Morris Chang "Globalization is almost dead and free trade is almost dead" Speech

[M. Y. Zuo]

You who else used a yearly culling system. Enron… Jack Welch was terrible for corporate America

I would add a caveat though, getting GE to weed out the worst performers at the beginning of his tenure was beneficial for morale, since headcount was enormous by then.

But instead of stopping or shrinking its scope afterwards, the practice was continued for decades, which destroyed a lot of the engineering camaraderie within GE.

Anyone with exposure to a corporate environment understands there's some degree of inertia behind decisions relative to the level it was approved by. The larger the company, the more inertia there is.

So it does seem like an odd thing for the thousands of GE executives, including Jack Welch, to ignore it over so many years.

Microsoft also similarly self-sabotaged with similar internal dynamics from what I heard of folks who were involved in 'stack ranking' in the Ballmer era.

 

[Paul2]

I would add a caveat though, getting GE to weed out the worst performers at the beginning of his tenure was beneficial for morale, since headcount was enormous by then.

But instead of stopping or shrinking its scope afterwards, the practice was continued for decades, which destroyed a lot of the engineering camaraderie within GE.

Anyone with exposure to a corporate environment understands there's some degree of inertia behind decisions relative to the level it was approved by. The larger the company, the more inertia there is.

So it does seem like an odd thing for the thousands of GE executives, including Jack Welch, to ignore it over so many years.

Microsoft also similarly self-sabotaged with similar internal dynamics from what I heard of folks who were involved in 'stack ranking' in the Ballmer era.

Smaller companies lack a culture, where culture = cult.

A rational action losses rationality over numerous repetitions.

The "Corporate governance" cult in English speaking countries is a great example.

Small companies = fast natural selection

Big companies = copying taking over learning can last longer before the inevitable crash

 

[Mooredaddy]

I would add a caveat though, getting GE to weed out the worst performers at the beginning of his tenure was beneficial for morale, since headcount was enormous by then.

But instead of stopping or shrinking its scope afterwards, the practice was continued for decades, which destroyed a lot of the engineering camaraderie within GE.

Anyone with exposure to a corporate environment understands there's some degree of inertia behind decisions relative to the level it was approved by. The larger the company, the more inertia there is.

So it does seem like an odd thing for the thousands of GE executives, including Jack Welch, to ignore it over so many years.

Microsoft also similarly self-sabotaged with similar internal dynamics from what I heard of folks who were involved in 'stack ranking' in the Ballmer era.

Any company has fat to trim at all times absolutely. But mandating it is just insane. Taken to an even worse degree are the places that have your coworkers rank you I.E Enron. Basically turning the whole company into a popularity contest and all the people with ideas or a voice worth a damn get weeded out for not “being a team player” I.e willing to play the game. Breeds toxic culture of betrayal and fear. Only bad managers are the ones who think this is a good idea.

 

[M. Y. Zuo]

By three methods we may learn wisdom: First, by reflection, which is noblest; second, by imitation, which is easiest; and third by experience, which is the bitterest.
- Confucius

Tis the common vice of nature, that we at once repose most confidence, and receive the greatest apprehensions, from things unseen, concealed, and unknown.
- Cicero

 

[hist78]

I worked for a company years ago where the CEO said “we need to recognize that no one hires perfectly so we lay off the mistakes every year which turned out to be around 5%. Brutal culture but very successful company until it wasn’t.

Those smart people in such organizations won't dare to ask a stupid question :

Why it's 5%? Why it's not 3.47% or 5.26% or 7.83%?

Sooner or later those smart, honest, and capable people started leaving such organizations.

 

[blueone]

Those smart people in such organizations won't dare to ask a stupid question :

Why it's 5%? Why it's not 3.47% or 5.26% or 7.83%?

Sooner or later those smart, honest, and capable people started leaving such organizations.

I can answer this one... the percentage of employees in the bottom rank is determined by the size of the budget cut you're looking to achieve. For some companies, and Welch's GE was supposedly in this category, the cut was sized by the assumption that the company could hire better performers than those in their existing x% in the bottom rank. For companies looking to cut their headcount budget for business reasons, the percentage is simply determined by the savings they're looking to achieve. There isn't the equivalent of an engineering rule of thumb in the reduction process.

 

[cliff]

Mr. Blue, is this due to VPs and directors ivory towers trying to hold onto their bloated salaries for another year or 2? They want the short term numbers to look good? I assume the they don't care about the long term. They don't own significant portions of the company.

Will HR help make these decisions on these 3-15% cuts? They are great at it. They are best able to filter out the technical biases. Diversity is our strength! Can we assume that the cuts will be made to American who have identified themselves as he/him workers? Note: It is difficult to cut employees from many European countries.

 

[nghanayem]

Not always. It can (and often is) just a reaction to current market conditions. GF didn't want to get out of the leading edge and dump many of their brightest minds, I'm sure Ford didn't want to shutter many of their American factories, ect ect. But if a company is to make money and be competitive against it's competitors sometimes unsavory cost cutting needs to be done (like amputating an infected limb). While I get the "long-term benefit vs short-term rewards" argument, sometimes the "long-term" is far enough in the future that not taking the "short-term" option can damage the long-term health of the business. The best management can do is to let folks leave in a dignified and respectful manner, as well as coming out stronger so that the folks who got laid off did get laid off in vain.

 

[blueone]

Mr. Blue, is this due to VPs and directors ivory towers trying to hold onto their bloated salaries for another year or 2? They want the short term numbers to look good? I assume the they don't care about the long term. They don't own significant portions of the company.

The decisions to do ranking and bottom x% cuts are usually made above the director and VP levels. I'm not sure if the boards of directors are in the decision process. A significant portion of a senior employee's compensation is usually in stock options of one type or another. The one exception I know of is Samsung. Many companies have minimum company stock holding requirements for VPs and above. I'm pretty sure Intel does, for example.

Will HR help make these decisions on these 3-15% cuts? They are great at it. They are best able to filter out the technical biases. Diversity is our strength! Can we assume that the cuts will be made to American who have identified themselves as he/him workers? Note: It is difficult to cut employees from many European countries.

HR always "helps" make these decisions, especially HR Legal people. Every developed country has laws dealing with discrimination. In the US, discrimination in hiring or firing based on race, religion, gender (including LGBT), age, national origin, disabilities, or being a veteran is generally illegal. I say generally because some aspects of affirmative action are legal, such as incentives for hiring candidates a company deems to be members of "under-represented groups", but not quotas. The US federal government, for example, is subject to a hiring law called "Veteran's Preference", which allows forces the USG to grossly discriminate against non-veterans for most job classifications:

Veterans' Preference

www.dol.gov

(Veteran's Preference reminds of "citizenship" status, in that profoundly silly movie "Starship Troopers".)

In my experience US employment laws at the federal and state levels are so complicated it takes an HR attorney for guidance.

Beyond specified legal protections in the US, if you're an "at-will" employee, and most engineers are (who aren't unionized), you can be terminated for any reason, or even no documented reason, so long as the reason isn't illegal. So, if you're employer feels you're too fat or too ugly... but I've read that some cities now have laws against firing people for frivolous reasons like this. It's amazing to me the country functions as well as it does, sometimes.

 

[cliff]

A company at an at-will state would not allow me to fire an incompetent female engineer (She had an MSEE degree. I guess you pay for them now). The lady in personnel told me that was the reason why I couldn't fire her. I have more similar stories that I will withhold. Lawyers are killing our country.

"It's amazing to me the country functions as well as it does"... fasten your seatbelt!

 

[benb]

There are three currents in fab employment, that I'm aware of:
1) The pragmatic stable employment model: Hire selectively from a variety of backgrounds, lay off people who are obviously incompetent, provide enough training so the inexperienced/mildly incompetent people survive, and pay well enough to retain people.
Pragmatic model in downturns: Some fabs invest in equipment orders (when they can get price concessions) and shed headcount, to a limited degree, during a downturn, when they have a strong financial position, and plan to gain market share.
2) European model: Hire selectively, have a higher level of training coming in, with a very willful and protected class of people who know a lot, are eager to prove it, and sometimes do prove it, although they can't be laid off.
European downturn model: Prioritizes protecting workers, furlough rather than laid off, with bankrupcy the only way to fire. This prevents investing during downturns, leads to lack of competitiveness on the upswing, and its how Europe got where they are, sadly. Given that they have the superior model with better training and quality.

In the HQs of Taiwan and Korea, there is another model:
3) Meritocratic tournament model: Everything determined by scholastic record from the college entrance exam, to college, to the job. Hyper competition, small mistakes punished with out mercy, big mistakes hidden, problems solved by making them someone else's problem (ie backstabbing).
Meritocratic in downturns: When problems can't be moved or hidden. Finally they solve them, "never let a crisis go to waste". There can be radical changes, new investments, closures. This hasn't quite started yet at Samsung, but is coming, I'm pretty sure.

 

[nghanayem]

In the HQs of Taiwan and Korea, there is another model:
3) Meritocratic tournament model: Everything determined by scholastic record from the college entrance exam, to college, to the job. Hyper competition, small mistakes punished with out mercy, big mistakes hidden, problems solved by making them someone else's problem (ie backstabbing).
Meritocratic in downturns: When problems can't be moved or hidden. Finally they solve them, "never let a crisis go to waste". There can be radical changes, new investments, closures. This hasn't quite started yet at Samsung, but is coming, I'm pretty sure.

I was under the impression that TSMC was more so 1 than 3. Samsung is infamous for number 3 though. I wonder if Samsung's US fabs operate in a similar manner to their ROK fabs. My guess is no, but I don't know for sure. On the topic of a SS slingshot, I hope so. My worry is that if their memory business continues to slip the money for logic might dry up before a rally is possible. While intel is still very much an unknown quality, things can get real ugly for Samsung if IFS starts eating their lunch as the "cheaper not TSMC leading edge fab". Hopefully TSMC and intel don't squeeze them out of the logic market, and their memory business recovers (but not too much since Micron and SK need a bit of an edge to be competitive with SS's scale).

 

[Paul2]

There are three currents in fab employment, that I'm aware of:
1) The pragmatic stable employment model: Hire selectively from a variety of backgrounds, lay off people who are obviously incompetent, provide enough training so the inexperienced/mildly incompetent people survive, and pay well enough to retain people.
Pragmatic model in downturns: Some fabs invest in equipment orders (when they can get price concessions) and shed headcount, to a limited degree, during a downturn, when they have a strong financial position, and plan to gain market share.
2) European model: Hire selectively, have a higher level of training coming in, with a very willful and protected class of people who know a lot, are eager to prove it, and sometimes do prove it, although they can't be laid off.
European downturn model: Prioritizes protecting workers, furlough rather than laid off, with bankrupcy the only way to fire. This prevents investing during downturns, leads to lack of competitiveness on the upswing, and its how Europe got where they are, sadly. Given that they have the superior model with better training and quality.

In the HQs of Taiwan and Korea, there is another model:
3) Meritocratic tournament model: Everything determined by scholastic record from the college entrance exam, to college, to the job. Hyper competition, small mistakes punished with out mercy, big mistakes hidden, problems solved by making them someone else's problem (ie backstabbing).
Meritocratic in downturns: When problems can't be moved or hidden. Finally they solve them, "never let a crisis go to waste". There can be radical changes, new investments, closures. This hasn't quite started yet at Samsung, but is coming, I'm pretty sure.

I would say Taiwanese share Korean trait of extremely low management mobility.

There are some 60-70 years acting CEO of non-family companies, some of which have global footprint.

Taiwanese switching jobs just 2-4 times in a career was something normal even a decade ago.

 

[ChrisGar]

I won't count that IBM's $1.5 billion payment to Globalfoundries totally as a prepayment for 7nm. At that time IBM can't find a suitable buyer for its semiconductor manufacturing division. Several companies, including TSMC, decided to skip the opportunity. It was that bad.

The $1.5 billion was a bonus IBM put into the deal to convince Globalfoundries to take over a division that IBM had no intention to keep.

You should google the reason IBM is suing GF. GF had committed to develop 7nm and supply to IBM. Then GF canceled 7nm. I believe IBM is suing GF for $2.5B. (the original $1.5B plus damages)

IBM had to scramble to redesign and fab with Samsung.
Lisa & AMD were able to fab with TSMC without missing a beat. I'm sure AMD had been already working with TSMC to do this before the GF 7nm announcement.

 

[hist78]

You should google the reason IBM is suing GF. GF had committed to develop 7nm and supply to IBM. Then GF canceled 7nm. I believe IBM is suing GF for $2.5B. (the original $1.5B plus damages)

IBM had to scramble to redesign and fab with Samsung.
Lisa & AMD were able to fab with TSMC without missing a beat. I'm sure AMD had been already working with TSMC to do this before the GF 7nm announcement.

Yes, I follow this IBM semiconductor manufacturing division spinoff from the beginning. The exact contract requirements and clauses are secrets between IBM and Globalfoundries. One big question is what IBM and Globalfoundries should get or be obligated from the $1.5 billion payment. It's not as simple as a prepayment for 10nm or 7nm.

IBM knew it otherwise there should have a penality imposed on GF clearly spelled out in the contract between IBM and Globalfoundries.

IMO, IBM's volume is just too small to ask GF (or even Samsung) to do something extraordinary.

 

[Daniel Nenni]

You should google the reason IBM is suing GF. GF had committed to develop 7nm and supply to IBM. Then GF canceled 7nm. I believe IBM is suing GF for $2.5B. (the original $1.5B plus damages)

IBM had to scramble to redesign and fab with Samsung.
Lisa & AMD were able to fab with TSMC without missing a beat. I'm sure AMD had been already working with TSMC to do this before the GF 7nm announcement.

From what I remember, IBM developed a 7nm version but it was not manufacturable. GF tried to do a more TSMC "like" 7nm and failed then did the big pivot to focus on milking the cows that they had in the pasture. There is still a lot of bull but no more new cows.

 

[benb]

Earlier there was a tread about the viability of fast following. While not exactly the same situation, GF is close enough to an example of what happens when you stop competing on the cutting edge and attempt to become a fast follower. Technology ages quickly in semis and revenue falls even faster still.

 

[nghanayem]

Earlier there was a tread about the viability of fast following. While not exactly the same situation, GF is close enough to an example of what happens when you stop competing on the cutting edge and attempt to become a fast follower. Technology ages quickly in semis and revenue falls even faster still.

I agree. GF couldn’t capture many new customers (and get the money to grow). 7nm was low capacity (10-15k depending on how big a loan they could get). Combine this with it coming out a bit after N7 and AMD jumped ship now that their agreement’s expired (leaving GF with effectively 0 leading edge customers to even fill up their meger 7nm capacity).

However I would argue that GF wasn’t a fast follower either, because they were only coming out a couple of quarters after N7. Not enough time to benefit from increased tool vendor and academic knowledge. That was simply GF going as fast as they could, not a deliberate attempt to let SS and TSMC do the bleeding for bleeding-edge for them.

As I stated I don’t think (and the market seems to support) that fast following working for foundries. The nature of the original post was whether a pure IDM could survive long term like that given their captive internal client.

 

[ChrisGar]

From what I remember, IBM developed a 7nm version but it was not manufacturable. GF tried to do a more TSMC "like" 7nm and failed then did the big pivot to focus on milking the cows that they had in the pasture. There is still a lot of bull but no more cows.

IBM only did early research on 7nm. (kind of like their 2nm now)
Gary Patton went from IBM to GF in 2015 along with a number of other IBM process development people. Their focus was on 7nm development. So "GF's 7nm" -- was really a direct follow-on to IBM's early R&D. Gary has been at Intel since -- left GF just after 7nm was canceled.

Gary did a number of industry presentations about this stuff during the mid/late 2010s.

 

[ChrisGar]

Earlier there was a tread about the viability of fast following. While not exactly the same situation, GF is close enough to an example of what happens when you stop competing on the cutting edge and attempt to become a fast follower. Technology ages quickly in semis and revenue falls even faster still.

But GF is not trying to be a fast follower. They are not working on 7nm. (at least they have made no public statements that they are working on it -- since canceling it 4 years ago)