Taiwan Semiconductor Manufacturing Company (TSMC), one of the world’s largest advanced computer chip manufacturers, continues finding its efforts to get its Arizona facility up and running to be more difficult than it anticipated. The chip maker’s 5nm wafer fab was supposed to go online in 2024 but has faced numerous setbacks and now isn’t expected to begin production until 2025. The trouble the semiconductor has been facing boils down to a key difference between Taiwan and the U.S.: workplace culture. A New York Times report highlights the continuing struggle.
One big problem is that TSMC has been trying to do things the Taiwanese way, even in the U.S. In Taiwan, TSMC is known for extremely rigorous working conditions, including 12-hour work days that extend into the weekends and calling employees into work in the middle of the night for emergencies. TSMC managers in Taiwan are also known to use harsh treatment and threaten workers with being fired for relatively minor failures.
TSMC quickly learned that such practices won’t work in the U.S. Recent reports indicated that the company’s labor force in Arizona is leaving the new plant over these perceived abuses, and TSMC is struggling to fill those vacancies. TSMC is already heavily dependent on employees brought over from Taiwan, with almost half of its current 2,200 employees in Phoenix coming over as Taiwanese transplants.
This last job (I’m a contract employee) will be my last in MFG. I was hired long term (2 years) to get a gsk/haleon site to add almost 40% more deliverables. 280 million units a year to 400 million. Reduce waste by 25%, CoA/CoE turn around down to 2 weeks from 6.
The labs, which operate almost entirely as a community (eg no real rigid structure, lots of senior empires) killed it. 7 day turn around which honestly now my mind. Packaging was a struggle once i pointed out that OEE can be improved by scheduling downtime rather than just oopsing it (strictly beancounter bullshit).
Manufacturing… Took my ideas, literally threw them in the trash in front of me and said they have experts from multiple countries, they don’t need my help. Cool, i still get paid so whatever. You wanna see the biggest dumpster fire ever… Laid off about 40% of the mfg work force, rolled out some bullshit trainings about operators and maintenance working to bring equipment “back to new” whatever the duck that means (means maintenance budget is gone) all while investing 0 dollars into repair and maintenance. Gear boxes leaking oil into overflowing catch cans for months. Vacuum traps actually pulling ingredients out of the batches, building more systems upon systems that they can’t validate. Cleans that won’t pass swabs, cleans that aren’t validated, processes that rely 100% on operators to transcribe SCADA data into an electronic batch record system.
Never seen anything like it but i know when a horse is dead and this one was dead before i got there.
As a software engineer, this is exactly how software works.
Everything is just a huge mess bolted and duct taped together, sometimes over decades. And it’s all way too complex to understand and crap like crowdstrike happens.
You can’t rely on anything anymore and I’m pretty sure, our highly interdependent world will come very close to collapse if anything major happens. Covid was a warning shot, but nobody heard it.
I don’t think there will be a collapse just because there is no meticulous maintenance or development. Most likely, in the future there will just be an accident or tragedy that will improve standards and safety.
If you want a collapse you have to pray that all the factors attack at the same time, because if only one does the attack they only strengthen humanity see Late Bronze Age collapse.
Look at crowdstrike. A tiny error disabled millions of computers for hours. Think about what would have happened, if this wouldn’t have been an error, but an actual attack.
Look at the supply chains of medical supplies. One major outbreak of some bacterial disease in India or China will lead to them stopping exports and since so many pills are produced there, a huge drop in global supply.
Look at the undersea cables. There are not that many and capable malicious actor could easily destroy a lot of them.
Look at the power grid. I don’t know about other parts of the world, but the European grid, spanning pretty much all countries in Europe plus turkey, has no plan for a cold start. If it breaks down, there’s gonna be blackouts for weeks.
Of course, none of that will end society, but that’s not how collapses work anyway. One event triggers another, and the combination leads to the collapse itself.
It is inherent risk and it is present everywhere. Just because there are bugs every now and then doesn’t mean there is a crash and you should also know that Linux was almost screwed by a backdoor that XZ Utils had, it doesn’t save open source.
The only thing you can do is to reduce it and if you don’t take precautions you will increase that risk.
Remove meticulous. I’m watching 15 year old equipment fail on a daily and the solution is to keep running it. Failing to plan is planning to fail no matter what mental gymnastics you pay a consulting firm to do for you after your layoff 10% of your open office mouth breathers and 50% of your neck down workers.
Businesses seem to have really gotten caught buying their own bullshit. If the numbers are so good and your OEE is so good, you don’t need that labor overhead. So they reduce the headcount. Problem is the numbers are all made up and someone whose ass was in the fire is now maybe safe for another few months. Multiply that by a few other “engineers” or whatever intern they pawned serious work off onto and you get a lemon.
I’ve been doing this a very long time and I’ve seen business struggle for all sorts of reasons. No one’s trying to steer the cart away from the cliff anymore. Why admit fault and get laid off when you can bullshit and get laid off? That’s worst case. Bullshit and keep your job? Gravy.