Hacker Newsnew | past | comments | ask | show | jobs | submit | capitol_'s commentslogin

That Trump makes us very motivated to stop relying on American tech.

This doesn't solve the issue that globalism caused. Europe doesn't make DRAM nor has the know-how to quickly bring factories online which usually take 10+ years.

We are tied to American economy and if AI companies start driving prices up not only DRAM but basically everything will become more expensive.


America doesn't manufacture DRAM either, this is all South Korea and Taiwan.

??? Micron has DRAM megafabs in both Idaho and New York state.

https://en.wikipedia.org/wiki/Micron_Technology


They don't "have them", they're building them.

https://www.micron.com/us-expansion/id

> Micron has already achieved key construction milestones on its first Idaho fab with DRAM output scheduled to begin in 2027.

https://investors.micron.com/news-releases/news-release-deta...

> Production is expected to start in 2030 with the fabs ramping throughout the decade.

Until they start outputting DRAM in any meaningful quantity, they're not relevant.


> They don't "have them", they're building them.

According to wikipedia Micron Fab 6 in Virginia started production in 1997 and is still operating


> "in any meaningful capacity"

Building a factory is one thing, they can have 50 of them built, but that doesn't mean much if all 50 together amount to like 0.1% of the company's output.

Once those factories scale up to 1-2%, then we can start considering that they've actually built a domestic supply, but that's a whole different goal than simply building the factories. Building factories is trivial. Making them output something is also "trivial". Scaling that up to a meaningful amount is a whole different, much harder goal to accomplish.


Are you sure? I'd think producing any RAM chips at all is much harder than scaling up a working production process.

Yes, I'm pretty confident you don't know anything about manufacturing at scale.

Say you use a magic wand and build 15 new state-of-the-art factories tomorrow. Who's gonna run them? Does any location in the US have enough qualified workers that can simply take over and produce RAM in them from day 1 with no major fuckups?

No, you need a ton of time to teach thousands of people how to run those 15 factories. To even begin to teach people, you need to have 1 factory up and running. That 1 factory is at first going to be run by some of their existent workforce that they temporarily migrate from South Asia. Only then can they start to teach local populace how to run those factories on their own.

This is why it's much cheaper to simply build an additional one in South Asia than it is to build more than one in a whole new location. South Asia already has a bunch of workers that know what they're doing because they've been doing it for a long time. Build a new factory, promote some of your existent workforce up the chain, fill the lowest positions with fresh graduates that are gonna be equally good every year and you're good to go. It's nowhere near that simple in a brand new location, where even the most optimistic scenario would take longer than a decade to produce a meaningful amount of output.


Wow, so it's almost like the workers are a key piece in producing any value at all.

Not to mention, given recent US immigration enforcement actions at various manufacturing plants, you can't even safely bring in overseas workers to train your domestic workforce...

It looks like it's still a big difference between how the US and EU are responding to the chip supply wars. The US is actually building their own manufacturing capabilities domestically while the EU is apparently doing nothing, which is unfortunate.

Infineon is _opening_ its fab plant in Dresden this year which was supported by around 1bn euros from the EU equivalent of the CHIPS Act. They started building this fab in 2023, while TSMC, who started building its fab in the US right after covid just delayed the opening to 2027

The fab that Infineon is building is vastly smaller in scale, and their tech isn't really relevant to this discussion. For instance, it doesn't produce CPU/GPU microchips or DRAM. Also only 300mm wafer technology, which isn't competitive for anything except for some narrow industrial use-cases. Glad to see the EU is doing it, but it's a completely different thing.

Pretty much everyone is on 300 mm wafers for everything now, and has been for a while. Are you perhaps reading this as 300 nm process (which would usually be called 0.3 micron)?

But in the context of what we are talking about it's still true that nobody in the EU is making cutting edge CPU/GPU/DRAM and there are no plans to do so either (including that Infineon fab).

Are those plants still functional after CHIPS act was axed? I thought they mainly produce in Asia now.

Well first of all, the CHIPS Act was not "axed", it is federal law passed by an overwhelming bipartisan majority of the House and Senate. It would take a complete reversal of congress to repeal it and it's still very popular among both parties.

Where do you get your information from?


> Well first of all, the CHIPS Act was not "axed", it is federal law passed by an overwhelming bipartisan majority of the House and Senate. It would take a complete reversal of congress to repeal it and it's still very popular among both parties.

DOGE cut basically all staff from the CHIPS Program Office, congress passed the money but Trump is choosing to turn it into a slush-fund the admin spends on industrial policy (such as buying a stake in Intel). Wolfspeed went into bankruptcy in part because the admin delayed CHIPS funding agreed by the previous admin [1] (it's unclear whether they received the grant now that they have left it).

[1] https://www.ft.com/content/4aac09f9-19df-401a-9ab3-ef14a47bb...


Once again, where do you get your information from? The only thing that doesn't exist anymore is DOGE itself.

https://wtop.com/government/2025/11/doge-quietly-disbands-8-...


Does eliminating DOGE magically put Humpty Dumpty back together again?

Just like executing a serial killer brings back all the victims, right? No harm done!

You hire someone to reqire a few outlets in your kitchen. They burn down half your house. You fire the crappy electrician.

What's the state of your house now?


If Doge disbanded but its actions haven't been undone, it doesn't matter if Doge doesn't exist anymore.

> after CHIPS act was axed

This is news to a lot of Americans! The 2022 CHIPS and Science Act is codified federal law. I think a lot of states (Arizona, Idaho, New York) would be very interested to learn that the funding for the infrastructure that they are already building has somehow gone poof.


Intel is now partially government owned(10%), they got rid of some of the milestones. The current administration has been extremely poor about communicating changes as well as constantly yanking funding (or threatening to) for projects - the chances of funding going poof are higher than ever.

> Currently, 100% of leading-edge DRAM production occurs overseas, primarily in East Asia.[0]

They make DRAM for cars, not computers, in the USA. They've promised they'll bring manufacturing onshore any time soon, which effectively means they'll wait until Trump forgets about it.

0: https://www.nist.gov/news-events/news/2025/06/president-trum...


That's not how it works, DRAM substrate (the actual chip that contains memory cells) is shared. It's only the packaging that differs.

American companies are driving global economy insane. Currently the American political administration sides with the AI companies since it gives the inspiration that the economy is doing well. If things start to go side ways, the US government can put pressure on its local companies like Micron to supply other fields.

Europe doesn't have local manufacturers. So it cannot exert control over the manufacturers to keep its internal / strategic market sane. All European hardware manufacturers have to put up with and compete in irrationality inflated prices.


And China with IXMT.

Europe has stopped making DRAM relatively recently (Qimonda).

This should have not been allowed to happen.


Guess who bought Qimonda's patents?

Who?

Micron

and facilities went to China's Inspur.

Isn’t there also basically 0 American DRAM?

Micron Technology, Inc. is an American semiconductor company that manufactures computer memory

They don't produce them within the US. They're building some factories to do so in the future, but as of now their output is 0.

However, the US government has / can have control over Micron's production. They are headquartered in the US. They have the intellectual property and know-how to erect a vertically integrated supply chain. Europe doesn't have this strategic investment.

The newfound desire to move away from American cloud providers isn’t related to pricing, it’s about the perception of growing instability within the American government, the perception of deteriorating freedom of speech, and the perception of an increasingly non-neutral business environment.

E.g., if I’m running a business in the US and I don’t kiss Trump’s ring (and pay bribes), if he becomes dictator for life in 2028, all bets are off for my business.

Both the EU and USA import the majority of their computer equipment, and the USA is placing heavy and unpredictable tariffs on those goods. It’s hard to argue that a business should bet that data centers will be cheaper in the US than in the EU if Trump is the last democratically elected president.

The most stable places to do business in 2026 are probably the EU and China.


You waited far too long.

one.. fully autonomous, self improving, replicating, general intelligence.

It means selling to all bidders, since it's information and not a tangible asset.


A prediction is not a claim.

Predictions operate on events that will happen in the future.

Proofs typically operate on things that already exist.


What on earth do the Direct Rendering Manager api have to do with commercialization?


I think they meant Digital Rights Management, but it's only a guess.


I have worked in the industry for a little over 20 years, i wouldn't be surprised if I have written between one to two million loc.

I would be surprised if it's significantly above that.


And yet, without the software for Linux gaming, there is no Linux gaming.

Very hard to falsify such a statement.


Software written for Windows, running with a translation layer on GNU/Linux.


The translation layer doesn’t really matter though, does it? If a user installs a game and it runs the same, the user doesn’t care about the translation layer inbetween. If installing and running a game on Linux is the same as running it on windows, there’s no reason to prefer one over the other for gaming.


It certainly does, because it allows game studios to keep ignoring GNU/Linux, even when they happen to have Android/Linux games written with the NDK, it is a Valve's problem.


With better performance than on Windows


Maybe in rare cases with few compatible games.


In some cases.


is there a point somewhere in this statement?


Not the parent or grandparent poster and not a gamer.

The echo in my mind from the statement was along the following lines:

I can do everything at work remotely from my Linux laptop as they use Microsoft365/Sharepoint/Teams/Outlook and all. I can just log in via Chromium and noone knows any different with one exception: the finance portal. I have to be on an employer owned Windows PC to do that one thing as it is the last 'native program' needed. Moral: enterprise-ish stuff is happening via the Web browser.

Steam et al financing WINE/Proton and generally hammering all the sharp edges out of the compatibility layer for running Windows software on Linux. Moral: Complex Windows native software can be run under Linux.

So, at some point in the future, does Microsoft just phase out Windows? Replace it with a really well engineered Linux with compatibility environment for legacy software?


Embrace, Extend, Extinguish/Exterminate? They already begun the Embrace phase: WSL.

The smartest Extend phase they could do would probably be a "Windows" GUI on top of Linux kernel, possibly with some customized locked-down systemd, to replace the aging X and the mess Wayland created. If it gets to be at least as functional as Win11 is, it will instantly wipe out the other two alternatives - Exterminate.


Check how many Linux contributors are on Microsoft's paycheck, including systemd author and some Rust people also related to Rust on Linux kernel efforts.

Microsoft already has their own distro.

And they don't need to bother with anything else, Valve with Proton, makes Windows, Visual Studio and DirectX the way to go for the large majority of game studios.

WSL on Windows, alongside Virtualization framework on macOS, are the Year of Desktop Linux, regarding the latops I can actually buy on a random shopping mall computer store.


Games work just fine through Proton already, except when they require kernel level anticheat. I'm fairly certain OP is just one of the purists who think it's not done "proper" until it's a Linux native port, which I wholeheartedly disagree with.


Why should Microsoft bother, when they have Visual Studio and Windows licenses that game studios gladly pay for?


A reality slap.


So no point to make then, cool, I can get back to playing games then


Make sure to use MAME as well, those arcade games are also Linux, apparently.


1. Nobody said anything about Windows games being Linux games. We were talking about Linux gaming, which is gaming on Linux. Which - yeah - emulators also contribute to

2. Above being said, translation is not emulation and has much less overhead So many pointless semantics to dismiss something genuinely good and useful


Translation is one form of emulation, because GNU/Linux still isn't Windows, at the end of the day.


No, it is not. Right there in the name of WINE.


I agree, if people just had refrained from building things in c/c++ that operated on data from across a security boundary we wouldn't be in this mess.


> That is a fairly simple question. The answer to it should be simple too.

"Is P equal to NP" is also a simple question.


The answer should be simple, too -- either yes or no. OP did not imply proving it would be simple.


I'm pretty sure that you won't be able to claim your 1 million dollars from the Clay Mathematics Institute by just answering yes or no.


That's why I'm going to go with a friend. I'll say yes, he'll say no, and we'll split the winnings. Easy money.


OP did imply that the paper contained the simple answer, though.

It’s easy to say that the truth is simple if you ignore everything about exploring whether or not a paper is an accurate representation of the truth.


650 GB fits in ram: https://yourdatafitsinram.net


what a pointless website. It would be nice if it at least showed appropriately sized cloud instances of just the same list over and over again.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: