Hacker News Reader

Hacker News

Comment by simonblack | Hacker News Reader

Full thread

simonblack

How can they make their chips commercially competitive? US input costs are a lot higher than both TSMC's and China's.

There is a reason that US manufacturers have outsourced their production to Asia for the last three decades.

View on HN

Well, chip manufacturing is mostly automated, so labor costs are insignificant.

Also, Intel is on path to regaining process lead, they can compete with TSMC and Samsung based on quality.

I'd say yes and no. TSMC has over 60000 employees. They're also probably the biggest siphon of fresh graduates in Taiwan. Like most Taiwanese companies people work like slaves(not Foxconn China level) and have fairly low salaries. Intel has twice as many employees, but then again they're also kind of all over the place.

I can't really comment how important labor costs are, but I don't think that they're insignificant.

so labor costs are insignificant.

When chip technology took off in the 1970s, the US semiconductor manufacturers had probably over 80% of the manufacturing share.

Why is that share so low today? Why is it that the US manufacturers would be lucky to reach double-digits in manufacturing share? How did they get to drop to so low in the first place?

The running cost of fabs is not the only cost of new fabs.

The US government is pouring a lot of money into subsidizing the cost of spinning up fabs. Additionally, China will help western foundries achieve success because they want Taiwan.

> Additionally, China will help western foundries achieve success because they want Taiwan.

Can you elaborate on this point?

I imagine the logic is that if the US has its own domestic chip production and isn’t so dependent on Taiwanese chips, then the US may not care as much if China invades Taiwan.

Terrible reasoning, but yes I think that's what he means

Ignores the fact the island itself acts a buffer to China's SE expansion/projection of power.

There is a lot of strategic value to the island, even without the chips.

That makes sense, and as someone else pointed out, Taiwan has strategic value by itself.

This, the CHIPS act was recently signed. Hundreds of billions are being invested in building onshore fabs.

The US also has a 5 times bigger military budget than the worlds next 5 biggest militaries combined. How do you explain the fact that the US has been years behind China and Russia in terms of hypersonic missile capabilities?

Money alone doesn't make the world go round. Trust does. People work for early stage startups for free because they trust the founders.

When I was researching my master thesis 90% of the papers authors were either Chinese or Indian of origin(probably of US universities). A significant number of Chinese-American scientists are making their way back to China for fear of unfair prosecution. Stigmatization of Chinese-Americans during Covid didn't help either.

I'd be willing to bet that Japans TSMC fab will be done faster than the US's

I'm not making any point on whether the US is likely to be successful or not, truthfully I doubt it, however I'm only pointing out that hundreds of billions of dollars were allocated for onshore fabs.

That's a fair point and a lot more than what Europe is doing. They have all the resources and the money that they need to succeed, but I feel like there is a need for a change in attitude for that to happen.

> How do you explain the fact that the US has been years behind China and Russia in terms of hypersonic missile capabilities?

I tell myself someone just had the clarity of mind to refuse to build a doomsday device (hypersonic warheads breaks MAD, or just becomes a new MAD once everyone has them). Happened before with Project Pluto [0] (among other reasons)

> The program was terminated in July 1964 by the Department of Defense and the State Department as “being too provocative”. It was believed by many that if the U.S. deployed a missile of such awesome power against which there was no known defense, then the Soviets would be compelled to do so [1]

[0] https://en.wikipedia.org/wiki/Project_Pluto

[1] http://www.vought.org/special/html/sslam4.html

Here's a NASA conference paper from 2001 on hypersonics[0]. So it's not that it hasn't been worked on, it's just that it never delivered any tangible results. Why? They spent trillions on planes and aircraft carriers that didn't really produce the expected results in modern warfare. Why is the moneyhose still feeding into that?

The question of morality is REALLY off-topic and always comes up when someone makes a point about the inefficiencies of certain systems, instead of actually addressing the main concern, which is why isn't the money being spent efficiently? And where are the checks to make sure it is?

But the only country that has so far used nuclear weapons and littered the planet with depleted uranium wasn't the Soviet Union was it? That country also threatened to use them in Vietnam[3] and brought the biggest nuclear weapon on its carrier to threaten India in 1971[1]. It also wasn't Russia that unilaterally pulled out of the arms agreements. Bolton just argued that Russia was violating it anyway, so there's no point keeping it[2].

Look, I'm not trying to bash that country, but hold your horses when talking about morals, in a discussion that is ultimately a game of power between superpowers. The reason why the US should stay a superpower isn't because it's so morally superior, it's because, if there isn't some sort equilibrium of power, one side is bound to become a bully. Just like how in a democracy there is an intention to have an equal separation of power which for the last decades has been gradually eroding, giving the executive ever more power, leading to all sorts of problems for society.

[0] https://ntrs.nasa.gov/citations/20020011010

[1] https://www.indiatimes.com/news/india/when-russia-stunned-us...

[2] https://www.npr.org/2018/10/23/659911920/bolton-affirms-u-s-...

[3] https://nationalinterest.org/blog/buzz/us-military-looked-us...

I just said "I tell myself" because its what I'd like to believe. Certainly the US is not in the business of demonstrating good morals.

Thanks for the links, particularly re: India, I wasn't aware of this chapter in the cold war at all.

I'm skeptical of Russia's useful military capabilities given that they're currently losing a war to a smaller country with ~20-year-old NATO tech. Turns out that 5x budget bought a lot more capability than Russia had.

Similarly, China's military hasn't been tested in decades.

Can you elaborate by which metric Russia seems to be losing the war?

Much bigger conventional army, 10 months later has yet to achieve any real progress towards victory.

They may not lose, but at the very least they are doing much worse than anyone expected given the relative disparity in military investments.

Russia is steadily losing the territory that they took in the initial invasion (e.g. Kharkiv, Izyum, Kherson), have made no new territorial gains in months, and have lost more soldiers in less than a year than America lost in the entire twenty-year Vietnam War.

They are not winning! And very much appear to be in the process of losing.

Well, subsidizing EVs made Tesla what they are, so maybe?

Um, you might want to let Intel know that the US is too expensive: https://en.wikipedia.org/wiki/List_of_Intel_manufacturing_si...

By my count, of the 7 fabs on the latest process, 5 are in the US.

Intel has been acknowledged to be having problems in keeping up with TSMC's progress towards 1 nm. They are falling behind.

Marginal costs in semiconductors are very low.

Robotics.

Reply on news.ycombinator.com