aaulia
Re:Running Gemma 4 26B at 5 tokens/SEC on a 13-year-old Xeon with no GPU
on Thursday July 16, 2026 @02:47AM (#48931118)
I tried Qwen MoE a while back. Using my 8GB RX470, somehow got 10+ token/sec, lot's of trial and error with llama.cpp config, and it's still slow to be used for my usecase. Even at 12 to 16 IMO it's slow. For chat, maybe it's enough, but for any other tasks, it's not viable IMO
Re:I think I have LLM burnout
on Thursday July 09, 2026 @12:11AM (#48840889)
LLM skew the time estimate tho. Now everybody expect stuff based on LLM work instead of normal human work. I/we can choose to solve problem normally, but the expectations have changed.
Re:The Doorman's Fallacy in Action
on Thursday June 25, 2026 @08:19PM (#48680923)
At first I thought you meant QR for payment, which is weird because most people (at least in SEA, where I lived) consider less friction and more convenient than cash or cards.
But it turns out you meant QR for menu, yes, hate them. Flipping through the menu is better experience overall. Opening menu on a phone is a chore, not to mention most menu web/app is crap. Lots of them are just link to pdf on google drive.
Re:The Token Compression Illusion: Why I'm Skeptical of RTK
on Friday June 19, 2026 @02:54AM (#48595614)
Tried it just now. The onboarding process could be better, for example guide user to pick the available models if providers is setup but it's not anthropic. Wasted a bit of time foguring out that the provider is detected, just he models was wrong.
But I like it, code processing is freaking fast.
Re:When Did White-Collar Work Start to Look So Bleak?
on Monday June 15, 2026 @10:26PM (#48549800)
Same, which is why I just go with my own pace and stop being FOMO. Fortunately my org is not that gung-ho about pushing AI.
Re:Doing Something That's Never Been Done Before
on Monday June 08, 2026 @08:01PM (#48454206)
Emulator?