No, no, no. Its a cpu, not a gpu! Its a cpu with tensor cores! Not the same at all!
/slight sarcasm (because its not the same, but at the same time fuck jensen and his bullshit)
Hey man, it's not even for you it's for the agents that will embody your soul and live your life for you. Think of the agents bro, there's nothing they can't do. The can read your email and use that to train other agents. Think of that! Agents running all the time doing everything for you bro. Like reading email, and uh.......sending email......and like automatically deleting some emails. It's not just default email behavior bro it's AI agents man it's totally different. Check out this code they wrote by looking at email threads. Agents bro they need this chip without it the agents starve.
A combination of a microprocessor and a graphics chip, developed with help from Taiwan’s MediaTek, it is designed to run AI agents locally rather than relying on cloud computing.
From a privacy perspective, at least, this has potential.
It will allow agents to navigate PCs autonomously, replacing humans’ traditional mouse and keyboard interactions.
Yeah, no. These things are still far too unreliable. Anyway, if you look at most sci-fi set in the future with voice control, keyboards (or at least their touchscreen counterparts) are still very much present.
The second part is wholly software dependent, so let's not conflate the two.
Having local hardware for local LLM (and other models too! there's plethora use cases for AI models, e.g. easily tagging people in your photo library, automatic subtitles for videos, even realtime stuff, we could even have models that automatically categorise photos and sort them into albums based on previous patterns, and so on) is awesome. Not having to trust some random third party with your data is awesome.
Blending that in with a specially written agent that can interact with stuff is not awesome. The two should be separate, but problem is, most users won't understand the benefit of this hardware without being given concrete examples of use cases like this.
I'm staunchly anti-genAI but consider other applications of traditional machine-learning less problematic. A purely local model for photo categorization seems, in theory, less objectionable to me. I'm sure models exist already but I'm purposefully out of the loop. Any suggestions for models I could look into? And just how much compute would something like that require?
SoCs like this will definitely be an improvement over power guzzling discrete GPUs. Not sure what kind of crack they’re smoking marketing LLMs as a replacement for a keyboard and mouse. I’m interested in self-hosting LLMs for agentic coding and this looks like a good fit for that, but I won’t even be bothering with OpenClaw, and proprietary Micro$lop is garbage is absolutely out of the question.
This reminds me of bitcoin so much. At first there was a rush for cpu then gpu then specific miners. Then specialty hardware. We are here now. So silly.
It's gonna be so fucking funny when the push to sell silicon that can run local models at 100 watts or less ends up destroying the business models of the companies that built out 100,000,000,000 watts of data centers.
30 Comments
Cyv_@lemmy.blahaj.zone · 98 pts · 80d
"It's a revolution, it's the best thing ever, it's magically the best at AI, it's a super chip!"
It's a mobile GPU. We have a lot of those Jensen, keep it in your pants for fucks sake.
DacoTaco@lemmy.world · 8 pts · 79d
No, no, no. Its a cpu, not a gpu! Its a cpu with tensor cores! Not the same at all!
/slight sarcasm (because its not the same, but at the same time fuck jensen and his bullshit)
ButtermilkBiscuit@feddit.nl · 2 pts · 79d
Hey man, it's not even for you it's for the agents that will embody your soul and live your life for you. Think of the agents bro, there's nothing they can't do. The can read your email and use that to train other agents. Think of that! Agents running all the time doing everything for you bro. Like reading email, and uh.......sending email......and like automatically deleting some emails. It's not just default email behavior bro it's AI agents man it's totally different. Check out this code they wrote by looking at email threads. Agents bro they need this chip without it the agents starve.
Bonesince1997@lemmy.world · 93 pts · 80d
No thanks.
yakko@feddit.uk · 46 pts · 80d
Every corporation for some reason: Nah you gotta
zewm@lemmy.world · 58 pts · 80d
Yaky@slrpnk.net · 44 pts · 80d
Isaac Asimov's future of angry men yelling at uncooperative robots is almost here.
devaly@ani.social · 40 pts · 80d
imagine all the people screaming at their computers at the office:
ABORT ABORT CANCEL YOU STUPID FUCK
LMAO
myrmidex@belgae.social · 28 pts · 80d
If only Huang's mom would've yelled that some 63 years ago.
isthereanydeal@discuss.tchncs.de · 2 pts · 80d
Chuckles upvote for you
pooterbroo@programming.dev · 1 pts · 80d
Australis13@fedia.io · 33 pts · 80d
From a privacy perspective, at least, this has potential.
Yeah, no. These things are still far too unreliable. Anyway, if you look at most sci-fi set in the future with voice control, keyboards (or at least their touchscreen counterparts) are still very much present.
gravitas_deficiency@sh.itjust.works · 20 pts · 80d
We have this already. They’re called APUs.
fonix232@fedia.io · 11 pts · 80d
The second part is wholly software dependent, so let's not conflate the two.
Having local hardware for local LLM (and other models too! there's plethora use cases for AI models, e.g. easily tagging people in your photo library, automatic subtitles for videos, even realtime stuff, we could even have models that automatically categorise photos and sort them into albums based on previous patterns, and so on) is awesome. Not having to trust some random third party with your data is awesome.
Blending that in with a specially written agent that can interact with stuff is not awesome. The two should be separate, but problem is, most users won't understand the benefit of this hardware without being given concrete examples of use cases like this.
UniversalBasicJustice@lemmy.blahaj.zone · 6 pts · 80d
I'm staunchly anti-genAI but consider other applications of traditional machine-learning less problematic. A purely local model for photo categorization seems, in theory, less objectionable to me. I'm sure models exist already but I'm purposefully out of the loop. Any suggestions for models I could look into? And just how much compute would something like that require?
zebidiah@lemmy.ca · 5 pts · 79d
Immich has a built in machine learning algorithms groups faces and runs on just about anything, including various raspberry pi
jabjoe@feddit.uk · 27 pts · 80d
The return of the fat client in the age of AI mainframes? Nvidia is selling shovels in the AI gold rush, so they don't care either way.
nbsp@programming.dev · 12 pts · 80d
Mainline linux drivers or gtfo
damnthefilibuster@lemmy.world · 10 pts · 80d
“… for Windows”? No thanks.
brokenwing@discuss.tchncs.de · 9 pts · 80d
dave2d said it better. Having agents in your laptop doing your work for you is revolutionary, but trusting Microsoft with it is the problem.
badgermurphy@lemmy.world · 8 pts · 80d
QuantumSpecter@lemmy.world · 8 pts · 80d
I don't think the industry realize the majority of AI output is youtube slop videos
zebidiah@lemmy.ca · -4 pts · 79d
I don't think society realizes the majority of AI usefulness is not in generating slop pics and video...
melfie@lemmy.zip · 6 pts · 80d
SoCs like this will definitely be an improvement over power guzzling discrete GPUs. Not sure what kind of crack they’re smoking marketing LLMs as a replacement for a keyboard and mouse. I’m interested in self-hosting LLMs for agentic coding and this looks like a good fit for that, but I won’t even be bothering with OpenClaw, and proprietary Micro$lop is garbage is absolutely out of the question.
mesamunefire@piefed.social · 5 pts · 80d
This reminds me of bitcoin so much. At first there was a rush for cpu then gpu then specific miners. Then specialty hardware. We are here now. So silly.
ceenote@lemmy.world · 5 pts · 80d
I saw zero genuine hype for this. State of the industry, I guess.
dogdeanafternoon@lemmy.ca · 0 pts · 79d
Probably cause apple already did it 7 years ago
Curious_Canid@piefed.ca · 4 pts · 80d
Now all that needs is a 16 Petabyte drive to store the database. Well, that and a buyer who wants all of that on their machine.
sturmblast@lemmy.world · 2 pts · 79d
Cash grab
GamingChairModel@lemmy.world · 2 pts · 79d
It's gonna be so fucking funny when the push to sell silicon that can run local models at 100 watts or less ends up destroying the business models of the companies that built out 100,000,000,000 watts of data centers.
FireWire400@lemmy.world · 1 pts · 80d
Laptops != PCs?
SnoringEarthworm@sh.itjust.works · 1 pts · 80d
This guy is wearing the cyberpunk act two villain starter pack.