TECH & Machine Learning

Joined June 2015
348 Photos and videos
GGs
1
1
5
prime.dev © retweeted
NVIDIA GH200 Module🥵🥵🥵 --- I didn't expect Grace to be so much bigger than I thought. Even though I saw pictures online all the time, it wasn't until I held it in my hands that I realized how huge it really is.
13
28
283
16,399
prime.dev © retweeted
WTF ...
179
1,552
39,768
625,845
😭 the better do rework
11
prime.dev © retweeted
CPU CPU CPU
6
10
145
7,446
ほうらここにも魔王ヨ-カンがいるぞぅ? TP=4はいいぞぅ?
2
1
8
531
😭😭😭...whats going on
4
A new era of PC. 25.0528, 121.5990
1
37
"AI" word sounds lazy when being used... but hearing an engineer using "Machine learning" in his talks sounds like a retired C, C etc.. developer
31
prime.dev © retweeted
Here’s a simplified illustration adapted for your level of understanding, hopefully this makes it easier for you to follow
42
233
1,951
34,202
prime.dev © retweeted
who are these so-called "stakeholders" you claim you are in discussion with???
2
2
726
HON. Hasn't finished handling NITA bill issue... noorr this come..ei.. data price reduction Korea 🥀🥀
🇬🇭 JUST IN: Effective June 1, 2026, transfers from MoMo wallets to bank accounts will attract a 0.75% fee per transaction, capped at GHS 5.
80
prime.dev © retweeted
Students who got Google AI pro free for One year. It's time to cancel the Autopay😃
137
122
5,471
730,811
I Forgot to Push the CSS files bro..🤦
The worst bios interface award goes to
52
prime.dev © retweeted
Invest in home lab, you won’t regret.
191
580
7,210
198,421
prime.dev © retweeted
You see what I was talking about last time? Cricket Wireless is using AT&T towers to offer cheaper plans to customers who want to spend less, even though both companies are owned by the same parent. So why can’t the government allow MTN to upgrade its equipment to higher speeds and lease its nationwide infrastructure to AirtelTigo while still prioritizing MTN users, just like AT&T does?
32
34
240
19,819
Notice a _Human action is required before AI Works....and if it's Automated you need a Human to Supervise...
1
18
One rtx Pro 6000 Blackwell, R9 9950X, 32GB DDR5 CL28 8000MTS, 2TB NVME 5.0, and 4K amoled monitor should be the highest to enjoy any open-source model (Flux 2 DEV,Qwen-Image & Z-Image , LTX, ... Deepseek,Qwen,Llama,Gemma, CHATGPT, etc..) personally and game when bored....
1
170
...you don't need any 128 GB RAM, two GPUs, etc.. Unified memories aint fast for high compute intense Models... Mac's ain't reliable for AI btw.._because if they were...Apple will have their own AI..that is why most people run quantized versions of the original models...
21