ry (@ry22aning) header
ry (@ry22aning) profile picture

ry

@ry22aning

17 yo || changing bits while understanding electron spins

 qbit |ψ⟩
Joined March 2023
  • Tweets1.6K
  • Following479
  • Followers697
197Media
Photo by ry (@ry22aning)Photo by ry (@ry22aning)Photo by ry (@ry22aning)Photo by ry (@ry22aning)Photo by ry (@ry22aning)Photo by ry (@ry22aning)Photo by ry (@ry22aning)Photo by ry (@ry22aning)Photo by ry (@ry22aning)
a theory, which says consciousness comes from the interaction between our brain and zero point field (omnipresent energy field, present in the fabric of space) In this model the brains acts as a receiver not a generator. crazy
Image from ry (@ry22aning)
2
32
ry retweeted
ry (@ry22aning) profile picture
ry@ry22aningOct 28
Replying to @cneuralnetwork
@cneuralnetwork
Image from ry (@ry22aning)
1
1
46
2,816
yo i thought the new ceo was a hardware guy but wtf is he doin 12 gb locked to 8gb 120Hz panel locked to 60Hz
1
3
210
the goal was the idea. not a product
ry (@ry22aning) profile picturery@ry22aningOct 23
china i love you
2
101
china i love you
1
2
176
dot file hell hole
1
96
keep the session open and be there 4 each other
4
99
How it feels to self-host a local LLM, running entirely in CPU on a Raspberry Pi 4.
107
34
10
1,319
103,736
Image from ry (@ry22aning)
4
112
youtu.be/QCVG8Rrms5s?si=NVz_…
I made a GPU in 24 Hours

I made a GPU in 24 Hours

Have you ever thought about designing your own GPU? I did to win (2...

youtube.com
1
5
265
gng if I don't end up doing shi like this, idk what will happen to me
1
43
bro i miss 2024, was shitposting all the time. anything i was getting my head into, what an idiot. gotta do it more
1
4
121
made first 15 chips using nand gate was pretty fun
3
102
Accelerator companies focusing on inference and tok/s at the expense of fundamentals is a serious mistake. LLMs are just a small part of all the things that will be built with AI in the future.
15
8
3
344
16,376
we don't really worry about the "how" bc it's enough to worry about the "what"
3
97
ry retweeted
ry (@ry22aning) profile picture
ry@ry22aningOct 14
5.6tok/s, alr I just wanna play around. Just the idea that i'm able to run it locally feels thrilling!
Image from ry (@ry22aning)
1
1
2
53
recently got a new laptop, i could finally use my old potato calculator to play now around. gonna make this rock think. THINK MONKEY THINK. i've 2gb of vram, w 8gb of ram, 117ssd. so would have to offload many layers before my gpu screams.
1
5
196
model: Qwen2.5 3B initially thought I'd use airllm, I've some old nvidia gpu (later found wasn't configured for cuda), so am using only cpu inference. But it took around 0.073tok/s 🥀 (due to layer by layer streaming) Tried it with ollama, 5.631tok/s
Image from ry (@ry22aning)
1
1
99
5.6tok/s, alr I just wanna play around. Just the idea that i'm able to run it locally feels thrilling!
Image from ry (@ry22aning)
1
1
2
53
revisiting de goat
Image from ry (@ry22aning)
4
94
Load more