David Lev Wilson

the 6000 pro will give you considerably better inference speed, but you won't be able to run nearly as large of models. What I've been asking myself, also trying to make this decision: would you rather sit there and wait a little longer for shit to generate, or get billions more parameters?

@0xSero But the real question is 2 or 3 of these or a 6000 pro?

Elite device. Stacks amazing Low power High memory Decent price

@picocreator all of the doorways would have to go up to the ceiling

Surprised we have yet to see a robotics startup pitching a household rail system, which skips all the current hardware problems of batteries and legs This lets it focus on the MVP, doing chores (in house), and hands.

@bryan_johnson Challenging yourself, feeding yourself, growing yourself, crushing your enemies’ hopes and dreams.

“So you spent $40K to run GLM 5.2 locally?” “Yes, Dave” “And you did all this to save $20 a month on Cursor” “Thats right, Dave”

@mike64_t Is this still mostly true on architectures that have more unified memory between the CPU and GPU?

Everybody should read this because it tells you how opinionated the GPU is about how exactly it wants work to be run, and taken to its logical conclusion it implies that e.g. your CPU-offloaded optimizer is not at liberty to run eagerly, but sit behind a message pump waiting for QMD sem-sets. It's perversely the *GPU* driving the *CPU*, not the other way around. The CPU has dispatched work statically, yielded control to the GPU and then aiding to preserve the CUDA graph dependency model facade as best as possible. cuGraphAddBatchMemOpNode with WRITE/WAIT_VALUE is really all you need here from an api perspective.

Neural CAs are amazing, but they've never scaled past low resolution. We propose a simple solution that allows an ~8x resolution boost with minimal extra parameters. The core idea: Treat cells as local neural fields instead of pixels. Try the demo: cells2pixels.github.io 🧵

🇺🇸

Happy Birthday America, Land of the Free. Don’t give up on the First, Second, Fourth, and Ninth Amendments. They’re unique, and your freedom - and the worlds’ - depends on them.

Hi! I’m Dave. Follow me along my journey of making music and learning AI