Notices by Kate (k4t3@comp.lain.la)
-
Embed this notice
Kate (k4t3@comp.lain.la)'s status on Sunday, 09-Feb-2025 07:50:00 JST Kate
@mia @lucy the worst part is, this effectively combines the worst of both worlds, the codegen/"JIT" doesn't do any optimizations, so it isn't even much faster than a well thought-out interpreter, whilst still carrying all the difficulties that implementing a JIT brings.
You could *easily* write a similarly capable VM in under 300sloc that's far more readable whilst still retaining similar performance.
What i've linked just heavily irritated me.
Even reading through that thing multiple times, I barely have any idea what those constants are meant to be / what those variables are supposed to represent.
What really tripped me up however is how they're just static globals that make all of the codegen that uses it prone to race conditions. -
Embed this notice
Kate (k4t3@comp.lain.la)'s status on Tuesday, 21-Jan-2025 02:56:35 JST Kate
@meso it's a psyop, there's no question the cia is involved here /s -
Embed this notice
Kate (k4t3@comp.lain.la)'s status on Tuesday, 21-Jan-2025 02:54:26 JST Kate
we need more people with the "<thing> is a mess, im gonna build my own with blackjack and hookers" mindset.
It's among one of the best mindsets one can have, best case something improves, worst case nothing happens. -
Embed this notice
Kate (k4t3@comp.lain.la)'s status on Sunday, 12-Jan-2025 07:27:37 JST Kate
@meso "goals" - meso -
Embed this notice
Kate (k4t3@comp.lain.la)'s status on Saturday, 11-Jan-2025 06:56:52 JST Kate
@meso will you let me drive? :meowPuffyPleading: -
Embed this notice
Kate (k4t3@comp.lain.la)'s status on Friday, 10-Jan-2025 21:11:28 JST Kate
@meso Not beating the AGP Allegations. -
Embed this notice
Kate (k4t3@comp.lain.la)'s status on Thursday, 09-Jan-2025 22:37:28 JST Kate
@maija are you on Prog?
Also, do you experience hot flashes? Whats your e2 levels? -
Embed this notice
Kate (k4t3@comp.lain.la)'s status on Thursday, 09-Jan-2025 22:15:51 JST Kate
Falling into the rabbithole of writing todo Lists of what I want to get done for the day, instead of actually getting anything done. -
Embed this notice
Kate (k4t3@comp.lain.la)'s status on Thursday, 09-Jan-2025 07:56:16 JST Kate
@meso what, the part that there are mtfs who are exclusively into other mtfs? -
Embed this notice
Kate (k4t3@comp.lain.la)'s status on Thursday, 09-Jan-2025 07:37:37 JST Kate
@meso no like, transbian whos exclusively into other mtfs. -
Embed this notice
Kate (k4t3@comp.lain.la)'s status on Thursday, 09-Jan-2025 07:35:31 JST Kate
@meso I once overheard a transbian complaining about lesbians in public, that was a weird experience. -
Embed this notice
Kate (k4t3@comp.lain.la)'s status on Thursday, 09-Jan-2025 06:50:59 JST Kate
@meso s/johnny test/mlp/ -
Embed this notice
Kate (k4t3@comp.lain.la)'s status on Thursday, 09-Jan-2025 06:34:01 JST Kate
@meso public display of affection?? in my AMERICAN cartoon????
no wonder all the kids are gay nowadays.. -
Embed this notice
Kate (k4t3@comp.lain.la)'s status on Thursday, 09-Jan-2025 01:59:59 JST Kate
@lain Herr Geruch -
Embed this notice
Kate (k4t3@comp.lain.la)'s status on Wednesday, 08-Jan-2025 21:44:45 JST Kate
need -
Embed this notice
Kate (k4t3@comp.lain.la)'s status on Tuesday, 07-Jan-2025 21:47:48 JST Kate
11 hrs of sleep, 13 hrs or something, of rotting in bed.. idk -
Embed this notice
Kate (k4t3@comp.lain.la)'s status on Sunday, 05-Jan-2025 20:02:11 JST Kate
@fence You can set the animation speed to instant in the KDE settings which gets rid of almost all of the obnoxious animations, would highly reccomend this. -
Embed this notice
Kate (k4t3@comp.lain.la)'s status on Friday, 03-Jan-2025 18:15:06 JST Kate
The 5 genders. -
Embed this notice
Kate (k4t3@comp.lain.la)'s status on Wednesday, 01-Jan-2025 09:08:15 JST Kate
<insert obligatory phrase that people say every 365 days> -
Embed this notice
Kate (k4t3@comp.lain.la)'s status on Sunday, 22-Dec-2024 10:19:42 JST Kate
@arcana @anemone @georgia @lucy LLMs are incapable of thought and reasoning, I feel like people often forget that all they do is autocomplete text.
There are inherent limits to how 'good' outputs, that aren't close-ish to some training data, can be, and it shows. LLMs can't do maths, Tasking an LLM to generate vertices for a 3d model that you describe in text would just go wrong, or even simpler, LLMs can't really do ascii art.
And even when they should be close to something in the training set, it can go terribly wrong if it's nothing that has been reiterated times and times again.
I tried using copilot when i wrote a small helper library to encode x86 instructions, and it was the biggest nightmare, the code it generated looked reasonable, but it was riddled with small bugs that took me more time to debug than if i had just wrote it myself.
If someone's job was receiving instructions such as in files [list of files], "center all divs with the id 'foo'" and fulfilling that, they'd be in danger, but that's not what software engineering looks like.
Real AGI could eventually replace engineers and reshape things in a way you describe, but AGI is still pretty far away, LLMs are certainly not intelligent.