Nikita Ivanov

[Nikita Ivanov @ViChyavIn]

@xah_lee @_forg__ Bro, you didn't know what undefined behavior was until a few days ago.

[Xah Lee @xah_lee]

@RoseSilicon whats UB? @grok

yes i saw the post, i think negative number input shouldnt be considered as bug.

[Nikita Ivanov @ViChyavIn]

@MaciejDominiak5 @SebAaltonen It's a wrapper around Vulkan and such.

[ɮiʀᴅᴏ @TLEbird]

Tim Curry’s best moment

[Konrad Reczko @reczko_konrad]

Depth-aware light injection in TypeGPU

I got a 448x448 monocular depth model down to ~8 ms on my M4 Pro across ~250 dispatches, which is fast enough to use in realtime :D
Since the inference is written directly in TypeGPU, I can just feed the depth buffer straight into the lighting pass. It never has to leave the GPU or go through any extra synchronization/interop step

Inference, lighting and draw all go through the same command encoder.

[Nikita Ivanov @ViChyavIn]

It goes from 30 seconds to 15 seconds... for a incremental build... Which is not really a success of any kind. And one of the "optimizations" is DISABLING DEBUG INFORMATION SO YOUR DEBUGGER BECOMES USELESS.

[Let's Get Rusty 🦀 @letsgetrusty]

How to decrease your Rust compile times by 50% - https://t.co/VDAsd9ro3P

[Nikita Ivanov @ViChyavIn]

@misterfitzie @_Felipe Would you be willing to loose borrow checker?

[Coobyk @Coobyk_]

GUID Partition Table

[Nikita Ivanov @ViChyavIn]

@jesse_makkonen Not a horror?

[Тsфdiиg @tsoding]

Programmers today: "boo-hoo your language sucks! 😭"
Dijkstra in the past:

[fintex @_yusufknl]

As someone who's been shipping LLMs since the GPT-2 days, this lecture on cross-entropy from a Stanford math grad is the closest thing to an ML PhD qualifying exam I've ever seen released publicly for free.

Everyone thinks language models predict the next word. They don't. They compress language. Once you see the math, you can't unsee it.

33 minutes. Bookmark & watch today.

[Ryan Fleury @rfleury]

The first way people usually learn code compression is by pulling common things out into helper functions, but in my view the complexity of adding a new entry point is often overlooked, and things quickly turn into a mess.

It’s a useful exercise to try compressing code without helper functions, and instead merging similar work by using loops, tables, and reordering. In my experience this leads to dramatically simpler architecture, and it’s a lot easier to keep the whole picture in your head.

[Nikita Ivanov @ViChyavIn]

@SwedPaul "Привет Мир" компилируется минуту :P

[Abakcus @abakcus]

1/3 + 1/9 + 1/27 + ··· = 1/2. The proof is in the picture. No words needed.

[Barji @BarjiYT]

How is this even a debate?

[Arika @Arikalabs]

devs be like "manual memory is unsafe" then run curl | bash from a medium article

[Alex Cohen @anothercohen]

I think about this video at least once a week

[Gabriel Dechichi @gdechichi]

- $15 billion dollar company
- ships entire browser with their application cause "native GUI too hard bro"
- javascript so devs don't have to reason about memory
- leaks memory anyway
- "let's just restart the application when we go above 4 GB"

this is a new rock bottom

[Computer Science @CompSciFact]

The C code below compiles and prints "hello, world".

[Vjekoslav Krajačić @vkrajacic]

Microsoft @Windows: "We're preloading File Explorer so it launches faster."
File Pilot: Compiles in less than 3 seconds. Launches instantly. Eats whole drive for breakfast.

Speed isn't a background task, it's design.
https://t.co/v4OY47UD4h