Entries for July 1, 2026

@onusoz · /2026/07/01 · 06:13 PM View on

(I am trying to run it on DGX spark)

@onusoz · /2026/07/01 · 04:59 PM View on

Is anyone able to run nvidia/Qwen3.6-35B-A3B-NVFP4 with the config suggested in the readme? It OOMs before it can start serving huggingface.co/nvidia/Qwen3.6…

Image hidden

@onusoz · /2026/07/01 · 12:31 PM View on

I know that my macbook local model benchmarks have started when my lap catches on fire Add to this list: "does not set the room on fire"

@onusoz · Jul 1, 2026

I keep seeing insanely expensive builds giving insanely impressive results These results don't matter What matters is, whether one can: - run a "SOTA level" model, whatever that is - with under 32gb VRAM or unified memory - in 5 parallel sessions - with 50~100 tok/s each - with enough leeway memory for other applications - in a system as cheap as $1000 That is our goalpost That is the threshold when open source AI will win

@onusoz · /2026/07/01 · 12:27 PM View on

I can't believe I'm asking GPT to use Claude to review It's almost as if there is a 9-month cornercutting cycle, a two-body problem between openai and anthropic

@onusoz · /2026/07/01 · 12:17 PM View on

I quite like this token speed simulator by @mikeveerman And I keep losing it, so hopefully I will remember to come back to this tweet:) Link: mikeveerman.github.io/tokenspeed

@onusoz · /2026/07/01 · 10:48 AM View on

I keep seeing insanely expensive builds giving insanely impressive results These results don't matter What matters is, whether one can: - run a "SOTA level" model, whatever that is - with under 32gb VRAM or unified memory - in 5 parallel sessions - with 50~100 tok/s each - with enough leeway memory for other applications - in a system as cheap as $1000 That is our goalpost That is the threshold when open source AI will win

@LottoLabs · Jun 30, 2026

That’s actually crazy A new frontier is found imo localmaxxing.com/en/runs/cmqznf…