I don’t think the industry realize the majority of AI output is youtube slop videos
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.
“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.
No thanks.
Every corporation for some reason: Nah you gotta
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.
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.
We have this already. They’re called APUs.
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?
Isaac Asimov’s future of angry men yelling at uncooperative robots is almost here.
imagine all the people screaming at their computers at the office:
ABORT ABORT CANCEL YOU STUPID FUCK
LMAO
ABORT ABORT CANCEL YOU STUPID FUCK
If only Huang’s mom would’ve yelled that some 63 years ago.
Chuckles upvote for you
deleted by creator
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.
“… for Windows”? No thanks.
dave2d said it better. Having agents in your laptop doing your work for you is revolutionary, but trusting Microsoft with it is the problem.
Mainline linux drivers or gtfo
I saw zero genuine hype for this. State of the industry, I guess.
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.
