matrik

6 ms·

1.
by matrik 2mo ago
For the same performance, one could even go about 50% cheaper with 16 channel ddr4 + a rtx3090 for prompt processing. But still, even for mid level projects API is orders of magnitude cheaper, since you don't need to set it up and main
2.
by matrik 3mo ago
And why should one prefer GitHub Copilot over OpenCode? Worse harness, more expensive prices, unreliable product strategy, limited model support, the list goes on.
3.
by matrik 5mo ago
I'm not sure this information is grounded, but I remember to have read somewhere the inference is indeed profitable. My personal experience is similar. Running 2x3090s draw 500-600W and you can locally run amazing models with a similar
4.
by matrik 5mo ago
Claude.ai uptime 89.2%. That's a stunning zero 9s.
5.
by matrik 8mo ago
At this scale, some bugs become features. I think fixing the bugs would need lots of conceptual work, due to the fact that there are millions of users of these apps.
6.
by matrik 1y ago
I find LLMs very efficient in a lot of things except writing big chunks of code that require some organization. They’re good at closing knowlwdge gaps, finding strategies to solve problems, writing code for defined scopes and even reviewing
7.
by matrik 2y ago
MobileNetV3 and EfficientDet are othwr possible alternatives to YOLO. I was able to get higher than 1.5 FPS on Raspberry Pi Zero 2W which draws 1W on average. With efficient queuing approach, one can eliminate all bottlenecks.
8.
by matrik 2y ago
what’s the name of the backend provider?
9.
by matrik 4y ago
Just out of curiosity, and haven’t read a similar comment here, why not create another Google account for ads? (I mean this account for ads looks dead and a separate account is unrelated with the new one.)
10.
by matrik 4y ago
It’s not better but more consistent than GCP cloud console.
11.
by matrik 5y ago
there is a way to count unique visitors without cookies and storing ip addresses indeed. have you heard of hyperloglogs? it’s very limited but works.