0xKitkat (@procrastiness) header
0xKitkat (@procrastiness) profile picture

AI ML| surviving @ntusg

singapore
Joined February 2026
Photo by 0xKitkat (@procrastiness)Photo by 0xKitkat (@procrastiness)Photo by 0xKitkat (@procrastiness)Photo by 0xKitkat (@procrastiness)Photo by 0xKitkat (@procrastiness)Photo by 0xKitkat (@procrastiness)Photo by 0xKitkat (@procrastiness)Photo by 0xKitkat (@procrastiness)Photo by 0xKitkat (@procrastiness)Photo by 0xKitkat (@procrastiness)
100 Token per second on 4x v100 with deepseek v4.1 flash.
Yeah 40 token per second on a seven hundred billion parameter large language model at q8 yes fuck the guardrail im the best
1
80
on 3k of hardware btw
1
16
Yeah 40 token per second on a seven hundred billion parameter large language model at q8 yes fuck the guardrail im the best
1
1
1
95
2/ Train yourself to go from 10 to 40 to 100 hours of focused work a week. I did this for months building an automated trading system from scratch. 5am-10pm 7days/week 99% of people need #WorkLifeBalance or whatever, and that's great. But I'm talking to the people on a mission.
34
411
79
3,522
286,431
Anyone tried these mods? worth the risk or not?
23
I named my ubuntu server "Triologay" and Opus would auto correct to Triology everytime. Very funny😂😂😂
24
What to run on 16gb of ram?
25
Will release an uncensored version of this within 3 days. Will be post trained with cyber tool use. Very good for pen testing and ethical hacking task. Also from now on i will ditch GGUF quant and only stick to EXLlama.
MiniMax M3.1 Flash weights are now public. 428B parameters total, 23B activated per token. Agentic performance: A generational leap in agentic tasks, coding, and cybersecurity — 76.2 on BankerToolBench, 73.7% on DeepSWE v1.1, 61.7 on OSWorld V2, 29.6 on Agents' Last Exam, and 82.2 on CyberGym. Efficient by design: MiniMax Sparse Attention and DSpark speculative decoding improve inference efficiency. Hugging Face: huggingface.co/MiniMaxAI/Min… ModelScope: modelscope.cn/models/MiniMax… API: platform.minimax.io M Plan: minimax.io/m-planMiniMax MCode for desktop: agent.minimax.io/download
46
In Claude We Trust
3
60
Well… I gotta say i was hella early.
We kept our commitment to bring you the Arduino VENTUNO Q at an introductory price of under $300 for our launch. Thanks to your incredible support, the board sold out twice! Due to rising component costs, VENTUNO Q moves to its official retail price of $399 starting today, October 19th, though continues to deliver a powerful computer for Physical AI and robotics at an unprecedented value. Read all the details in our CPO Marcello Majonchi's blog post and grab yours now: blog.arduino.cc/2026/10/19/a…
1
54
DEEPSEEK now charges $0.75 per 1000 websearch request.
2
134
Oct 2026 Dorm room lab update.
3
61
welcome home lil guys
2
46
Thank you @0xSero for the amazing references on your hf. Helped me a lot while setting this up!!!
25
Qwen 3.8 Flash Next say hello world from the 3x V100 + 256GB DDR4 rig!
1
3
196
0xKitkat retweeted
Success is not certain, but excitement is
1,179
1,604
219
21,039
1,019,812
Fking hell they quantized my boy OPUS 5.5😭😭😭😭
3
48
v100 32gb testing.
45
v100 32gb testing
33