Sign in

mira

@inkreas.ing
119 followers 287 following 183 posts

she/her blog.inkreas.ing

PostsRepliesMedia
mira @inkreas.ing · 29/09/2026
why does my boring business hotel have trans-colored walls and egg-shaped mirrors???
020
mira @inkreas.ing · 28/09/2026
i don't like that it's dark outside when i wake up
021
mira @inkreas.ing · 24/09/2026
waoh thats a cool waffle picture!
130
mira @inkreas.ing · 22/09/2026
what do you mean with the comparison to tool calls? like what changes if tokens are cheaper than tool calls? would you for example put the files into the llm instead of making it call grep? sorry i really didn't understand that section 😅
100
mira @inkreas.ing · 21/09/2026
First i tried a pretty cheap power supply, but i think that one leaked some of the AC voltage onto the negative pin when the circuit was open, which was a bit uncomfortable. Inserting the needle does not get easier if a tiny bit of AC is on it!
010
mira @inkreas.ing · 21/09/2026
I upgraded my electrolysis setup from a (dying) 9V battery to a lab power supply. The higher voltage allows for a consistent higher ampere. Especially the consistency is great because i now can set the amp while not using it and still be relatively accurate.
120
Reposted by mira
opale !! @n1ark.acouphene.net · 06/08/2026
girl who has a weirdly religious relationship with Census Designated by Jane Remover
162
mira @inkreas.ing · 05/08/2026
Hii!! :33
020
mira @inkreas.ing · 31/07/2026
and cell (which in the end means unsafeCell) is only used with unsafe (as the name suggests lol) and so you either have to make unsafe harder to write or it doesn't give you anything.
020
mira @inkreas.ing · 31/07/2026
i don't even have a shipping notification yet :(
010
mira @inkreas.ing · 31/07/2026
With rust having noalias on &mut is there a advantage for it in safe code? As you said noalias is stronger. Adding it to raw pointers in safe code is also useless because you can't deref them anyways. And you would have to figure out some way to disable it in a wide enough area around unsafe.
050
mira @inkreas.ing · 27/07/2026
soooo, who wants to clicker train me so i can see in the dark?
070
mira @inkreas.ing · 22/07/2026
is it still boymoding if i wear a femtanyl shirt?
060
mira @inkreas.ing · 20/07/2026
yep! tbh i didn't understand the proposed tree borrows extension completely, but they're on now V3, so maybe i will give it another attempt.
120
mira @inkreas.ing · 20/07/2026
righs
040
mira @inkreas.ing · 20/07/2026
- work still ongoing when matching on a Option<NonNull> not being allowed to write 0 in it Also of course the safety invariants didn't change. You can't give a &uninit to safe code.
160
mira @inkreas.ing · 20/07/2026
as far as i understood the discussion it came down to a lot of additional UB with limited optimization potential. There are some "carveouts" where optimizations are useful: - &! is uninhabited (because it does not depend on the "content" of the memory)
270
mira @inkreas.ing · 20/07/2026
sorry this link is better: github.com/rust-lang/un... here the discussion actually finished. (if you ignore enum discriminants)
github.com
Should validity of a reference depend on the *contents* of memory in any way? · Issue #414 · rust-lang/unsafe-code-guidelines
This is one of the successors to #77. The question is: for a reference to be valid, what (if anything) do we require about the data it points to in memory? Note that the Rust reference currently an...
040
mira @inkreas.ing · 20/07/2026
this is actually not UB, so miri not "catching" it is correct: github.com/rust-lang/un... github.com/rust-lang/un...
github.com
Document current justification for not requiring recursive reference validity (in particular, `&mut uninit` not being immediate UB) · Issue #346 · rust-lang/unsafe-code-guidelines
This post is a draft of my understanding of the problem space, and justification for &mut uninit not being UB to hold or pass around between functions. What is UB? This is a very brief summary; see...
280
mira @inkreas.ing · 11/07/2026
i thought of my post when i saw this one lol bsky.app/profile/mado...
010
mira @inkreas.ing · 11/07/2026
"oxide computers should give me a rack" post, but i just want money for top surgery
1293
mira @inkreas.ing · 09/07/2026
i think the threshold is probably not in % of net worth, but in total net worth
030
mira @inkreas.ing · 30/06/2026
new bors picture makes me happy everytime i see him
150
mira @inkreas.ing · 25/06/2026
the heat is cooking my brain this was the worst injection i ever did
030
mira @inkreas.ing · 21/06/2026
🐇🐇🐇
me in a bunny-top
0190
mira @inkreas.ing · 15/06/2026
should i debug the C++ library that just segfaults immediately or rewrite the little stuff i need into python?
020
mira @inkreas.ing · 13/06/2026
is that about me lol? or just coincidence?
120
mira @inkreas.ing · 26/05/2026
currently just listening to music and zapping myself. very fun times
030
mira @inkreas.ing · 26/05/2026
DIY hair electrolysis is happening!! wrote a little blog about my circuit and setup. blog.inkreas.ing/electrolysis
blog.inkreas.ing
DIY hair electrolyis
292
mira @inkreas.ing · 20/05/2026
mirror selfie in a staircase. i have light, purple pants, a shirt with a bit purpley print, a light green jacket and my bag.
020
mira @inkreas.ing · 19/05/2026
replacement 2 aquired (i love crubit)
ping glowing catears
040
mira @inkreas.ing · 19/05/2026
whoooo!!
020
mira @inkreas.ing · 19/05/2026
replacement aquired
i hold cheap black car ears in my hand
010
mira @inkreas.ing · 18/05/2026
shit i can't hear anything my ears are completely closed up
000
mira @inkreas.ing · 18/05/2026
soon!!
train announcement screen showing a train going to utrecht
010
mira @inkreas.ing · 13/05/2026
don't have a real oscilloscope, but the startup looks acceptable to me.
000
mira @inkreas.ing · 13/05/2026
it says "bottom view" on the datasheet.... i was being extremely stupid
020
mira @inkreas.ing · 13/05/2026
with it now reversed it seems to keep the current pretty well with different output resistors. 100ohms poti is pretty good and i am accurate to about 0.05mA. will upload the schematics once i actually used it. (this is for diy hair electrolysis btw)
010
mira @inkreas.ing · 13/05/2026
the lm334 chips i've ordered seem to have the opposite pin-out of the datasheet. (or i am extremely stupid)
110
mira @inkreas.ing · 13/05/2026
gist.github.com/inkreasing/d...
gist.github.com
laptop benchmark results
laptop benchmark results. GitHub Gist: instantly share code, notes, and snippets.
010
mira @inkreas.ing · 13/05/2026
No, pretty much default profile. I am not yet at the stage where i measure performance lol. (you asked about rust on 8gb before and i said that i once had a LTO configure compile of a bigger app get oom killed)
110
mira @inkreas.ing · 13/05/2026
i didn't upgrade to 8gb it was already like that when i got it. just wanted to make clear that it's not completely 2009 lol
110
mira @inkreas.ing · 13/05/2026
It's a audio app and the audio is glitching a tiny bit on the debug build at this point. But i will probably try to configure cargo to optimize my audio code even in debug mode, which should solve that problem. This should let me keep the good incremental times, because it's a separate crate.
010
mira @inkreas.ing · 13/05/2026
no this incremental is still release. I just didn't delete the target folder. I don't think the performance between debug and release matters for UI. (also while it is all rust i don't use a GUI library, just wgpu so i don't know how that would go)
210
mira @inkreas.ing · 13/05/2026
i have 8 gb ram (upgraded) and a "Penryn" 2.26 GHz Intel "Core 2 Duo" (copied from the internet lol) also a ssd. if you try to run it on a hdd it would probably be horrible.
110
mira @inkreas.ing · 13/05/2026
i just ran it and a clean release build took 10 and a half minutes. after a small change an incremental build took 20 seconds. So still possible, just not for trying out things.
110
mira @inkreas.ing · 13/05/2026
oh sorry i see you are talking about a very different kind of device. i still think there are worlds between "only available to well-off devs" and "single-core 500Mhz 265MB RAM". A ten year old computer is more powerful than that and enough for the rust compiler.
030
mira @inkreas.ing · 13/05/2026
Without incremental it takes like 5 minutes, but i only have to do that when upgrading compilers or libraries, which isn't that often. Of course release builds are a different thing, but i don't need them for development.
120
mira @inkreas.ing · 13/05/2026
i don't think this is the case. I develop a desktop app (slowly starting to integrate AccessKit btw, thank you for that) and on a 2009 macbook i have incremental compile times under 5 seconds, which to me is very usable. rust analyzer diagnostics or check builds are maybe 1 second.
320
mira @inkreas.ing · 10/05/2026
came out to my mum. went pretty good (as expected) but she still said a bunch of weird stuff (as expected)
070