Microsoft Acquired Nearly 500,000 NVIDIA "Hopper" GPUs This Year

[Dr. Dro]

Madness. Utter madness

The pinnacle of AI technology. Thanks for investing trillions in this, Jensen, Satya, Elon and Co.

 

[mb194dc]

It's just as mad as when people said a computer in everyhome was madness or the cloud was madness. AI and not doing any processing locally is the future.

Yet Microsoft (and Qualcomm) are also pushing "AI" enabled computers for home use, not that many seem to be selling. Probably because people buying them are going WTF is this.

For me it's just another mass hysteria bubble. I guess they've been "conned" in to thinking they're buying in to "AGI", when LLMs are nowhere near that, get things wrong consistently, hallucinations, because they fundamentally aren't intelligent in any way at all. Wouldn't even surprise me if eventually they're not allowed to call these systems "AI", because it's blatant false advertising, Machine Learning ok.

I wouldn't rule out some use case coming along for all this hardware eventually, but equally it wouldn't surprise me if all of it ends up being given away or in landfill because there's nothing on the front end to pay and keep the datacentres running.

Similar to blockchain, even 15 years later, all it's being used for is speculation, create a "coin", pump and dump it., only difference with Bitcoin is it was the first one. There's no productive utility to the technology at all.

 

[TechBuyingHavoc]

For me it's just another mass hysteria bubble. I guess they've been "conned" in to thinking they're buying in to "AGI", when LLMs are nowhere near that, get things wrong consistently, hallucinations, because they fundamentally aren't intelligent in any way at all. Wouldn't even surprise me if eventually they're not allowed to call these systems "AI", because it's blatant false advertising, Machine Learning ok.

It is telling to me that everyone uses the chat LLMs for stuff that doesn't really matter or that has a high level of BS and artifice. Think of areas like performance reviews, interview questions, corporate emails, informal complaints, etc, etc.

No one is using chat LLMs for the stuff that actually matters to people, where the content of the message is critical, where there is no play-acting or artificial posing layered on top of the message. Think of areas like romantic conversations, chatting with friends and family, personal messages to loved ones, etc, etc.

That is enough to tell me that LLMs do have a use in society, mainly for reducing the time spent on menial tasks or pointless communication but no role in what actually matters. It would be better if we just cut out all the stuff in society that "AI" is currently helping us to manage and save all that time and effort on what actually matters (which we rightfully are not using AI to manage).

This is very far away from all the AGI hype that is being pumped out like propaganda out there.