Dhavan

The latest 𝕏 algorithm has been published to GitHub github.com/xai-org/x-algo…

With apps like Superwhisper and Wispr Flow, imagine having a subscription for your keyboard. You want to backspace many words? Sure! Ten bucks!

I have been playing too much with different models. It is now confusing and yet illuminating how different models behave for different uses

looks like I won't have to pay for commit messages at least :P

Revived my old 2080ti, loaded Gemma 3 12B with llama.cpp. This is where we are at: prompt eval time = 76.59 ms / 17 tokens ( 4.51 ms per token, 221.96 tokens per second) eval time = 32043.16 ms / 1598 tokens ( 20.05 ms per token, 49.87 tokens per second) total time = 32119.75 ms / 1615 tokens

@_HooJR @antirez Totally! Great stuff here

@antirez ur killing it man 🔥

If the correctness tests will pass, expect a sensibly faster DS4 inference speed for DGX Spark, and especially a lot flatter prefill as context increases. Soon in the repo if everything goes as expected.

@Sentdex GM!

Goooooood morning!

Done with things for now

It cost me ~US$1 on OpenRouter to fetch latest from meshtastic upstream, compile and flash. GLM-5.1. it made no errors which was great.

@__tinygrad__ @APompliano I would LOVE this!

@APompliano We need stacks of GPUs in every house, not really big stacks of GPUs controlled by companies who are trying to extract value from us.

We need more data centers, not less.

hot take: unrestricted social media algorithms are as dangerous as weapons of mass destruction you can literally reprogram the minds of billions of humans

@GaryMarcus I've recently started using it for all general searches, automating basic tasks. Experimenting with code as well, but not as much as codex / claude yet.

Who out there is using Hermes, and for what?

@ksg91 @amazonIN 💀

I think @amazonIN thinks I am Ambani!

DS4 running on DGX Spark (GB10 / CUDA), private branch for now. 12 tokens/sec, the memory bandwidth is limited in this system, at 270GB/sec. But prefill is ways more alighed to M3 Max at ~200 t/s. I'll release when more mature, but it is almost sure that it will get merged.