Crest of Light

Oh how disgusting, T-Mobile is testing an NVIDIA TX PRO 6000 Blackwell Server Edition GPU attached to one of their cell towers to create something called, "Physical AI". They are also working with Nokia on attempting to execute this nightmare...Apparently, it's supposed to supercharge LLM systems, by increasing responsiveness and make them generally faster at performing tasks.

I really don't see the appeal because it would allow this malignant tumor given birth to by Silicon Valley and Peter Thiel's machinations access to critical infrastructure. Which is honestly terrifying, giving how bad LLMs are at doing basically everything, save for compromising anonymity of people online.

Once I pay off my device, I will be calling it quits with T-Mobile because I want no part of a company messing with shit like this.

https://www.phonearena.com/news/t-mobile-nvidia-nokia-5g-physical-ai_id178964

#Tmobile #FuckAI #technews

Mar 19 at 2:34amWeb
Show threadChristian Rickert5d ago

@WanderingInDigitalWorlds

I can see a future where gamers open the cell towers like a can of sardines: arrrrr! 🏴‍☠️

Show threadCrest of Light5d ago
@christianrickert The only way to fight back against the LLM scourge is to forage for parts from their data centers. It would have to be a massive flash mob, to limit their ability to track people down and punish them for finding the right PC part for an upgrade. ROFL
Show threadChristian Rickert5d ago

@WanderingInDigitalWorlds

In a bit these data centers become obsolete (AI bubble burst) - they're going to sell the hardware for scrap soon enough.

Show threadCrest of Light5d ago

@christianrickert I'm sure, as the pressure isn't dying down...I just feel a bit of civil disobedience would hasten it along...For fun, naturally.

As these muppets can't hold a narrative that is going to win over anyone. If you've ever heard Sam Altman or Jensen Huang talk...I wouldn't be too interested in LLM powered tech.

Given how badly the shit they do peddle works...That's also another reason for my lack of interest. If only they didn't try to market it as something LLMs could never be. If they were smart, they could've won.

Show threadChristian Rickert5d ago

@WanderingInDigitalWorlds

Lucky for us (and the environment), they aren't smart.

I see AI as a "one-trick pony" - it's useful as a tool for specific tasks, but that's about it. Neither smart technology, nor sustainable, nor is there any path forward to a generalized intelligence: they already siphoned literally the knowledge of the world and didn't make further progress.

I think it's fair you call these criminal fraudsters muppets. - It's a show they put up after all.

Show threadCrest of Light4d ago

@christianrickert I agree, that the environment and us, are lucky for the fact these muppets aren't intelligent.

They chose the wrong path to create an AI; still feel it will be a possibility, but not for at least 150 years of concerted human effort, involving many scientific disciplines. It will take generations to get to the point that AI is truly a possibility.

LLMs were never the way to that end; the only purpose it served in that respect, is to sell the illusion of AI rather than the actuality.

Proven by the fact they scraped, stole, and consumed so much human created information and all we got was a sexy suicide coach ChatGPT...

A Claude that needs heavy edits and reworks in order to extract use able code from it's outputs. A skilled human could create that code faster, with less junk, and barely any hallucinations (unless they were up for nearly 24 hours, might get loopy then).

LLMs do have uses, but the core problem is that they are being used incorrectly, to force humans out of jobs to maximize profits. Also the data is being collected without strict ethical standards, which truly alarms me in a big way.

Edit: I wanted it include a picture that made me chuckle. As I'm a nerd.