And mbin!
tal
joined 2 years ago
“Apple choosing to partner with a Chinese military company would be a grave mistake,” John Moolenaar, the Republican chair of the House China committee, told the FT.
“Helping the [Chinese Communist Party] succeed in its plans to dominate critical supply chains will make our country’s tech industry and economy more dependent on China at a time when we must build secure tech supply chains with our allies,” Moolenaar said.
If Micron wants a protected market, then I think that it's not unreasonable to also require that it come with an obligation to make enough memory to meet demand. Maybe they will in the long term, but they clearly are not doing so now and won't be for the next 18 months.
CMXT is not going to refrain from producing memory if Apple doesn't buy from them. Memory is a commodity. As long as there is someone in the world who wants to buy memory
and there is a lot of pent-up demand out there
CMXT is going to be able to sell their memory.
If the Chinese government wants to subsidize memory production, great. Right now, we don't have enough, so that's solving a problem we have using China's resources. If Micron starts to get in trouble down the line, if they can't produce at competitive prices, and we view it as a national security imperative to have domestic memory production, then look at protection.
“It’s shooting pure unadulterated propaganda into our veins,” says another government employee. “Maybe Fox’s editorial standards are too high.”
Necessity is the mother of invention, I suppose.
President Donald Trump’s approval rating has dropped to its lowest point during his two presidencies, according to a new poll.
A poll from the American Research Group, out Monday, found that approval for Trump stands at 30%, the lowest figure recorded across both of his terms. It is a 1 percentage point drop from May, when 31% of Americans were in support of the president’s performance.
I commented here quite some back that I did not expect that Valve would subsidize the console and that it would be for this reason, that subsidizing the console means a razor-and-blades business model, and a razor-and-blades model requires a closed system, where one has to purchase additional product specifically from the vendor of the initial product. They were making an open system, where this isn't the case.
https://en.wikipedia.org/wiki/Razor-and-blades_model
The razor-and-blades business model[1] is a business model in which one article is sold at a low price or even given away in order to increase sales of a complementary good, such as consumable supplies. It is different from loss leader marketing and product sample marketing, which do not depend on complementary products or services. Common examples of the razor-and-blades model include inkjet printers whose ink cartridges are significantly marked up in price, coffee machines that use single-use coffee pods, electric toothbrushes, and video game consoles, which require additional purchases of accessories and software not included in the original package.[1]
Although the concept and the catchphrase "Give 'em the razor; sell 'em the blades" are widely credited to King Camp Gillette, the inventor of the double-edged safety razor, Gillette did not in fact follow this model, nor did it invent the razor-and-blades model, although it did pioneer the production and sale of disposable razor blades.[1][2]
In more recent times, video game consoles have often been sold at a loss while software and accessory sales are highly profitable to the console manufacturer. For this reason, console manufacturers aggressively pursue legal action against carriers of modchips and jailbreaks due to a belief that the resulting possibility of unauthorized or prohibited copying causes a loss in profits. Particularly in the sixth generation era and beyond, Sony and Microsoft, with their PlayStation 2 and Xbox, had high manufacturing costs. As such, the companies sold their consoles at a loss and aimed to make a profit from game sales.[9][10] Nintendo had a different strategy with its GameCube, which was considerably less expensive to produce than its rivals, so it retailed at break-even or higher prices.[11] In the following generation of consoles, both Sony and Microsoft have continued to sell their consoles, the PlayStation 3 and Xbox 360 respectively, at a loss, with the practice continuing with the concurrent eighth and ninth generations of console hardware.[12][13][14]
If you're on Linux and aren't particularly fussed about any issues stemming from power loss, you can probably run it under eatmydata and it'll disable SQLite calls to fsync().
Well...but @MangoCats@feddit.it isn't asking about the spike, but about the absolute price.
PC Part Picker's memory trends page unfortunately only shows the past 18 months. But we can hit archive.org's Wayback Engine.
First of all, here's a current level for DDR5-5200 2x16GB:
So about $500 for DDR5-5200 2x16GB.
They only started tracking this category back in early 2022-ish. It looks like it was about $380 then. Adjusted for inflation, that's $435.14 in 2026 dollars. So it's probably never been that expensive.
However, that was also when DDR5 was pretty new, and it looks like it started out expensive.
If we look at DDR4, which might be more interesting, since we can go back further and avoid the initial spike:
Looking at DDR4-3200 2x8GB, it's come down a bit, but looks like it peaked at about $190.
Inflation-adjusted, that's $144 in 2019 dollars.
It looks like that was about April 2019 when DDR4 exceeded the peak from the last few weeks.
Microsoft's problem, I think, is in significant part that they are the big commercial player trying for a local AI play. Like, your local Windows machine does AI inference. In Anthropic's business model, the inference is cloud-based.
Local is more hardware-intensive, because the capacity utilization of hardware is going to be lower for local AI. If you stick a piece of hardware in a datacenter and lots of people share it, you need less hardware, because when one person isn't using that hardware, another can be. If you would use local AI hardware 1% of the time, then it costs only about one-hundredth the amount from a hardware standpoint to have people sharing parallel compute hardware in a datacenter as to do everything locally. So as long as hardware prices, like shortages of memory, are a constraining factor (or cooling, for that matter, or maybe power if you're talking about laptops on battery, all of which have cloud-based approaches getting an advantage) Microsoft's going to have a harder time of it than the cloud guys.
Microsoft (and local AI in general) does better if people really want low-latency or always-doing-work load, or reliably always-available services, or services where data privacy is critical. There, local AI has the advantage over cloud-based or at least erodes the cloud-based advantage. Right now, I think that that's just not generally where the state of affairs is. Could change in the future, but I think that they're just going to have a hard time of things in the near term. My guess is that Microsoft's relative potential improves as memory prices come back down.
I think that running local LLMs would be great. But the simple fact is that for most users, it's just too costly to make sense for a lot of applications with current memory prices.
I got a Framework Desktop, 128GB, specifically to do local generative AI stuff, in 2025. At the time, the system was $2,500, which is already going to be pricey for a number of people for a single-purpose computer. In the months since that shipped, the price on the exact same hardware configuration has gone up to over $6,500. That's just not a price that a lot of people are going to be willing to pay for a PC. If component supply rises and prices drop back down, then I think that the calculus changes for local AI.
AI companies are acquiring more memory than the entire rest of the world uses. If we want to do the same thing that we could do in the cloud locally and have capacity utilization of 1% on that hardware, then we need a hundred times as much memory as we do with a cloud-based compute approach. That's...a kind of staggering number.
EDIT: Oh, one major exception that could favor local compute, if R&D produces some major improvements in this direction. Right now, the way most models work, we have a very-expensive-to-generate model that's static, unchanging. This model does not change as one does inference on it. This makes it fairly efficient for a single compute node in an AI datacenter somewhere to have this one model loaded onto it, and then be used by many users who are all wanting to use the same model.
However, if the same model isn't being used by many users, then the (expensive) cost of loading a new model onto memory attached to parallel compute hardware for each prompt has to be paid.
If someone comes up with major improvements that can be derived from mutating a model, from updating it as it is used
and I would say with some confidence that human-level AI will require some mechanism giving it more ability to learn after the initial training phase is complete than is presently the case for LLMs as used in 2026
this is something that may greatly shift the balance in favor of local AI. It's something that we, as humans, do.
If I were Microsoft, I might seriously look into advanced AI R&D, where models are updated during use. It's not just that it's an area with potential, but that it's an area that might advantage Microsoft's strengths (control of the local computer, doing local compute) relative to other major AI companies. It only takes one major breakthrough that makes everyone very badly want to have an constantly-updated model to drastically alter the economics of AI compute.
Though...if that happens, as I point out above, that'll likely set off an even greater RAM crisis than happened with cloud AI compute...
In fact, IIRC from an article I was reading a while back, the very first industrial robot was used on an American auto assembly line.
searches
Yes, Unimate, and it was even on a GM line, same as here.
https://en.wikipedia.org/wiki/Unimate
Unimate was the first industrial robot,[1] which worked on a General Motors assembly line at the Inland Fisher Guide Plant in Ewing Township, New Jersey, in 1961.
Devol, together with Joseph Engelberger, his business associate, started the world's first robot manufacturing company, Unimation.[7] Devol's background wasn't in academia, but in engineering and mechanics, and previously worked on optical sound recording for film and high-speed printing using magnetic sensing and recording. Engelberger's ultimate goal was to create mechanical workers to replace humans in factories.[8]
The machine weighed 4000 pounds[9] and undertook the job of transporting die castings from an assembly line and welding these parts on auto bodies, a dangerous task for workers, who might be poisoned by toxic fumes or lose a limb if they were not careful.[4]
Just expend some more compute time on doing compression and we'll get those filesize numbers to a workable level.
$ stat -c %s enhance.png
276773
$ convert enhance.png enhance.avif
$ identify enhance.avif
enhance.avif AVIF 1164x558 1164x558+0+0 8-bit sRGB 14391B 0.000u 0:00.000
$ stat -c %s enhance.avif
14391
$
zoom and enhance
$ identify enhance2x.avif
enhance2x.avif AVIF 2328x1120 2328x1120+0+0 8-bit sRGB 32448B 0.010u 0:00.000
$ stat -c %s enhance2x.avif
32448
$
zoom and enhance
$ identify enhance4x.avif
enhance4x.avif AVIF 4656x2232 4656x2232+0+0 8-bit sRGB 50758B 0.000u 0:00.000
$ stat -c %s enhance4x.avif
50758
$
Okay, that last one took 17 minutes to upscale on my GPU, so I'm not going further. But I'm using SD Ultimate Upscale, which is tile-based, so in theory that could be farmed out over a collection of GPUs and parallelized. Just need more compute hardware.
But as to filesize, that's under 50kiB.
Yeah, that's true. But...if a website derives its revenue from ads, a non-ad-viewing user
and users who switch browsers so as to use an ad blocker would presumably be blocking ads
loses that website money rather than generating it. Like, for such a website, the issue would be how many Firefox users that do view ads would be lost if their website didn't work on Firefox.
Those of you who remember the early 2000s, when Internet Explorer had very high marketshare, probably remember a number of websites that didn't work with other browsers.
I think that the concern is whether a number of websites might stop working with Firefox if Chrome users consistently represent ad revenue and Firefox users generally don't.
I use Firefox, but it has relatively-limited marketshare in 2026. A lot of people just browse on mobile devices and on Android devices, and there Chrome's probably the default browser.
https://gs.statcounter.com/browser-market-share/
Chrome at 70.25%.
Safari at 15.72%.
Edge at 5.14%.
Firefox at 2.19%.
It's the mlmym frontend, rather than just a theme.
Some instances run it locally as an alternate at a different URL; lemmy.world is one such.