3 ms·

I worked in semiconductor manufacturing. The hardest stuff is troubleshooting the state of the live system and performing root cause analysis. In my time at Sam

I worked in semiconductor manufacturing. The hardest stuff is troubleshooting the state of the live system and performing root cause analysis. In my time at Samsung, we politely referred to the facility as "bootcamp for engineers". Indeed, I learned how to troubleshoot incredibly complex systems after just a few months of direct exposure. The #1 skill I learned was to not ever make an assumption about a thing. "Trust but verify" was burned into our souls. Perhaps the best way to train the next generation is to create actual bootcamps on-site where young professionals get a chance to try and answer questions like "How did lot ABC123 end up with unexpected CU exposure while it was sitting inside a N2-purged storage area?" The only tools they get to use: historical logs of every event that occurred and a copy of grep. In my experience, the deep PhD-style education was absolutely not essential for the day-to-day. Only a few people need to master the idea. Once the complex academia is converted into a physical reticle or CMP process recipe, you then need an army of people to reliably execute on that standardized vision. I think someone coming into semiconductor manufacturing would have an easier time if they were dropped right into it on day one. The theory is nice but we found it to be mostly a waste of time when someone from manufacturing was calling us on the phone and complaining about a stuck lot.

josaka 4y ago
Same fab, same experience. I would add that a deep understanding (and ability to explain) statistics was more critical than understanding the underlying device physics for most roles I encountered. Enormous amounts of noisy data, characterizing several hundred process steps, is harnessed to apply massive pressure on engineers to "fix" problems that might just be a statistical blip. Economic stakes of minor yield deviations are so high that you often have to act as if something might be wrong before signals reach statistical significance. I bailed after four years, but it was great experience to start my career.
drivers99 4y ago
> The #1 skill I learned was to not ever make an assumption about a thing. "Trust but verify" was burned into our souls. There's a book I like called "Stop Guessing" (Nat Greene) about that.
- TeeMassive 4y ago
  Link?
  - drivers99 4y ago
    The name and title should let you buy (or borrow) it, but I just found this preview: https://books.google.com/books?id=umXEDQAAQBAJ&printsec=frontcover#v=onepage&q&f=false https://books.google.com/books?id=umXEDQAAQBAJ&printsec=fron... Here's the US Amazon link: https://www.amazon.com/Stop-Guessing-Behaviors-Problem-Solvers/dp/162656986X https://www.amazon.com/Stop-Guessing-Behaviors-Problem-Solve... Not affiliated.
gumby 4y ago
> Perhaps the best way to train the next generation is to create actual bootcamps on-site where young professionals get a chance to try and answer questions like "How did lot ABC123 end up with unexpected CU exposure while it was sitting inside a N2-purged storage area?" In other words, apprenticeship. I strongly agree, there should be much more of it in our industries.
- mrguyorama 4y ago
  Companies used to invest in their workers, but after MBAs convinced themselves that workers are just replaceable cogs, nobody wanted to front that expense, so they've spent the past fifty years crying that colleges don't prepare people for jobs, as if jobs have completely transferable requirements, or as if the point of college is to get a job.
  - anon291 4y ago
    Actually companies stopped investing in workers and outsourced education to colleges because the US courts ruled that aptitude tests (the means by which companies formerly would select people for jobs without requiring college degrees) were racist. While not banning them per se, the decisions made it such that suddenly, businesses could now be held liable if the test produced 'discriminatory' outcomes. This is the case: https://en.wikipedia.org/wiki/Griggs_v._Duke_Power_Co https://en.wikipedia.org/wiki/Griggs_v._Duke_Power_Co. Since then (whether you agree with the ruling or not), companies have been reticent to use aptitude tests. Instead they outsource selection to colleges, which conveniently do the exact same thing as the SC was trying to end. Except now, the companies can foist the cost of training onto the employee, since the expectation was no longer that you could get a job out of high school and learn on the job. Unfortunately, the case IMO has had deleterious effects. It should have been expected, given the racial dynamics of the time, that any aptitude test may have produced unequal results. However, the proper response is the government improving the education system so that these disparities disappeared, rather than making them illegal. As this article itself states, if an equal mathematics aptitude test were run today by a company seeking to train up engineers and other STEM-inclined workers, then we would expect the test to produce racially disparate results (because according to the article Asian and white parents, who are better off, invested in private tutors for their children). No company will do this test then, because it will certainly be found to be discriminatory once they get the results back, and suddenly they're potentially on the hook for millions. Thus, companies instead outsource the selection to colleges who suddenly manage to whitewash the fact that the graduates of their STEM programs are disproportionately asian (and to a lesser extent, white). But this is why apprenticeships are all but impossible in the US. Without the college degree to exculpate you, you will be certainly be found to have discriminatory hiring practices if you are picky with your entrants.
meh8881 4y ago
Sounds like a good opportunity for ChatGPT if it’s log based.

josaka 4y ago

Same fab, same experience. I would add that a deep understanding (and ability to explain) statistics was more critical than understanding the underlying device physics for most roles I encountered. Enormous amounts of noisy data, characterizing several hundred process steps, is harnessed to apply massive pressure on engineers to "fix" problems that might just be a statistical blip. Economic stakes of minor yield deviations are so high that you often have to act as if something might be wrong before signals reach statistical significance. I bailed after four years, but it was great experience to start my career.

drivers99 4y ago

> The #1 skill I learned was to not ever make an assumption about a thing. "Trust but verify" was burned into our souls. There's a book I like called "Stop Guessing" (Nat Greene) about that.

TeeMassive 4y ago
Link?
- drivers99 4y ago
  The name and title should let you buy (or borrow) it, but I just found this preview: https://books.google.com/books?id=umXEDQAAQBAJ&printsec=frontcover#v=onepage&q&f=false https://books.google.com/books?id=umXEDQAAQBAJ&printsec=fron... Here's the US Amazon link: https://www.amazon.com/Stop-Guessing-Behaviors-Problem-Solvers/dp/162656986X https://www.amazon.com/Stop-Guessing-Behaviors-Problem-Solve... Not affiliated.

gumby 4y ago

> Perhaps the best way to train the next generation is to create actual bootcamps on-site where young professionals get a chance to try and answer questions like "How did lot ABC123 end up with unexpected CU exposure while it was sitting inside a N2-purged storage area?" In other words, apprenticeship. I strongly agree, there should be much more of it in our industries.

mrguyorama 4y ago
Companies used to invest in their workers, but after MBAs convinced themselves that workers are just replaceable cogs, nobody wanted to front that expense, so they've spent the past fifty years crying that colleges don't prepare people for jobs, as if jobs have completely transferable requirements, or as if the point of college is to get a job.
- anon291 4y ago
  Actually companies stopped investing in workers and outsourced education to colleges because the US courts ruled that aptitude tests (the means by which companies formerly would select people for jobs without requiring college degrees) were racist. While not banning them per se, the decisions made it such that suddenly, businesses could now be held liable if the test produced 'discriminatory' outcomes. This is the case: https://en.wikipedia.org/wiki/Griggs_v._Duke_Power_Co https://en.wikipedia.org/wiki/Griggs_v._Duke_Power_Co. Since then (whether you agree with the ruling or not), companies have been reticent to use aptitude tests. Instead they outsource selection to colleges, which conveniently do the exact same thing as the SC was trying to end. Except now, the companies can foist the cost of training onto the employee, since the expectation was no longer that you could get a job out of high school and learn on the job. Unfortunately, the case IMO has had deleterious effects. It should have been expected, given the racial dynamics of the time, that any aptitude test may have produced unequal results. However, the proper response is the government improving the education system so that these disparities disappeared, rather than making them illegal. As this article itself states, if an equal mathematics aptitude test were run today by a company seeking to train up engineers and other STEM-inclined workers, then we would expect the test to produce racially disparate results (because according to the article Asian and white parents, who are better off, invested in private tutors for their children). No company will do this test then, because it will certainly be found to be discriminatory once they get the results back, and suddenly they're potentially on the hook for millions. Thus, companies instead outsource the selection to colleges who suddenly manage to whitewash the fact that the graduates of their STEM programs are disproportionately asian (and to a lesser extent, white). But this is why apprenticeships are all but impossible in the US. Without the college degree to exculpate you, you will be certainly be found to have discriminatory hiring practices if you are picky with your entrants.

meh8881 4y ago

Sounds like a good opportunity for ChatGPT if it’s log based.