uaruss

Lifelong learner. Safety is an illusion. If you want to be great, you don't have choices. It takes what it takes.

[Mario Nawfal @MarioNawfal]

Japan's national football team vs 100 kids

That goal line is insane 😂

Writer: Ian

[El Programador Senior @5eniorDeveloper]

Esta es la mejor manera de ver el mundial!

[uaruss @uaruss]

@edgecase411 @houston_flood @CamCutApp Computer controlled car didn't run into anyone. Human controlled car failed to manage the situation. Pretty clear which is safer.

[uaruss @uaruss]

@LibertyRushmore @newstart_2024 Not knocking it at all, actually intrigued!

[uaruss @uaruss]

@Combat_learjet Blocked has been used for a long time. When I hear it, I shut up and wait for ATC. Works. Simple.

[uaruss @uaruss]

@LibertyRushmore @newstart_2024 Sardines and water for 5 days? Crackers?

[uaruss @uaruss]

@elonmusk Looks like Starship landing on the right!

[uaruss @uaruss]

@FluentInQuality Who wouldn't want to work for this man.

[uaruss @uaruss]

@TravisMcDonald @0xSero Any chance you rna the 3.5 version of that previously? If so, how do thy compare?

[uaruss @uaruss]

@TrungTPhan This is the key ->"our feeling is that the day that somebody working in Apple decides that they can't make a difference anymore, is the day we've lost"

[uaruss @uaruss]

@sudoingX Awesome! I agree w/ you on opening for AMD. The hardware in the R9700 is a steal, just need a few more months for the software to really crank!

[uaruss @uaruss]

@sudoingX @xai Anyone had success with Grok 4.1 fast reasoning and Hermes? It was pretty much a disaster when I tried to run it recently (significantly worse that local qwen models w/ hermes)

[uaruss @uaruss]

@freeroll88 Miss your posts. Hope things are well!

[uaruss @uaruss]

@sudoingX @ItsTomWoolley Are you manually switching models or soemthing else? I'm using one of the Carnice qwen models (64 GB VRAM), but still too often default to basic Claude code or Claude, and ignore hermes, for critical work.

[uaruss @uaruss]

@freeroll88 that had to be about as textbook as they get!

[uaruss @uaruss]

#AMD R9700 cheats: (1) Fan issue: See RDNA3 gpu_od/fan_ctrl sysfs interface (2) Power limits: Undervolt by 0.75mV and set power limit to 315W (3) For MoE on multiple cards, do this: https://t.co/BvqX9hpPEM

[uaruss @uaruss]

@sudoingX I just got Carnice-MoE-35B-A3B-Q6_K running on my dual AMD R9700s, llama.cpp, ubuntu. So far, great experience with Hermes! The dense model would be interesting!

[uaruss @uaruss]

Thanks! I'm running this on dual AMD R9700 GPUs on llama.cpp with the config below (if it's helpful for others). So far, seems much better than than non-carnice version!)

-m /home/***/models/Carnice-MoE-35B-A3B-Q6_K.gguf \
-ngl 999 \
--tensor-split 1,1 \
-fa on \
--cache-type-k q8_0 \
--cache-type-v q8_0 \
-c 131072 \
--port 8081 \
--host 0.0.0.0 \
-np 1 \
-t 8 \
--temp 0.6 \
--top-p 0.95 \
--top-k 20 \
--min-p 0.0 \
--presence-penalty 1.5 \
--repeat-penalty 1.0 \
--jinja \
--chat-template-kwargs '{"enable_thinking": false}' \
--reasoning-format deepseek

[uaruss @uaruss]

@gophillionaire @Teknium @Google @openclaw @NousResearch Ha! I've got it running on dual AMD R9700s on llama.cpp. Had to tweak a llama.cpp bug (submitted a PR), but it's been stable for ~2 weeks since.

[uaruss @uaruss]

@Teknium @gophillionaire @Google @openclaw @NousResearch Why not an MoE model (e.g. Qwen3.5-35B-A3B)?