ASML dethrones Applied Materials, becomes largest fab tool maker
tomshardware.com
tomshardware.com
https://news.ycombinator.com/item?id=18463181
Apparently their software development processes are terrible.
The price would probably drop slightly if they made 10x more machines, but they would still end up earning more. And the world would be a better place, too.
So I am left wondering ... what is capping their output? What is the bottleneck?
To commit engineering to achieving 10x manufacture rate would probably require giving up the lead on next generation processes.
Personally, I think there must be an end to this. There are steps we are going to learn to more reliably replicate, and reliably reassemble. Machines won't become smaller, but other heavy industries have learned to do faster, de-skilled installations so semiconductors should eventually as well. However, I don't think we'll get the 'good for the world' benefit because by the time every country can make leading edge semis, some other critical, supply-restricted part will become what's contested.
On the absolute forefront, it has to be artisanal.
But we can learn to scale making the machines that were cutting edge two years ago.
Maybe ASML just doesn’t do this because they leave it to other companies? Either way, somebody should be the one that’s always two years behind but makes 2000 machines a year.
And of course, when you remove a bottleneck, you find another. But that other bottleneck will be wider.
And then, these machine are indeed more like physics experiments than a device. If you compare it with that, ASML is actually quite fast.
And the market knows all of this, hence their valuation.
And maybe not two years, but 5, but you catch my vibe here.
Zeiss
And maybe not the cutting edge, but whatever was cutting edge 5 years ago.
Not really that old on this context though.
First, I'd like to say that skrebbel has reacted very gentlemanly to my comment considering how hard hitting it was, my complements to his balance in this. I can only assume he may possibly have been burnt on something he put a lot of effort into, I can relate to that. I've been there too.
But ASML is a very large company and parts of it that I have worked with have very impressive testing capability with a large farm of servers testing all code against regression and progression before any integration with release code. I'm happy to say that I've personally improved parts of this farm in both quality and performance of the testing environment (even though I wasn't part of that group).
The challenge is to be able to maintain and improve upon that when the software is so huge and growing. I poke and push at every opportunity I can. I'm sure that people in the company will recognise this :)
On a personal note and I think a lot of people here can relate to this, you can do a lot for quality simply by recognising the efforts of people correctly, the improvement in motivation translates to improvements in quality. In that area I think things are indeed changing since 2018. Just last week I was very pleasantly surprised to see in a group close to mine the announcement of a new program to do just that. To recognise the efforts of people that go the extra mile in both effort and innovation. Very nice to see.
Now it's best for me to shutup.
AND... now that have your attention. Perhaps I can implore upon the readers here to help me get my new project into the main discussion page of hacker news? Cheeky bastard that I am :)
It's really interesting. It's a sound localization project that combines microsecond Raspberry Pi system clock with sound recording to do really accurate Raspberry Pi based sound localization. I've been trying to get it to the main discussion on hacker news for ages with no success.
Links here:
https://hackaday.com/2023/12/30/localizing-fireworks-launche...
https://medium.com/@kim_94237/tdoa-sound-localization-with-t...
Appreciate it :) I don't take a lot of pride in HN comments and definitely not in ones that I wrote 6 years ago :-)
Appreciate the insights, very cool to see how good it's all getting.
Why? Because they're hot on the stock market since the pandemic? Nokia's SW was the same kind of shitshow when it was dominant. IIRC a long time ago someone on HN wrote here that compiling Symbian OS at Nokia took 2 whole days and Nokia management saw no problem with that.
To me, it's exactly what I expect from a HW company, from personal experience. SW is seen just as a necessary evil, another item on the BoM. Oh, and there's a bunch of useless processes designed as jobs programs to keep some useless managers employed, where each of them needs to review your change and give their green light despite them not being up to date on the technical side for >10 years.
I know this because it's exactly the same at another major Dutch semiconductor spin-off from Philipps I was at in a past life.
Just because ASML is hot right now, doesn't mean they value and employ top SW talent, because they don't sell SW, they sell HW and that's what their customers value, and so ASML values physicists and traditional engineers, not SW devs.
I used to work in automotive and all unit tests didn't require the finished car, just the SW.
Nasa doesn't hava another Voyager probe in their lab floating around in zero gravity to run unit tests on before sending the SW patches, they use simulators.
For calibration you do need the final production HW, but unit tests shouldn't, so maybe there's a confusion here about the type of tests ran.
JPL keeps a working duplicate of their Mars rovers on Earth and tests software updates there. So if the Voyager zero-gee environment prevents a useful duplicate, that means testing software in simulation is considered inferior to using a (massively expensive) physical duplicate like for the Mars robers.
https://mars.nasa.gov/news/8749/nasa-readies-perseverance-ma...
It’s really unique for such a competent machine builder to not have this, and it’s 100% due to their excellence being in physics and mechanical engineering, a culture of treating software as an afterthought for too long. Note, I don't think ASML still treats software as an afterthought, but they did for a long enough to make it really hard to catch up.
I've seen files with 13k lines of if/else/switch, how do you test that shit .-.
As a Dutchman: no, we are not when it comes to businesses software. Dutch companies are very conservative and always choose the safe option. They are modern in that they updated the slogan 'Nobody ever got fired for choosing IBM' to the current century.
It is the main reason that I have only worked at companies with a healthy mix of different nationalities for the last ten years. Doing software in a Dutch monoculture is suffocating.
To me is specific to there because while I’ve had weird experiences somewhere else, it was not as consistent, so in my own experience (guess I need to specify the obvious) it's more likely to end up in weird codebases there than somewhere else, again, in my experience
Was always wondering why people say on their blog that opinions are their own
Slightly kidding, but only slightly because I kind of agree with the idea that people shouldn't be writing software but designing systems. All this big talks about unit testing, management styles etc. and at the end we have this software all around with huge security holes and terrible bugs. Maybe the people partying on Fridays right after pushing untested code to production are having it right. Their machines work.
A bit like us designing cat toys: Although we can design some successful ones by observing cats and trying out different things, more often than not we design it to our liking and the cat will enjoy the package more than the toy itself.
They are just a domestic flight away from Eastern Europe. And just a train ride away from London.
Just because they won a golden ticket selling products based on US research given to them, in a market impossible to enter by new players, is no proof their culture is superior to other companies. It could be they were at the right palce right time.
Pretty sure Nokia also though their corporate culture was superior when the iPhone dropped due to their huge market cap in the phone business.
>Surely they would have opened offices in different countries if Dutch culture was holding them back.
They do have offices is many other countries other than NL.
I don't care about parties either but if I'm going to have to work with you I need to know you fit in a little, creating software is collaborative. In real life, in real companies you are not solving leetcode problems all the time - so why hire based on that? Person A is super intelligent but abrasive and person B is half as smart but super easy to work with. 100% of the time I pick person B.
> having awareness of software principles is really worthless
This is nonsense, you are already expected to know this
> I've seen files with 13k lines of if/else/switch, how do you test that shit .-.
I've seen those too, but don't pretend that's unique to a specific country. There are shitty software developers everywhere.
My wife for example is a doctor. They have cabinets of 4 GP, 1 of them as we found out is a proper sociopath with very unstable personality when things are not perfect. He is driving whole cabinet which employs 10 people to the ground very effectively, wife is running away and hoping it will collapse only after she got out legally. If it wasn't for psycho moves of this guy that cabinet would thrive. While he is consistently being reported as a great doctor by his patients. She is moving into another cabinet where head of it understood it extremely well, and is super picky about people from personality perspective.
People here on HN love stories about experts saving the world, they as experts see themselves in that position. In real life, thats hardly ever the case, most long term problems come from people and not how you solve technical challenges. Once you covered this by far the most important aspect, then of course professional excellence is next step. Never make the mistake of changing the order of those 2, ever.
Having a brilliant misfit at the wrong position can tank a team.
Lets assume we are talking about a reasonable candidate with good social skills, and a higher-than-average tech skills (that would be a charitable interpretation of the original comment).
The hiring process I like to have in place is:
1. HR Interview: see if there is a fit with the company 2. Skill check. Do you have the problem solving, hard skills and communication skills needed for the job. 3. Team fit. Do you like to drink a beer (or non alcoholic) with us and we with you. Can you hold a conversation about something else than work.
6 people are involved. 2 per step. Everyone has a veto. I as a manager do not have special power. I've had people fail most of the time on #3. E.g. a guy that did not talk the woman of the team. She veto-ed. We were really looking for his skills, but he was a jerk.
I don't think you would have passed #3 either. You can have l33t coding skills, but if you cannot make it work in a team it is basically useless for anything of size.
Why is that relevant for work? Are you looking for labor, or for a date/drinking buddy?
Social and communication skills are important but I can be professionally communicative for work ralted topics, and still not want to have a beer with you after work, because it's my free time and I have other interests than you and other things to do.
Unless the company choses to have the beers/drinks during working time in which case by all means, if you pay me to drink on the clock I won't say NO.
This is exactly my frustration
I'm not suggesting should offer your free time to help cultivate this, but in my experience it does help with building a good work relationship with your colleagues - either outside of work or even at lunch - which then improves the outcome of the labor.
Ah I had party at my place, and friends of coworkers invited wanted to have me at their home for weeks before I left the country because thanks to me improved the amount of friends they had despite being born there
I can't tell you all the things where I have had a confirmation that I am great with almost all kind of people.
Except landlords and hiring managers.
Why are you even in a technical interview process?
Even if someone is not capable of having a beer or social interaction, what's the issue?
You're there to punish people who have different personalities, have a different background? Your company is so shallow that it can't digest dealing with someone different? Are you so soft skinned? Is the system now punishing introverts?
Ah as working part of a team, my last manager in NL said they were glad to have me because I was the only one non introvert that would ask questions during meetings. They let me go because I wouldn't go to social events anymore. You know what were the social events? People mimicking italian gestures, talking about cappuccino and pizza with pineapples (I am italian).
Probably the Peter Principle.
>You know what were the social events? People mimicking italian gestures, talking about cappuccino and pizza with pineapples (I am italian).
I chuckled out loud from that.
Yeah, forced social gatherings at work are the worst. The problem is if those gathering are part of the "company culture" then refusing to go will automatically ruin all your promotion perspectives in the company, as most other people also attend just to save face, rub shoulders with upper managers and show conformism.
If you don't go to those shitty gatherings but most people are, then you're signaling you're a non conformist, and rigid companies don't like to promote non conformists.
Overall the NL/BE work culture is not desirable from experience.
As with all large companies there will be great bits and bits that can be done better and the larger the company gets and the more critical the software is the slower some parts will be done. There's always room for improvement. And ASML has improved immensely over the last 10 years. In the teams that I've worked in the caliber of the people has been extremely high. The code reviews are rigorous and that's a good thing. There are a lot of extremely smart people working for ASML. To call it Dutch is interesting, most of the teams I've worked in are international with less Dutch than foreign people in them.
ASML has changed from a much less interesting company when I joined to a very interesting company with respect to the software stack. Yes, I believe there's still lots of room for improvement but that's the case with all big companies.
But heck, I don't know about everything, I've only worked there 12 years.
ASML is impressive because they managed to ship despite the code having been a mess. It’s a mad accomplishment of both engineering and organization (they solve the messy code problem by simply employing 10x the programmers than they might need if the code was better, and somehow that actually works! That’s impressive organization, cause conventional wisdom eg Mythical Man Month suggests you can’t do that)
I think it’s more that not many ASML people are on HN.
EDIT: there’s a commenter named hcfman here in the thread who works at ASML and says that my comment is garbage (which I take to mean that it’s way outdated). Consider updating your impression accordingly, I see no reason to doubt what they’re saying. Last I checked ASML really did want to improve the software situation, looks like they did.
Seems crazy that there are no alternatives for these four companies in the high end AI semiconductor supply chain.
Those 4 companies aren't without competitors. But if you pick the best in each category there is (funnily enough) only one option and right now chaining them together gets a noticeably better product than anyone else can build. That isn't normal but it also isn't that weird for new products.
If you want what Nvidia was building a few years ago there are several of options.
I won't rule out a direct competitor, but they have a substantial moat.
They face some pressure from indirect competition like Canon Nanoimprint though.
There are a bunch of thin bottlenecks and oligopolies in global supply chains. It is a big and complex world. But that leads to a situation where three things are true at once:
* There are companies that are expensive to replace.
* No single company is truly irreplaceable.
* On the cutting edge, it is common enough for there to be only one company that knows how to do something.
There are several companies in this space that are truly irreplaceable. Zeiss is definitely one.
Oh favoritism and bribery! More like, small market that can deal with the fact that the vendor can only promise small deliveries. But also Intel trying to land a whale.
> virtually no fabs today can run without equipment from Applied Materials, ASML, KLA, and Tokyo Electron
> ASML’s China Sales Surged Despite Secret Dutch Deal With US
https://www.bloomberg.com/news/articles/2024-01-25/asml-s-ch...
Edit: note to self, don’t beer and then HN.
> Secondly, ASML could sell sophisticated tools to Chinese customers for most of calendar 2023 as sanctions against the Chinese semiconductor sector kicked in only in September and only for one tool. By contrast, sales of tools to Chinese clients by Applied Materials were, to a degree, impacted by the U.S. export rules introduced in October 2023.
Maybe control-f is not the best strategy here.
In the long term you don’t think the stock price is a function of a company’s success? Surely while a a company does very well financially and the financial outlook is very good, the stock price would be higher than when the fortunes change?
That’s because there’s no guarantee the company will thrive in the long term, nor that you didn’t overpay to begin with. That does not change the fact that share price is a function of the business’ financial success.
Also I've been wondering why our Prime Minister has been so enthusiastically following US foreign policy lately. Turns out he's eyeing the NATO boss seat, and the US has the most important vote.
I think we do have slightly more influence on the US than our size would merit.
There is some truth in parody ;-)