I come from a CS background. If you leave school and enter the workforce at SAS, you will be at a disadvantage insofar as you are looking for a workplace that you hope to learn something from or one that you hope reinforces good engineering practices. The semiconductor industry, from the vantage point of an experienced systems thinker, reinforces awful engineering practices. All the things that SE folks take for granted—version control, e.g, as a discipline/tool to name one example, for regression analysis—say goodbye to any of these things, along with any expectation of competence from your colleagues.
The semiconductor industry sources heavily from the defense sector, and this has deleterious, catastrophic effects.
A close friend (from a non-CS engineering background) I know to still be at SAS. We don't talk about finances, but at around 8 years in, I estimate that he just recently (since COVID) crossed the six figure threshold. He's in a fairly senior role. Meanwhile: "What's x86?" — a direct quote from a conversation not too long ago.
He has no real professional experience outside of SAS, was not set up in an environment that rewards "extracurricular" experience, so he doesn't have any of that, either, and he's effectively stuck as a result of the "golden" handcuffs that the path at SAS has led him on.
Don't work in the semiconductor industry, folks. What you can expect to be is a monkey pressing buttons on a machine (that your supervisor—and their supervisor—doesn't understand, and nor will you; you pass butter). Make no mistake: there's an abundance of low-hanging fruit. But you won't have, as Gerry Weinberg put it, "permission" to pick it.
By all means, if your prospects/ambitions look to be something like working for the state/federal government, learning nothing over your career, and cashing mediocre checks for doing mediocre work along the way, then by all means go for it. Otherwise, it is likelier that you'll be hobbled by your time in the industry. And not for no good reason. If I'm evaluating someone and it came up that they'd done any significant time in the world of semiconductors, the net impression it'd have on me (having hash an insider's view of it) would be one worse than if they said that they'd just gotten their degree and had spent the last N years flipping burgers, cf Dan Luu.
Don't do semiconductors, folks.
PS: sorry, drunk.
In fact, an entire generation of SWEs who forgot the concepts of simplicity and efficiency, coding a never-ending avalanche of rubbish bloated apps is why we continually need new silicon.
Fab engineers in the US are fairly disposable because their existence isn't based on market needs but really as a hedge by the US to be independent. That's going to suck for you because TSMC isn't going to value you that much because you're not really business critical, you're just there to get government subsidies. In Taiwan it's another story and it's almost competitive with good SWE jobs.
Why do you have this idea ? I'm not sure it sounds more "just" than saying, for example, that tasks that are physically more challenging should pay more, or humanly more challenging should pay more, etc..
The truth is we don't need fab engineers because the applications we need to build require already "solved" hardware problems which are mostly supplied en-masse by a the few, large companies.
But, the deeper truth is that we don't need any of it... I have been recently and somewhat pleasantly surprised by how little I truly need. E.g., a 200$ chrome-book satisfies most of my needs. I have much, much beefier computers, hundreds of gigs of ram and multi-core processor, etc...all a waste. (I work as a software engineer and regularly find myself maxing my cores for things I don't really need).
If I could build my own hardware and have it actually do things, man, that would _useful_. Unlike this giant rickety tower of babel we have built...
Recent front page submission comes to mind XD
It's okay just to say that you like the money. You don't always have to use the largest words. It's also okay for someone to say that software engineers are paid lots of money. Perhaps too much.
-> how about starting a semiconductor company where people are paid more?
I'm assuming you're being facetious given the enormous logistical and financial costs of starting a somewhat up-to-date semiconductor company. But, maybe that's the point.
I'm not an SWE.
> You don't always have to use the largest words.
Some words have specific meanings, rather than just being longer synonyms for short words.
> Perhaps too much.
Again, no. This was my point — this is like saying that the solution to homelessness is to take away the homes of everyone with homes, so that everyone is now equally homeless.
Don't pay the people who are already getting a good deal less; pay the people who are currently getting a bad deal more.
> I'm assuming you're being facetious given the enormous logistical and financial costs of starting a somewhat up-to-date semiconductor company.
No. Just fucking do it. Start another company that's less complex but is an obvious money earner; get rich; reinvest your wealth into a semiconductor company. Should only take 10 years. If you care, do it.
If Elon can do it — and Elon is not the smartest man — you can do it too.
(To be clear, you don't have to start a company that competes in the wider semiconductor market; you just have to start a company that competes in a niche. Like, for example, 5G radios. But your goal does have to be to compete; getting acquired means you no longer have control over your own business strategy, incl. profit margins, and therefore can no longer make your company "the one where people get paid well.")
I can agree with this, definitely. I still don't think that it's that easy to start a semiconductor company for the average person considering the requisite expertise and scale of capital necessary. However, outside of the basic accounting, it's not really my area of expertise.
You're correct that their software practices, by and large, are absolutely atrocious. In general, I found that if a team had people who could code, they'd just write their own programs and stay in the Shadow IT realm. I taught myself JS/PHP/SQL while I was there, and built a shitty CRUD app to replace some even worse Excel sheets others were using.
However, from SAS' perspective, they aren't making software. All of the machines have their own control systems (which are usually running XP and are terrible), and they need tools that can connect to those, retrieve data, dump it into a DB, and run queries on it. It never needs to leave the site, you don't have to worry about surprise traffic, and scaling issues are never going to get to the point of needing thousand-node clusters or the like. What they have works.
Re: pay, IME engineers of any kind would start at ~$80K fresh out of college, which in 2017 wasn't that bad for Austin. Engineers also got promoted very quickly. Yes, pure software jobs rapidly outpace SAS salaries. Hell, I went from a supervisor role making $86K to an Associate SRE role (I have 20 years of hobbyist Linux experience, I just never did anything with it - also I got a Master's in SWE while I was at SAS) elsewhere, and immediately went past $100K.
It's a fascinating company that is also incredibly stressful, the pay isn't as good/fast as others, and to my knowledge remote work doesn't exist. Still, someone has to do it, else we wouldn't have chips. Also, they employ some _crazy_ intelligent PHds. They just exude brilliance. It's awe-inspiring watching them talk through problems.
Semiconductor fabrication is the most advanced engineering humanity has ever attempted, so you will reach your threshold eventually solving problems that might not have answers (depending on your role). Not everyone wants to do that though, and the world needs implementers and those focused on details.
Pay? Could be higher for sure.
Their friend also stayed there for 8 years while the poster presumably quit sooner.
For the vast swathes of the country where the available jobs are some mix of retail/warehouse low wage drudgery, long haul trucking, drug dealing / fraud / other crime, or welfare, working in semiconductors might be a welcome and attractive option.
I don't know, I have seen numerous PHDs who can't competently write any piece of code, I'm beginning to think it's a feature, not a coincidence.
So for the set of people who can be fab engineers, live in the USA and speak English, yes, the FANG engineer option is always there.
Here’s an example: https://www.ziprecruiter.com/c/HRL-Laboratories/Job/Wafer-Pr...
I remember a professor sympathetically talking about how little clusters of people just split up the homework load together like it was a team project from random classes and then tell each other how to solve it to each other to save time.
CS was just another science degree in comparison.
You might not get an interview at any given company in a particular year, but you should be able to get some interviews over a couple years if you have decent GPA or work experience. You probably won't be placed on a top-notch team, but you'll get in.
Could the tooling be better? Sure. Our version control system is pretty old, but the company has made a massive push over the last few years to modernize all of it, and when we transition to that at the end of this year I think we'll be in a pretty great spot. Some teams are already using the new tooling and it has gotten very positive reviews. I have a great relationship with my team and my manager. I get to work on interesting stuff (mostly embedded systems), and I've had the opportunity to learn a small part of one of the most advanced machines humans have ever built.
I'll say the thing about getting people from defense is true, but not so much in software. If anyone has other questions feel free to just ask.
But going back to your point, there are certain areas within semiconductor manufacturing that are software engineering jobs. They tend to be closely related to what's called physical design in the design part of making a chip. Or EDA coders. It's a cool area. But the bigger part of semiconductor manufacturing is working in a fab, which is atrocious grueling work, which is quite boring and dangerous.
Better than most people.
When I mentioned Dan Luu, I was specifically referring to the recent thunderstorm on Twitter where he mentioned the uncanny valleyness of doing work for unsexy companies and that it's actually more like a blemish on your résumé than a non-zero positive contribution.
What does this part mean, exactly? Are photolithography lasers just repurposed Naval anti-aircraft lasers or something?
-> He has no real professional experience outside of SAS, was not set up in an environment that rewards "extracurricular" experience, so he doesn't have any of that, either, and he's effectively stuck as a result of the "golden" handcuffs that the path at SAS has led him on.
-> Don't work in the semiconductor industry, folks. What you can expect to be is a monkey pressing buttons on a machine (that your supervisor—and their supervisor—doesn't understand, and nor will you; you pass butter). Make no mistake: there's an abundance of low-hanging fruit. But you won't have, as Gerry Weinberg put it, "permission" to pick it.
-> By all means, if your prospects/ambitions look to be something like working for the state/federal government, learning nothing over your career, and cashing mediocre checks for doing mediocre work along the way, then by all means go for it. Otherwise, it is likelier that you'll be hobbled by your time in the industry. And not for no good reason. If I'm evaluating someone and it came up that they'd done any significant time in the world of semiconductors, the net impression it'd have on me (having hash an insider's view of it) would be one worse than if they said that they'd just gotten their degree and had spent the last N years flipping burgers, cf Dan Luu
All I'm understanding from this is not "don't go into the semiconductor industry" but rather that you're an awful friend to be trash talking your friend. It seems like you just called their career a waste of time and implied your friend to be more poor (since when was 6 figures not a decent amount of money?) than you and less knowledgeable ('what is x86'). I'm sure you'll agree that your friend is neither poor nor stupid, as you seem to imply here for their working in the semiconductor industry. I'm sure your friend would be delighted to know that they're a monkey pressing buttons without any knowledge of what's going on or that you estimate a fast food worker has a better background for transitioning to software engineering.
You haven't talked to your friend about finances, but you can magically estimate their savings? I understand that you might be able estimate within an order of magnitude based on context, but you don't know how much this person has saved nor do you seem to consider that they might spend their money on personal emergencies, sick family members, or kids. That your estimation of your friend's savings should be taken as evidence for poor compensation in the semiconductor industry is laughable, honestly. But, hey, I at least appreciate your earnest opinion.
You're right. I appreciate the correction. I didn't think of that as referring to the salary.
Their latest equipment had some issues that were only showing up on production lines, so they wanted engineers to babysit the equipment at the fabs, and write patches on the spot for any issues observed. They repeatedly stressed how much money their customers lost for any downtime, and how there must not be downtime.
This would be an interesting employer for many trades.
It's a very high-energy place with more going on than you can ever hope to keep track of. Even if the technology doesn't interest you, the cultural interactions should. You could have some days with Japanese, Koreans and Americans translating 3-ways in a conference room. Opportunities for business travel are probably restricted more now, but when I was there they practically begged us to travel to the Korean sites in hopes that we would learn more faster.
Some roles can be far more exciting than others. Someone has to work security and manage the facilities. That said, all the important roles seem to be the compelling ones too. I don't know of anyone who had a hard time getting another super cushy job after time at Samsung, so it's also great for long term career prospects.