaré suggested?

2025-12-25 11:02:56 UTC

- reply

"lump in the carpet" is a great analogy

2025-11-28 22:30:57 UTC

- reply

It's honestly shocking how charitable the people in that thread are, with some spammer who basically just dumped tens of thousands of lines of code that he has not even bothered to understand (let alone write) and expects them to review it...

2025-11-27 21:25:38 UTC

- reply

As a French, we were actually asked to solve the French version of this problem in middle school: how many layers are in a croissant?
(ok, the problem was actually much simpler (obviously — we were in middle school) since it was just about integer powers, but I'm pretty sure it can be made to resemble what's discussed in the paper if you start thinking about whether you fold from the left or from the right at each step)

2025-07-07 22:41:54 UTC

- reply

It's half past midnight and I'm nerd-sniped 😭

2025-07-07 22:32:34 UTC

- reply

Ohh I missed that 🙄 I thought it had as many steps as elements in the cycle, which is obviously false

2025-07-07 22:26:04 UTC

- reply

Unless I'm missing something, you need as many steps as you have elements, minus the number of elements that remain unchanged:
- You "obviously" need at least as many steps, because otherwise there would be an element that does not change
- You need at most as many steps, because you can decompose any permutation into a product of cyclic permutations with disjoint orbits, and you can apply your little algorithm for cyclic permutations for each of those

2025-02-22 16:14:05 UTC

- reply

A related observation:
"A burrito is like a functor: it takes a type, like meat or beans, and turns it into a new type, like beef burrito or bean burrito."
https://blog.plover.com/prog/burritos.html

2024-11-30 13:02:24 UTC

- reply

Ok but I still don't get why he should "see her at a close distance"?

2024-11-30 11:50:46 UTC

- reply

> Bob will eventually see her at a close distance
I don't think so. From Bob's conceptual perspective, Alice's progress is stalled near the horizon only as long as Bob is a distant observer. Once he approaches the black hole, there's a time on his clock at which he can say she's crossed the horizon, even though the light of Alice crossing the horizon hasn't reached him yet.

2024-11-30 11:20:43 UTC

- reply

I assume "question 2" refers to your question about Alice being flattened in my scenario where you follow her from a short distance. But in that scenario I was assuming you were following her from close enough (compared to the size of the black hole), at the same speed.
If you go slow enough and leave enough space between some ships and you that you start to see them flatten, I think (not 100% sure of that) you will still see them flat (but not infinitely flat!) while you're crossing the horizon. But I may be wrong, and you might see them progressively get (some of) their thickness back as you're approaching the horizon.
Either way, I don't see anything paradoxical?

2024-11-30 07:06:40 UTC

- reply

Correct
@npub1nf4…nqe4 @npub12cu…d38p

2024-11-30 00:40:02 UTC

- reply

I think you need to make it clear if the observer is going into the black hole with the object or not.
1. If not, then yes they will be seeing the object flatten above the horizon to avoid crossing it.
2. If they are going into the black hole too, then things will look normal, and they will see the object cross the horizon just while they are crossing it themselves (see my other toot).
3. If they follow the object, but accelerate backwards enormously at the last minute, to avoid crossing it, then they will first be seeing a normal image of it before it crossed the black hole (see my previous toot), then after the backwards acceleration they will suddenly see it flattened because their relative speed will have become suddenly high.
You can try to imagine all kinds of scenarios like that, and the answer will generally be somewhere on the spectrum between 1. and 2.
@npub1nf4…nqe4 @npub12cu…d38p

2024-11-27 21:27:26 UTC

- reply

Just so we're clear: you're suggesting that the bacterium in my eye shoot a tiny laser towards my nose, on which a tiny mirror will have been placed, allowing the laser light to bounce on it and get back to my eye; and that the bacterium measure the time (on its clock) between the moment it shoots the laser and the moment it gets its light back. Right?
Now you say my nose is "receding", and my eye is "running towards" it — but what frame of reference are we talking about now? I thought we were talking about my local frame of reference, in which neither my eye nor my nose are moving! If we're talking a distant observer's point of view, we've already established they do indeed think I am flattened!
@npub1nf4…nqe4 @npub12cu…d38p

2024-11-27 21:07:44 UTC

- reply

What experiment do you suggest I do, once near the horizon, to measure the thickness of my nose in my frame of reference?
@npub1nf4…nqe4 @npub12cu…d38p

2024-11-27 18:34:19 UTC

- reply

No, the bacterium is also flattened (from the point of view of a distant observer), along with its local environment, including my nose, and thus it does not notice that my nose is flattened — there's no local experiment it can perform to determine that my nose is in fact flat, any more than it could in space away from the black hole.
It's the exact same in special relativity. If you and I were flying close to light speed relative to each other, from the perspective of each of us the other would be flattened, yet we wouldn't think of ourselves as flattened.
@npub1nf4…nqe4 @npub12cu…d38p

2024-11-27 18:07:20 UTC

- reply

Your first paragraph is correct *as long as you're looking from far enough from the horizon*.
The second paragraph is incorrect. Locally, you are still in euclidean spacetime before, during, and after your crossing of the *event horizon*; this only becomes false as you approach the *singularity*.
Sure, if you're close to the horizon, people far away from the black hole do see you flat and redshifted, but you don't agree with them!
@npub1nf4…nqe4 @npub12cu…d38p

2024-11-25 23:47:50 UTC

- reply

The first ship does actually "wait for you", i.e. its clock slows down relative to yours and it slows down so much it appears to stop at the even horizon — that is, until you go to the event horizon yourself.
(It also gets redshifted more and more to the point you can't see it anymore — again, until you go to the event horizon yourself)

2024-11-25 22:28:56 UTC

- reply

Alessandro Roussel, aka ScienceClic, also makes great GR- and black hole-related visualizations
https://www.youtube.com/watch?v=wrwgIjBUYVc

2024-11-25 19:16:45 UTC

- reply

Ahh you're right, I didn't think about that. So I guess what I initially thought was a volume is actually a half-line. When will shrinkflation stop 😢
@npub1glg…tqzy @npub1x4d…6lp6 @npub12cu…d38p @npub1hnp…26ht

2024-11-25 19:02:47 UTC

- reply

Thank you, I corrected my reply. Am I (locally) correct now talking about a half volume instead of a volume, and saying Bob "sees" the horizon everywhere in front of him as he's crossing it?

2024-11-25 18:18:02 UTC

- reply

If Bob could see the horizon, at the moment he crossed it, he would see it in front of him, overlapping him, and behind him all at once, like an invisible wall passing him at light speed. When Bob crosses the horizon, it appears to him everywhere — like a volume instead of a surface. But Alice only appears at one point in this volume, in front of him.

2024-11-25 18:07:58 UTC

- reply

He sees her in front of him, but by the time the photons coming from the event "Alice crosses the horizon" reach Bob, the horizon also reaches him.

2024-11-08 15:17:12 UTC

- reply

It certainly depends on your definition of "implicit", but for what it's worth, I did have to scratch my head about this specific dependency (if I remember correctly).
pkg-config works if the dependency is installed on your system, but it's useless if you don't have the dependency installed already, right?

2024-11-07 17:42:00 UTC

- reply

I am guessing the lack of examples is a consequence of library authors relying on the types too much

2024-10-17 05:58:30 UTC

- reply

I see :) (and to be clear, I wasn't criticizing you) It's just you tend not to share cutting-edge theoretical talks on here, and to target a slightly less technical audience, so it's funny seeing you kind of assuming, in this post, that your followers are at that level (even though I know many actually are)

2024-10-17 05:51:47 UTC

- reply

Not gonna lie, your "He also keeps prerequisites to a minimum. Well, okay: if don't know quantum field theory don't watch this."
got a chuckle out of me.

2024-09-29 03:44:38 UTC

I know it's totally subjective and I'm completely biased in favor of the latter, but #Rust feels so much "bigger" than #Haskell
Sure, depending on your background Haskell might be very hard to grasp*, and GHC is huge and complicated, and a lot of Haskell code hurts the brain, but in terms of core concepts and syntax, I actually feel it's pretty straightforward :
- Basic function application is as simple as can be. The one complication is infix operators/functions and precedence rules
- Pattern matching is immediately understandable
- At the type level, you need to know about algebraic data types and record types
- And then there's a sugar layer on top of that, with a couple of constructs that you need to know: string literals, if-then-else, do notation... only the latter of which may be confusing.
I feel like the rest is either building on that, or not essential to the language?
Whereas in Rust you have to grasp:
- Most of that too (functions, some operators, algebraic data types, records, pattern matching)
- Plus general imperative control flow structures
- Plus low-level stuff like pointers and references
- And then Rust's own innovations (everything related to borrowing)
- ... and some confusing, ad hoc sugar too; see https://cheats.rs/#language-sugar
*Not getting into the debate of whether one of Haskell or <insert your favorite imperative language here> is easier to learn as a first programming language here