@thetechitout: This might look like a compact PC, but it’s built for AI at a whole different level. Powered by the Ryzen AI Max with 128GB of unified memory, it shares resources seamlessly across CPU, GPU, and NPU—all working as one system instead of separate parts. 🔥Try my FREE PeakPost, it tells you the exact best times to post on any platform: https://besttimetopost.netlify.app That unified design delivers major gains in speed and efficiency, allowing large AI models to run entirely offline, no cloud, no latency, no data leaving your device. It marks a shift away from relying on distant data centers and toward running powerful AI locally. That’s the promise of edge computing: faster performance, stronger privacy, and full control over your data. For developers and creators, this changes how AI is used day to day, bringing advanced capabilities directly to your desk instead of a remote server. 👉 Like, Comment, Share, Follow: @TheTechItOut
TheTechItOut
Region: US
Tuesday 21 April 2026 14:17:23 GMT
Music
Download
Comments
AiPriest :
The next 10 yrs gonna be wild
2026-04-23 15:36:57
84
Natron308 :
200B params doesn’t impress like it used to.
2026-04-23 20:28:30
50
justthefacts03 :
what's the cost?
2026-04-23 02:52:50
34
Tower Lord :
Reminds me of a 10GB hd from the 90’s. Give it a couple years and we’ll be running AI models on a device the size of a thumb drive.
2026-04-24 00:52:20
22
KOCMOTPOH :
😂 yet Apple of all delivers way better performance and power efficiency.
2026-05-31 06:33:33
1
Mir :
No cuda? Sorry...I am out
2026-05-31 09:02:59
0
Code :
200b parameters at one token per second…
2026-04-24 20:52:31
24
the_nightman :
128GB unified memory can run qwen code and that (from what I hear) is competitive with opus 4.5. 2K is too much money though when you can use opus 4.7 for $20 a month
2026-04-26 01:45:20
2
Maru🌹⚜️🏰 :
Why make it small when it's completely unaffordable to the average person?
2026-04-23 09:41:21
0
syzguru :
instead of APU call it unified memory, but its getting not faster then
90GB/s is fast, 1900GB/s on a gpu is faster
2026-05-30 23:00:02
0
ClipsGlitched :
did nvidia just quietly destroy all gaming pcs?
2026-04-23 17:14:11
4
wendyinfutures :
Things like this - and continuing innovations in computing hardware - is exactly why the giant data centres being built everywhere are likely stranded assets before they are even complete.
2026-05-31 09:35:45
3
Nunya :
Wake me up when it's native to my phone 🥱
2026-04-23 22:36:43
7
bendrinking :
nothing but a spark
2026-04-23 10:56:02
5
ale :
como el DGX Spark de nvidia ?
2026-05-31 02:08:44
0
AISweeties :
it can LOAD the models, "run" might be a bit subjective depending of what you consider acceptable tokens per second
2026-04-24 00:49:00
5
lueybatchiluey :
How much? Can buy a Mac mini for $499 🤔
2026-06-07 12:50:57
0
Charly Pro Account :
👍 next level🚀🚀🚀🔥🔥🔥😎😎😎
2026-06-07 07:28:38
1
WheelzUp :
Nvidia has the DGX Spark, which I assume competing with this device.
2026-04-23 19:16:10
3
Slooth :
Don’t mwan it’s gonna be fast. Mine can do 200b. But it might take an hour.
2026-04-24 12:11:47
3
Joseph Simkoko :
Can I donate that machine and how do I get it?
2026-05-31 06:28:43
1
rest less :
Big AI companies are gonna go bankrupt soon. Good!
2026-05-31 01:39:07
1
Rektor :
How does it compare to a Mac mini beefed up?
2026-04-24 23:29:37
2
Time To Wake up :
It's important because those guys who are using subscriptions in 18 months won't be able to pay the monthly $500 fee. 😳 That's why.
2026-04-25 00:11:47
2
🇳🇮 PapaGus 🇺🇸 :
“That’s the internet, Jen.”
2026-04-25 01:01:15
1
To see more videos from user @thetechitout, please go to the Tikwm
homepage.