411–420 of 423 posts
Earlier quoted context omitted.
> 1GB of RAM and everyone could be happy I remember a time when I was surfing the web with 256MB and a 5KB/s modem.
True, but I don't want to go back to that. Bloated web pages suck. But hi-res, true color images and videos don't suck.
And I prefer text over video. Cat videos excluded, of course.
Earlier quoted context omitted.
> But, most people wouldn't. How do you know?
To me it’s so obvious that it doesn’t require testing. The market shows it; there are nearly no production vehicles at volume with crank windows.
Earlier quoted context omitted.
Food calories are cheaper to convert into something useful. It's not like GPUs, once bought for peanuts, turn into perpetual motion machines. They need power, cooling, a whole infrastructure built around them. GPUs would have taken the world by storm already in the roughly 30 years since they've been around. Even for GenAI it's likely ASICs take over at some point if we really care about performance.
GPUs have taken the world by storm. There's one in almost every computer, and they make up the bulk of supercomputers! If you put a 75% discount on these powerful GPUs there will be a long line of non-AI-company purchasers.
Yes, lots of companies will buy them for cheap, but these AI beasts also have OpEx costs. Not every alternative use is worth the money and there are 0 guarantees that the alternative costs cover the gap. NVIDIA sell 80% of GPUs for AI now.
I think people don't realize just how big this bubble is.
Inflation metrics seem to fail to capture the increase in leading-edge tech products of the past 18 months.
They seem to fail to capture a whole lot of things. Supposedly $1 in 2000 is worth $1.88 in 2025. So 88% inflation over the 25 years. Meanwhile the median home price has increased by 150%. Family insurance by 350%. Median college tuition by 225%. Childcare costs have risen by 200%. But sure. We can buy super cheap 65" tvs now. Hurray for us! Literal kings who lived hundreds of years ago couldn't possibly imagine a wo…
Earlier quoted context omitted.
You have to wonder whether they realy intend to use the ram, or they just spotted an opportunity to corner an essential market and choke/extort in resale.
Do they have the extra money to play games like that? I thought they were pulling back on ads because of the community reaction even though they wouldn't be introducing it now unless they needed it.
Earlier quoted context omitted.
About fifty teraflops. That's the meaning of intrinsic calor - the device can do what it can do, regardless of market conditions. Today it has the value of fifty teraflops, and tomorrow it still does, unless it breaks. However, intrinsic value cannot be measured in dollars.
And yet we're talking about electronics here, they don't have sentimental value and just because compute capacity is unused there are no guarantees that it will be used, even at a per unit cost approaching €0. I'm sure that farmers during the Great Depression were also consoling themselves with the "intrisinsic caloric value" of their corn.
There are also intrinsic costs, mostly power consumption.
Somehow another sign that our world is collapsing. We have more and more wars, less free trade and international spirit, freedom and privacy is going down the sink, "everything is going down", and now ram is a good example of things that use to be abordable and with price going down, becoming expensive luxury good. One could say that RAM price is just a special temporary event not related to our down trend but a deca…
It certainly seems clear that a purely upward trajectory of peace and living standards in the West was not supposed to last. History has its ebbs and flows, I'm sure we have a long way downhill to go, but I suspect there are peaceful times on the other side.
Earlier quoted context omitted.
When OpenAI fails, all the AI projects everywhere else will also be killed. Such is the nature of bubbles. With any luck the farm land where the datacenters were built will be sold back to the farmers for half off and they get a free barn out of the deal.
It's more likely that overcapacity is put to work in a plan B, like cheap cloud virtual desktops. Why spend effort on spying and tracking users when their whole desktop computer is in your data center?
Earlier quoted context omitted.
I wonder if semi-reliable RAM could be made to work for training. After all gradient descent already works in a stochastic environment, so maybe the noise from a few flipped bits doesn't matter too much.
Also, depends on the nature of the error. If only a small memory range is affected, you could patch the kernel to avoid it.