Craig St. Jean

Buying domains (especially in bulk) just so you can park it and sell it for an absurd amount should be illegal

04.08.2025 15:30 — 👍 0    🔁 0    💬 0    📌 0

Some stats:

Ollama with Llama3.2 3B Q4_K_M, M1 got 66t/s, M4 got 116t/s, the 7900 XTX got 119t/s.

MLX Q4_K_M, M1 got 104t/s, M4 got 192t/s

llama.cpp Q4_0 for the 7900 XTX with ROCm got 109t/s, Vulkan got 112t/s
Q8_0 with ROCm got 105t/s, Vulkan got 103t/s

13.03.2025 01:27 — 👍 0    🔁 0    💬 0    📌 0

Sadly, the experiment did not work. With AMD=1, Exo loaded properly but wouldn't respond to any prompts. Also tried the amdfix branch.

When running with the M4 Max (128GB) and the M1 Max Studio (32GB), it ran, kinda, but the M1 kept disconnecting. When it did work, it was extremely slow.

13.03.2025 00:14 — 👍 0    🔁 0    💬 0    📌 0

Trying out Exo for an LLM cluster, funny comparing my AMD RX 7900 XTX (24GB VRAM) which shows as 106TFLOPS to my M4 Max (128GB RAM) which shows as 36TFLOPS

12.03.2025 22:21 — 👍 0    🔁 0    💬 2    📌 0

I love this

13.02.2025 05:41 — 👍 42477    🔁 7410    💬 3156    📌 561

Hot take, if you have a YouTube video or a training video of some kind and the voice is a robot voice, I’m not watching it. Even if it’s content I want to see. I don’t mind if your English isn’t great, I’ll take it any day over robot voice.

10.02.2025 14:00 — 👍 0    🔁 0    💬 0    📌 0

Finally tried #Noita and now I can’t stop playing it

06.01.2025 23:18 — 👍 0    🔁 0    💬 0    📌 0

Why is google like this?

01.11.2024 17:38 — 👍 2813    🔁 86    💬 111    📌 13