Pardox (@pardoxAI) header
Pardox (@pardoxAI) profile picture

GenAI/LLM addicted,clusters of gpus Thinking hard about the future

Joined March 2026
Photo by Pardox (@pardoxAI)Photo by Pardox (@pardoxAI)Photo by Pardox (@pardoxAI)Photo by Pardox (@pardoxAI)Photo by Pardox (@pardoxAI)Photo by Pardox (@pardoxAI)Photo by Pardox (@pardoxAI)Photo by Pardox (@pardoxAI)Photo by Pardox (@pardoxAI)Photo by Pardox (@pardoxAI)
Replying to @uctive
When The multi-stream aggregate numbers will be higher than the pure single-stream decode speeds?
1
1
138
Replying to @Yarilo7brigada
stop hook me with the term of "f*cking insane:"This is f*cking insane:
97
Replying to @uctive @coldniko
What hardware ?
1
1
17
the 4 bit will work also ?
1
337
Wait what. that free ?
1
214
Replying to @Morrowmake
Amazing job man!
34
Replying to @ShenSeanChen
Love your explanations
1
1
68
Replying to @Mr_Buzzoni
that slop AI ?
81
Replying to @GauriTripa94282
what the best repo to use for each one of them do you can share?
32
Replying to @nash_su
i think web app will be amazing !!
Replying to @TheAhmadOsman
You mean China models
1
85
Pardox retweeted
Jensen Huang in the Matrix, learning how Chinese spies work Wait till the end 👀
🤖 Made with AI
1,807
15,650
2,214
88,808
8,314,979
Replying to @witcheer
you using llm-wiki ? if yes you can share more about how you use it ?
1
1
19
Replying to @TheAhmadOsman
youtube video soon ?
1
1
56
Replying to @SteezeyCS2
There a lot of big AI that like local AI
1
9
Replying to @elder_plinius
But how you verify its true ?
1
130
Replying to @superalesha
I'm not surprised at all. Good Chinese models are on the way.
8
392
Replying to @superalesha
Obviously, they will do it. They don't want to see local AI evolve, that's the biggest threat to them.
1
2
1
106
4,859