Hacker News Reader

This is an downright evil take on the current situation. The supply chains are so complex that no single country is capable of replicating them entirely. It starts with the fact that the required natural resources are distributed around the globe in a way that no country has access to all of them. The production chains from resources to finished machines are downright byzantine. And this becomes recursive with the need for specialized tools and their own production chains along the way. You need trains amd trucks and ships to be able to build semiconductors, for example. Except for maybe China pr India, there is no country that has the manpower to cover all of this domestically. The supply of workers and training falls far too short.

Any Western strategy that sees this as both "us vs. them" and also pursues reduced international collaboration is bound to lose bitterly in the long run.

The result is either a silent collapse of that country's economy or the start of an ill-conceived war of conquest to gain by force what the country cannot supply itself.

> Any Western strategy that sees this as both "us vs. them" and also pursues reduced international collaboration is bound to lose bitterly in the long run.

The problem is that Europe does not have a choice here. The Greenland steal crisis is on hold, but not fixed. America clearly shown it will abuse any ties there are - it will lock accounts to tech to bully and get what they want. It will use tariffs to bully countries to make laws, release presidents friends criminals from prisons, you name it.

Meanwhile, America seems to take Russian side in Russian expansion. Meanwhile, America is just cause major oil issue and potentially triggered next refugees crisis. Meanwhile, America clearly shown it does not even pretend to care bout war crimes and international law at all. It is sponsoring afd and other fascist parties all around the Europe while openly insulting Europe. Maybe it is too late for disconnect, but not trying would basically be a suicide for Europe.

It would be great if it was not "us vs them". But it is "us vs them". Trust toward American made Europe super vulnerable.

The US isn't the navel of the world. It is one country that is slowly removing itself from international trade and the international scientific community.

The European Union has many friendly trading partners left in the world and is also receiving an influx of previously US based talent. The trade decisions of the US aren't forcing the EU into isolationism. This is where your argument goes wrong IMO.

The US government has announced that it plans to actively support extreme right wing parties in the EU. If this comes to pass, it is a direct attack on political freedom in those countries, separate from any economic policy decisions. I don't know how well EU countries can defend themselves against this in the short amd medium term. Some counties have better defenses than others. But I see virtually all of them struggling.

> It is one country that is slowly removing itself from international trade and the international scientific community

While asserting itself militarily. This is the Roman Republic —> Empire transition.

> It's absolutely insane that Claude Code can spit out a week's worth of business automation tasks in half a day. And do it at relatively high quality in low-defect rate languages like Rust.

> Europe won't be able to catch that.

You think Europe won't be able to use Claude Code? If Claude Code is the one reaping the majority of the benefits of "spit[ing] out a week's worth of business automation tasks", then it's not worth much to the business. If Claude Code isn't the one reaping the majority of those benefits, then...Europe can use Claude Code too and reap the benefits for their business as well.

Well, if it becomes strategically advantageous to bar Europeans from doing so, why would we (Europeans) be permitted continued access?

Weird how all the American social media companies continue to try to operate in Europe in spit of the massive fines they keep on racking up in court. They can't help themselves, if there is money to be made they got to get in there.

Social media is a tool to shift public opinion and extract cash from the general populous. Claude, however, is actually useful.

The US government and software industry do not want Europe using Chinese AI for similar soft power reasons. A 1-billion strong market acclimating to Chinese *aaS is a net-loss for the US - see the panic about Canada allowing a few thousand Chinese cars

Open models are, at worst, a few months behind SOTA closed models. This has been the case since 2024, and there's no indication that it's going to change.

You don't need anybody to permit you access.

You can, in all seriousness, thank Meta and the Chinese for this.

> Open models are, at worst, a few months behind SOTA closed models.

What open models are "a few months behind" Claude Code Opus, Google Nano Banana Pro, or ByteDance Seedance 2.0?

Another big problem with the open models is that the ecosystem targets consumer GPUs rather than running in the cloud. The future is a thin client world with stuff at the edge being generations behind in performance.

> Claude Code Opus

Kimi K2.5 is quite literally just a few months behind, at worst. In many cases it already outperforms Opus 4.5 (and reliably outperforms Sonnet), and it's not really that far from Opus 4.6.

I'm less familiar with the video/image side of things.

> Another big problem with the open models is that the ecosystem targets consumer GPUs

No, this is something of a misconception. It's quite difficult to run a model like K2.5 truly locally. But any nation-state or sufficiently large corporation can get its hands on HPC clusters and run it freely, without needing permission from anybody.

"A few months" is an incredibly long time when the gap is widening on a daily basis.

You really do not understand how interdependent everyone is. The chip machines Taiwan uses come from Europe, for example.

The US kneecapped itself for no reason.

> The chip machines Taiwan uses come from Europe, for example.

Yeah, the EUV photolithography machine, but not much else. American companies like Lam Research and Applied Materials are the leaders in thin film deposition and etch, KLA Tencor is the leader in metrology, and Synopsys and Cadance are the leaders in EDA (though there's also Germany's Mentor Graphics).

> The chip machines Taiwan uses come from Europe, for example

The US, not Europe.

ASML's EUV and High NA EUV production is all done in California via US DoE joint ventures (specifically Cymer LLC [0]). Additionally, their metrology IP is Taiwanese [1] as part of ASML's acquisition of HMI back in 2016 with Taiwanese approval [2].

ASML is the capital partner because in the early 2000s, the US government wanted to prevent a duopoly forming between Nikon and Canon for photolithography as part of an antitrust battle.

And the next generation of lithography tooling coming into Taiwan is being funded and developed by Japan [3] due to MUFG, Mitsui, Mizuho, and SoftBank becoming the primary capital partners for Taiwan's electronics industry [4]. This is also why TSMC is expanding in Japan and Taiwanese players are transferring IP to Rapidus.

Additonally, all the packaging, testing, and design work - especially leading edge nodes - is done in Asia, the US, and even Israel but not Europe.

---

Personally, I think Europe is too far behind at this point for the EUV and High NA EUV boom - Taiwan, the US, Japan, South Korea, China, and others deployed significant amounts of capital and subsidizes in the late 2010s and early 2020s and worked to build IP partnerships for front-end work with players like TSMC (US, Japan), UMC (China until 2019), PSMC (Japan, India), and Samsung (US).

The EU had a shot but Intel rolled back their Germany expansion plan in order to double down on 18A in Arizona, and TSMC decided to double down on Japan. Additonally, all the backend work is done in Taiwan, South Korea, China, ASEAN (Malaysia, Thailand, Vietnam), Japan, the US, and even India now because Micron, Samsung, Amkor, and others transferred their IP there, and design is primarily concentrated in Taiwan, the US, China, Israel, and India and with Malaysia and Vietnam likely to become much more prominent in the next few years due to Arm and Marvell respectively.

What the EU can instead do is concentrate on power electronics (already a strong suite) and compound semiconductors (already a strong suite) and target a leapfrog technology like 2D semiconductor design and fabrication which is still in it's infancy and also has applications for quantum computing. The EU already has the capacity for "legacy" (but still critical) semiconductors but is too late to the game for sub-14nm fabrication.

And based on the kind of fundamental research and funding I've been seeing in the EU, this is the strategy that appears to be increasingly adopted within the EU - but this is something the US, China, Japan, South Korea, Taiwan, Israel, the UK, Singapore, Canada, and even Vietnam and India are doing as well, and both French and German initiatives risk being misaligned due to mutual industrial competition. The fact that French players in the space would rather collaborate with Singaporean [5], Korean [6], American [7], Israeli [8], and Canadian [9] partners to develop IP instead of with other partners within Europe, it shows issues around misalignment.

On a separate note, as I mentioned before on HN, the French seem to be on the right track - other European states other than the UK less so. French players are much more ruthless and "American" in attitude.

[0] - https://www.cymer.com/

[1] - https://www.asml.com/en/company/about-asml/hmi

[2] - https://www.asml.com/en/news/press-releases/2016/asml-obtain...

[3] - https://www.nikkei.com/nkd/company/us/SNPS/news/?DisplayType...

[4] - https://asia.nikkei.com/business/tech/semiconductors/japan-l...

[5] - https://www.pasqal.com/newsroom/three-new-partnerships-signe...

[6] - https://www.pasqal.com/newsroom/pasqal-expands-into-korea/

[7] - https://gov-pritzker-newsroom.prezly.com/governor-pritzker-a...

[8] - https://www.quantum-machines.co/blog/pushing-the-boundaries-...

[9] - https://www.lemonde.fr/en/economy/article/2024/09/25/quebec-...

The laser source, not the rest of the machine.

The US used this agreement to bar ASML from selling the machines to China in 2018.

The machine is a clump of metal without the light source.

Its also useless without the rest of the machine, which is likely the absolutely most complex thing we can build, except maybe things like the stellarator.

The metrology is coming from Taiwan and California (HMI) as well. The Veldhoven campus "only" does final assembly (which should not be underestimated either - it's complex and high precision work).

But it's the light source and the metrology that is the blocker.

Edit: can't reply

> And mirrors from Carl Zeiss

Absolutely! But note how Zeiss/Trumpf is not ASML. If the US DoE changes the terms of the Cymer partnership and pressures Taiwan (who have just purchased $8B in US military equipment and whom the EU logistically speaking cannot protect) to revert the HMI acquisition, ASML is over.

Additonally, a lot of the muscle around Zeiss/Trumpf's mirrors is also at the Zeiss office in the Tri-Valley because of their partnership with LLNL.

And both China [0] and Japan [1] are in the process of building an Ex-ASML supply chain for EUV, NA EUV, and DUV, and will likely reach that point by the late 2020s to early 2030s.

[0] - https://www.reuters.com/world/china/how-china-built-its-manh...

[1] - https://www.nikkei.com/nkd/company/us/SNPS/news/?DisplayType...

And mirrors from Carl Zeiss. Heavy lifting on "only".

Where can I see the DOE having a stake in Cymer?

Department of ENERGY not the DoD.

The EUV IP which Cymer owns was originally part of EUV LLC, which was an LLNL [0], Sandia [1], and Intel [2] initiative as part of a CRADA. Cymer eventually began working on EUV as well building on EUV LLC and SEMATECH's work, and was eventually purchased by ASML after the Dept of Commerce and DoE backed their acquisition in 2013 [4], but included Cymer in additional CRADAs [3], ongoing projects [5], and maintain Cymer as a separate operating unit [6] within ASML.

It is these CRADAs that allow the DoE to exert its muscle on IP ownership and knowhow, as they are essentially a license of IP and personnel on US DoE terms [7] while also allowing for private partners to commercialize.

[0] - https://www.llnl.gov/article/27641/euvl-partnership-makes-it...

[1] - https://newsreleases.sandia.gov/partners-unveil-first-extrem...

[2] - https://www.intel.com/pressroom/archive/speeches/EUV91197.HT...

[3] - https://www.osti.gov/biblio/1413999

[4] - https://www.asml.com/en/news/press-releases/2013/asml-comple...

[5] - https://www.llnl.gov/article/52226/llnl-selected-lead-next-g...

[6] - https://bidenwhitehouse.archives.gov/wp-content/uploads/2021...

[7] - https://www.directives.doe.gov/directives-documents/400-seri...

I don't see a stake by the DOE?

> They're going to be a decade behind if they keep it up.

42 European here.

I've heard my whole adult life that Europe is ten years behind USA.

That doesn't feel that bad though. Being bleeding edge comes with the thrill of the avant garde prestige. But it does also mean you take the downsides of navigating the unexplored unknown in your face with no one to help with turn key solution when it happens.

If it means 10 years buffer on big social seismic troubles, that doesn't sound too bad if there is indeed an efficient shelve. That's not necessarily the case on every matter though, like global climate change is going to impact everyone, no matter the political isolation, and if a direct military aggression happens, it can be hurtful no matter how prepared is the society.

I'm not a fan of this "us vs. them" framing.

Arguably the greatest threats to the US's future is ourselves. If we fundamentally corrupt who we are as a nation we've already lost before the competition with rivals has even begun.

Our significant tech advances could become tools of our own downfall if they violate our values or undermine the social mobility of the American dream.

Frankly I think the people pushing this competitive mindset (particularly against the EU) are trying to mislead otherwise intelligent builder-sorts to not pay attention to the looting & destruction of American values.

>Europe won't be able to catch that. If we cut the chip supply right as things take off, China might not either.

https://www.asml.com

I would be worried about the US militarily capturing Eindhoven if Europe dared to cut them off.

Unlikely, but also not needed. The major ruling parties in The Netherlands are always strongly pro-US. Even if the US would attack the International Criminal Court in The Hague and kidnap its judges, the VVD (the liberal party) and the PVV (extreme-right party) would find ways to defend it.

There is no way they would go as far as cutting the US off from ASML. At least not without a significant shift in Dutch politics.

ASML is under US export control though. Probably because there is US tech in there.

>>Europe won't be able to catch that. If we cut the chip supply right as things take off, China might not either.

My immediate thought is - why is it a race. Like holy shit imagine if we could actually work together instead of having this mentality of "if we work hard the other countries won't catch that". As someone who grew up in the golden age of globalization and rise of the information superhighway, the way countries are just siloing themselves and treating everything as a zero sum game is both sad and scary - that's exactly how you lead the world on a path to another world war - telling yourself that you don't need anyone else and in fact you need to beat them to the punch and everyone else is your opponent. If an alien race was looking at us right now they'd be shaking their heads.

People tend to choose extremes. Either a total globalization or zero-sum games only. Everything in between bears risks of a cognitive overload so should be avoided at all costs.

> People tend to choose extremes.

Corporations, along with greedy, selfish people and also perhaps ignorant people too, tend to choose extremes.

The rest of us are remarkably good at compromising and finding common ground.

I think when Europe failed to handle Ukraine on their own, the optics started looking pretty bad.

They're handling Ukraine better than America is "handling" Iran

I honestly don't even understand what you mean by that. Europe has accepted an extraordinary number of refugees from Ukraine giving them immediate and full rights to live and work within its territory, keeps donating billions of dollars worth of goods to keep Ukraine going, provides training to Ukrainian military personal and ramped up its military production specifically to bolster Ukraine.

How has Europe failed here? Did you expect Poland to start shooting missiles towards Moscow, or something else?

If the US wasn't around, Russia would likely be rolling around most of Eastern Europe today. From 2022 to 2025, America and Europe went dollar for dollar in support of Ukraine. And many of the Euro dollars went to purchasing weapons and defense from the US.

Euro leaders are terrified by this. Europe, which is supposed to have global parity with the big players, was just as much (maybe even more) of a paper tiger as Russia. But unlike Russia they have a big friend who has lots of money and lots of weapons.

For the average European who has only ever know life under the protective wing of NATO (read: America), this is just Tuesday. For the versed European leader, who needs to ensure their country can be safe if America goes away, it's totally harrowing. Even more harrowing when your population doesn't understand the need for defense.

><>If the US wasn't around, Russia would likely be rolling around most of Eastern Europe today. From 2022 to 2025

No one is saying otherwise. Still, I don't see how you can describe Europe as failing. Unless, like I complained in my original comment - you see the world as some kind of race where only one country is "winning" and everyone else is losing(failing).

>> America and Europe went dollar for dollar in support of Ukraine. And many of the Euro dollars went to purchasing weapons and defense from the US.

In terms of military spending, yes. In terms of refugees support and civil expenditure, Europe was obviously ahead. And again, this isn't a race to see who can spend more.

>> Even more harrowing when your population doesn't understand the need for defense.

As a Pole, I don't see this sentiment at all - Poland really understands the need to defend itself, with our spending being allocated accordingly - and importantly, new funds are coming with the specific provisions to prevent them being spent on American technology.

>> of NATO (read: America)

Obviously, America has the largest fighting force on the planet. But are we just going to pretend that UK, France and Germany don't have capable armies which would be capable of defending our eastern flank if needed, including with nuclear deterrent? I know Americans are being fed the stories of NATO being entirely US and no one else, but obviously even cursory inspection shows that this isn't true?

I have been a nay sayer on LLMs/GPTs in general having tried many, but recently Ive been shepherding a fairly complex code build through the latest opus model and its quite impressive.

It still gets things wrong occasionally but the time its saved me has been substantial. Im starting to enjoy it.

I recently built a reasonably-complex embedded controls project using codex and an esp32.

Starting with systems stuff like "Set up vscode with whatever it needs to work with codex and talk to an esp32," and ending with "Now add a web interface with persistent tunables that always runs in both AP and station modes," my prompt inputs were very terse.

And it'd just kind of go forth and just do it. It'd even design and run its own tests.

I never once looked at the code. For all I know, the code doesn't even exist.

And it works. I'll be using it in the field (in the proverbial middle of nowhere) all next week. I have every expectation that it will behave itself.

(I did spend a lot of time defining and refining some ground rules with AGENTS.md, but in theory I get to re-use that effort for the next go.)

Just to follow up: I used it all week, in the field.

No issues to report.

The controls-widget behaved perfectly fine and I still have not seen the code. At this rate, I may never see it -- it seems to be a rather comprehensively complete project.

> America [...] needs to make friends with all of its former allies. And it needs to import PhD students.

America, in the form of the Trump administration and a Trump-subservient Congress, just spent the last year completely destroying trust on these issues and it would take decades of sustained effort to rebuild it.

>>Edit 3: What I'm saying is that these tailwinds might put America back into the position it was in post-WWII. Manufacturing, tech, and science powerhouse in all the places that matter. Peers a generation or two behind. That's literally where America was after the war, and it looks like we could be teeing up for a repeat if it all doesn't unravel first.

Let us know when Trump is finished sacking the ivory towers these imported PhD students are supposed to land at.