It is a bit of an odd article, they spend half of it describing this as a remarkable turn of events, and the other half describing why it isn’t really an apples-to-apples comparison and how the companies aren’t in direct competition.
I've lived in NL for a while, few years, changed few companies, it's a place where processes are more important than actual code quality, where people stay in the same job for 10-15 years and everything below that is seen as a bad thing, people is hired for the culture primarily, having awareness of software principles is really worthless. I ran away as the last company I interviewed with, after solving a coding puzzle in 30 seconds, finishing in 15 mins something which had to take 45 mins, asked "Are you available to come at the office on Friday for company parties?". So I can only have nightmares about software practices in NL companies
I've seen files with 13k lines of if/else/switch, how do you test that shit .-.
I'm a hiring manager in the Netherlands. I would be happy that you ran away, because you would not have passed the bar.
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.
> Do you like to drink a beer (or non alcoholic) with us and we with you.
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.
It might be relevant, because labor isn't just about working in a silo. A significant part of work might be done in a team setting, where communication, feedback loops, constant interaction, collaboration, code reviews, team spirit, and more are important.
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.
I have been my whole career a great team fit, I’ve had bosses inviting me for Christmas at their home when I couldn’t travel to my home town, I’ve had coworkers inviting me to their child baptism and their wedding, I only don’t pass the bar for social interaction when its enforced by clowns out of context
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).
>Why are you even in a technical interview process?
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.
And this kind of nonsense is basically why EU tech scene will forever stay uncompetitive. I find it interesting that ASML got to where it is from the remnant of some Philips labs but I’m not sure they will manage to stay in top for long if some other tech center decide to go and build something competitive.
Overall the NL/BE work culture is not desirable from experience.
Good thing that soon all the software would be written by AI then.
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.
Haha right, but maybe you still want to have some control over it - just to be sure that it fits you first. Calibration and quality control if you will. At some level
High enough the thing becomes art because you have the have empathy which is something that machines can only pretend since they are operating differently than humans.
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.
This is a company with over 40,000 employees and a market cap of over $100b. Surely they would have opened offices in different countries if Dutch culture was holding them back.
They are just a domestic flight away from Eastern Europe. And just a train ride away from London.
>This is a company with over 40,000 employees and a market cap of over $100b.
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.
This seems like a very one-sided opinion/experience. While I've certainly had my own share of odd experiences and heard similarly odd stories, this doesn't really strike me as specific to The Netherlands as a whole.
Of course I can only express one sided experiences, did you think you were reading a post from Him? :D
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
I'm going to defend this method of hiring a little bit. To me, having at least a bachelor of IT already proves you have software engineering skills. The coding tests are just a quick check if you haven't lied or anything. The first month is trial basis anyway (both ways) so if the candidate is not on-par technically it's a quick goodbye.
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.
You interpreted the interview in the most charitable way, and the commenter in the least charitable way. Why do you think they are possibly abrasive? Why do you think asking about Friday night parties, drinking alcohol or swearing (happened to me) is a high-value signal?
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).
You chose social convenience and niceties over quality work. That’s fine for an established company with secure revenue but in the long run it’s not going anywhere.
If the less nice but more competent people gets hired by one of your competitors you basically provided the ammo to kill your business. But I guess whatever, most of the timer hiring managers are found in companies who stopped evolving.
Yeah this is typical junior code ninja opinion. Folks with 20+ years under our belts know damn too well how humane aspect is more important than literally anything in long term.
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.
Completely agree. And as someone who didn't fit culturally as often as I did fit, this is a two way street (ignoring the very extreme ends of the spectrum).
Having a brilliant misfit at the wrong position can tank a team.
> I thought that the dutch would be more ... modern?
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.
It's a pretty large gap between enterprise organizations that have lost the capability to do anything but buy mostly terrible software in the open market and organizations that maintain some amount of ability to make things. If you're at one of the former, expect to be locked into a digital landscape that's decades out of date.
Tbc, based on the comment that looks to be driven almost completely by the fact that they are making a tiny number of incredibly expensive bespoke machines, so they can't just keep extra machines lying around to do unit tests with (both because each machine is expensive, and because the results wouldn't necessarily generalize to other machines). So it sounds like a constraint of the field, not bad choices.
Why do you need the actual production machine for unit tests? Why can't you build a SW simulator for them and run unit tests on them?
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.
> 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.
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.
I don't work on the machines but I can provide you an example on why this wouldn't even work for simple machines like CNC machines. You could tell the machine for instance to move the drill out of boundaries, too hard, run it into a wall... All things that you can't catch with a software only unit test because the machine you're sending commands to doesn't really have any idea what else is going on in the space within it. So you can give it commands which are fine software wise but once they reach the real world, they can break stuffm
Many other machine builders in the area have large parts of the hardware replicated as a software simulator so they can test the code as much as they wish. It’s not uncommon to begin load testing the software (incl lots of error scenarios, simulate a robot arm getting stuck etc) before the first prototype of a machine got put together.
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 love these threads. Everyone will chime in about how they know better how to run the company, with suggestions on the processes that they don't know and so on. Meanwhile the story is "business is #1 and booming". Elegant software stacks aren't always the most important part of the business.
Exactly. ASML made a profit of about $2,050,000,000,- in 2023. I doubt they are very interested in complaints about how some random dude expected more time to run his tests.
I think they are pretty serious about high quality software development. There are often questions and concerns from (tech/non-tech) management about build times, compute costs, software product’s performance etc.
PS:I work at ASML.
No, most of them want to make a ton of money. They're hanging out virtue signaling their cleverness on a forum owned by a Silicon Valley startup incubator for a reason.
Note that I wrote that in 2018 based on hearsay from well before that (I live in the area and have friends who worked at ASML at the time). Things might’ve changed in the ~10 years since then (for better or worse!).
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.
i advised one of their software-making vendors on software-organisation+culture and similars, about 10y ago. It went pretty well, but the funny thing was, The tools and environment they had to use (because of ASML) to check/test stuff, was same as it was in Motorola 25 years ago (Sun CDE /Common Desktop Env? something like that). I felt like dropped in a time machine :/
From that comment I concluded that software patching speed isn’t a bottleneck there. It sounds frustrating but to be terrible it needs to significantly impede the product or business relationships, which isn’t at all happening.
Stability and long product lifecycles are incredibly important to them. They're pretty much the opposite of move fast and break things or deploy early/deploy often.
Absolutely. What makes a lot of sense for one organization would be absolutely insane for others. ASML has incredible innovation in some areas but I'd expect them to prioritize stability in others.
Honestly, if you think about the availability guarantees that these FAB's have to deliver, I'm kinda surprised they don't run off of IBM mainframes. Can't believe it but it would be a perfect use case. High bandwidth I/O, paired with ungodly, over-the-horizon-radar like availability.
I find it strange that they are making so few machines.
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?
My understanding is that production and installation do not scale since the machines are essentially replications of a successful physics experiment environment. Manufacture is a long, multi-step process requiring technical expertise (which is limited). Then, machines are re-assembled at the customer site--again, carefully and with a large technical staff.
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.
Asml does this, learning how to scale the cutting edge of two years ago. This is what scaling cutting edge physics looks like. Academic groups sometimes take the better part of a decade (or more) to replicate a setup by another group. There are few components you can buy off the shelf in this space: you'll have to understand, design and make them all yourself. Market's way too small to suppory competing suppliers. Takes longer than two years to farm all of that out and create a market of sufficient size. Plus, buyers tend to not want (or pay) for one or two decade old stuff, so that market (for part suppliers) may never quite materialise.
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.
But they are! It just takes way longer than two years, both to scale production in house (creating the first takes way longer than than, creating the second in 2 is already fantastic), and to create any supply chain market which others could use to compete. Canon surely has been trying, and failing, because it's hard.
Part of the answer was in previous comments: they are relying too much on “nice” humans instead of rationalizing processes.
This is very common in the EU and in fact trying to do stuff that will lead to less people needed for a given output will quickly get you in trouble.
The other day there was someone calling EU communist and on a “culture” point of view this is pretty close.
We seem stuck on old paradigm of everyone has to do something even if it cause more problems in the long run.
I believe they could absolutely produce more machine if they had too but they don’t have to care since they have no real competitors on the high end stuff so this is what we get…
There simply aren’t enough people that can do the job and not that many Germans are interested in taking an apprenticeship so they can polish glass for the rest of their life.
There's not going to much meaningful pushback on this comment because any (ex-)ASML employee is going to be super hesitant to share anything. So feel free to think what you want.
Are you sure? Every ASML person I know loves sharing these stories at parties. It’s not a secret or anything. Stock wont go down if people find out that the code is a mess, everybody who cares about this already knows it.
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.
The comments in that post are full of untruths. ASML is a huge company so no one person can talk about all of it.
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.
Yeah there are a lot of silos. Partly that's by design, partly that's a an arguably bad part of it's culture. I'm in a position where I can see a lot of them work in isolation, all choosing to use their own tooling: programming languages, frameworks ides, version control systems... Indeed no one team works quite like the other.
Do you have other insight? A counterpoint? Anecdote that supports a different narrative? Anything, at all, that would contribute to the conversation and support your "garbage" comment?
It's not accurate as I said I other comment. Software is tested not just on the machine nowadays. I think you'll find most asml employees hesitant to share more, I'm not sure what I can say either.
Okay I'll bite. Usually with large and important companies like ASML one has to be very careful with what you say. Usually it's better to say nothing at all. This time I felt I had to jump in and defend ASML as it was painting rather a misrepresentational picture.
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.
I totally get that you need to be careful about what you say, and I especially appreciate you taking some extra time to write this out. This comment is 1000x more interesting than "that story has a lot of garbage".
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.
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.
a bunch of useless processes, if only.. they started using Scaled Agile (or SAFe) a few years ago, which slows down software development by a factor of five or more and adds non-productive fulltime roles such as 'Release Train Engineer', 'Scrum Master' and 'product owner'.
They are physicists and hardware folks, right? I would expect messy code written by convoluted processes for nightmare applications, that somehow works.
I expect that if their costumers had to choose between ASML keeping machines around for software testing or shipping every machine the can build, they would want the hardware.
That perspective is a little wrong; those are the 4 companies on the edge of a wave of technological improvement. It isn't unusual for there to be unique companies in those circumstances. For example when Apple hit the market with the iPhone back in '07 there were similar things going on. One of the things that hamstrung Apple's competitors was that nobody in the supply could manufacture equivalent parts.
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.
Zeiss and ASML currently have no competitors in the EUV lithography space as everyone else dropped out. (There was a big switch from lenses to mirrors)
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.
Another way of framing your opening is "a bunch of other companies are on the verge of figuring this out, but the market isn't big enough". EUV hasn't even been the done thing for all that long, I gather it is only the tech of choice for about a decade.
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.
When the MacBook Pro switched to Intel processors, Intel made an entirely new top bin for Apple and for something like six months they had exclusive right to it.
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.
> 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.
I don't want to brag, but buying ASML stocks was probably the smartest financial decision in my life.
Also, AMSL & AMAT aren't competing in some areas. AMAT provides tooling & services that assist ASML.
Same, I sold at the peak. Don't hold on to your winners for too long, the stock price is not entirely a function of the company's financial success. A lot of it is just on whims and cycles.
> Don't hold on to your winners for too long, the stock price is not entirely a function of the company's financial success
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?
Sure, but there is no guarantee long term stock price is above what you paid for it. Also, the market can be unreasonable, in both directions, for quite a while. Dividends are what you epupd be looking for in long term stock investments.
> but there is no guarantee long term stock price is above what you paid for it
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.
A function? Yes. Realistically linked? No, absolutely not. Added caviat: expected financial success, current and past success are either irrelevant or already priced in.
The US is not far from doing that. See the Hague Invasion Act.
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.
Well there are various degrees of enthusiasm. Netherlands also claims to be a fighter for human rights, yet it immediately dropped UNRWA funding after unproven allegations about a handful of people (out of thousands) and thus letting civilians starve to death...
Is there any need? We licensed them specifically the tech because it would have been unreasonable to have given it to the established players and because they bought SVGI.