Hacker News new | comments | ask | show | jobs | submit login
GE Powered the American Century, Then It Burned Out (wsj.com)
61 points by haaen 32 days ago | hide | past | web | favorite | 25 comments

The article gives the answer early on - the company is still in a phase of change from selling GE Capital - for 20 years they were used to using Capital as a source of excess cash to paper over other temporary market issues elsewhere in the company - they no longer have that ability - so now the cyclical nature of normal business will show directly on their balance sheets and statements.

GE will probably be just fine - in a decade.

I'm dubious that business has a cyclical nature. It's hard to distinguish a random walk from noisy waves.

Demand for certain goods and commodities are cyclical - I design dispatch consoles, we tend to get customer verticals in clusters - public safety, power/utilities, transportation and it continues in a cycle, based on replacement intervals (on a 20 year lifetime) - I know Motorola sees similar cycles too - they'll oscillate between public safety and commercial markets.

I know there are similar business cycles in aviation, power generation and industrial controls.

If I understand you correctly, you're suggesting that the demand for dispatch consoles follows a wave with a 20-year period, because the device must be replaced approximately every 20 years? I can believe that for an individual buyer, but for the market as a whole to follow the same period, that'd imply that every buyer made the bulk of their purchases at around the same time. That seems unlikely.

I agree with your assessment. GE was a conglomerate and as such the various businesses in its control operated on different cycles.

Absolutely agree...also didn't a similar thing happen with GM when they got rid of their internal financing division (GMAC)?

Well, maybe their "rank and yank" fostered a low teamwork environment or they fired the people who would actually of helped them with the next big thing. Frankly, their management contribution to American business has to be one of their darkest legacies.

My employer is famous for hiring folks at executive levels from GE and I can tell you without doubt that most of them had been some of the the worse execs I have came across. I can easily see GE as frothing with some of the worst senior level clueless "talent" out there.

I think GE used to be great under Jack Welch because he ruthlessly demanded results. When you operate in that mode virtually anything works because everything other than success gets brutally weeded out. You can start any number of new businesses and the evolution system implemented by ruthless pruning eventually will yield successful species. After Welch left, GE decided to use most of his principles as blueprint except one: demand top notch results and ruthlessly weed out everything else. One company that operates like GE under Welsh these days is Amazon. It would be very hard for any exec to spend 5 years at Amazon without showing amazing results. On the other hand I know plenty of execs at big 5 tech who have sailed through for as much as decade by simply "managing up".

This seems like a "darned if you do, darned if you don't" mentality

Allowing managers to coast and play office politics sucks but also squeezing them doesn't seem very positive (though I'm sure it's great for results)

Non-paywalled version here.


This is turning into the modern day 'first comment.' HN should really just generate this link as a feature.

Edit: My bad, this has been covered: https://news.ycombinator.com/item?id=17320147

Copyright Infringement As A Service (CIASS)— I like it!


11,000 words of gossip about GE CEO promotions with no insight into the actual business problems.

Makes the authors look like incompetent wannabe "journalists."

The author does write about GE's business problems. GE capital, as mentioned above. And the Alstom acquisition, which was a too expensive bet on the wrong horse.

Nice and short, and I tend to agree.

The article ended with an invitation to people who know about GE or some such to comment, but apparently need to be a subscriber to the WSJ to do that.

Subscribe to the WSJ? When I was at FedEx, I subscribed to the WSJ, got the copy each day, but soon concluded that the content was not very useful, had serious flaws like your "no insight", and have nearly totally ignored the WSJ since.

"Gossip"? I agree. And I have an explanation that applies to the WSJ, Forbes, etc.: The writing is not to be informative but a manipulation of the assumed emotions of the readers. The manipulation is to write the stories so that a reader can imagine that they are C-level or BoD people, have a vicarious emotional experience, and, thus, tell themselves that they are learning about business. No, they are learning about applied formula fictional, manipulative story telling. The stories concentrate on gossip, cut of the jaw, grimace on faces, style of eyeglasses, and other personality fluff as if that is what is really important; no, it is what is assumed a reader, maybe the guy in the mail room, the floor cleaning squad, or a security guard, can best find interesting. It is as if the assumption is that only a tiny fraction of the population cares about serious business information but nearly everyone has emotions. Manipulation of emotions, not meaningful information.

A good fraction of the comments mention the stuff that GE has long believed that "a good manager can manage anything" or some such. I heard this strongly when I worked at GE. To me the problem with that claim was that it called for management with too little knowledge of the business, just as in your "no insight".

Yes, early in my career, I was at GE: I was the applied math, statistics, digital filtering, fast Fourier transform, numerical analysis, curve fitting, etc. guy at the HQ of GE Time Sharing. Looking back, the place was devoid of "insight".

In the WSJ article, they keep mentioning GE Power. Okay, that has to do with gas turbines. Yup, GE got into that field early in the history of aviation for turbines for superchargers for airplanes at high altitude with thin air. There the need was just desperate.

Sooooo, GE has gone now nearly 100 years with gas turbines. More generally apparently a pillar of the business has been urgent demands from the US military: (1) The engines on the FedEx planes were from GE in Lynn, MA and originally for a USAF drone. (2) The bigger part of GE aircraft engines was in Ohio and did the big high bypass gas turbine engines, IIRC, originally needed for the Lockheed C5A.

[For subsonic flight, high bypass works better than pure jet. Why? For throwing mass m out the back at velocity v, pay (1/2)mv^2 in energy but get momentum change, which is what moves the plane, mv. So to get the most momentum from the given energy from the fuel, want lots of m and less of v. So of course do throw some hot air out the back but also have a big fan, ducted propeller, grabbing huge volumes of air and pushing it out the back around the core that is burning the fuel, that is, bypassing that core. Of course, Pratt & Whitney, Rolls Royce, and some others around the world also are into aircraft jet engines.]

Theme: GE was really slow to come up with new businesses. That is, their Power division is from initiatives nearly 100 years old. It's been a good business, but there have been too few such while the rest of business charged ahead, e.g., with computing while GE missed out.

GE was early in both transistors and computing. Some of their computing was the computer for the MIT Project MAC, MULTICS. Suddenly some in GE saw that their transistor business was no longer just a profitable cash cow, that technology progress, really integrated circuits, would mean new R&D and capital expenses. So, the generalist managers with "no insight" bailed -- got out of the transistor business and sold the computer division to Honeywell. But MULTICS was a big deal, the first with a lot of stuff, e.g., attribute control list security (ACLs) still important, and lots more. IIRC we can trace much of the architecture of the Intel chips, 386 on, back through Prime Computer (a super-mini computer based heavily on MULTICS and started by some Honeywell engineers) to MULTICS.

Lesson: Due to the generalist managers with no insight, GE missed out on the future of computing.

For more, in principle, there's no good reason GE could not have done what Intel, Microsoft, Cisco, Google, QUALCOMM, Facebook, Amazon, Apple, etc. have done.

Lesson: GE just did NOT do well cooking up good new businesses internally. Yes, the generalist managers would buy businesses based on whatever, sell divisions based on whatever, but for all that financial musical chairs, with only a few exceptions, e.g., GE Capital, long essentially an unregulated bank, just would not think hard about the future of new directions in business.

Yes, there are some close parallels with IBM. I worked there, too, in an AI project at their Watson lab. For one of the problems we were trying to solve, I did, later published, some applied math, a colleague called "radical, provocative", that was much more powerful than what we were doing with AI. I also proposed a research direction, in applied stochastic optimal control, the field of my Ph.D. dissertation, but IBM and Watson management didn't want to let me do. Uh, stochastic optimal control can be a massive user of computer hardware and in a sense an ambitious direction for AI. Since then the Princeton ORFE department has been working hard on that direction.

Lesson: IBM's Watson management and IBM more generally blew a good opportunity. Reason? A big one is the "myth of the generalist manager".

Suspicion: Long the internal GE office power politics was really severe with heavy emphasis on severity and appearances over reality, lots of gossip, etc. The WSJ article hints at this, and, besides, it's standard and called "goal subordination", i.e., fight with the guy down the hall, subordinate the goals of the company to those of personal promotion via political infighting. One of IBM CEO Gerstner's early remarks at IBM was that it was IIRC "the most inwardly directed, arrogant, process-oriented company" he ever saw. From my time in both GE and IBM and the WSJ article, I have to agree with that remark of Gerstner and suspect it applies to both IBM and GE.

Fun stuff: Simple, O(n) solution to the Google lake volume puzzle:


>Due to the generalist managers with no insight, GE missed out on the future of computing.

Of course, you could argue that any number of companies that were successful as computer companies in the 1970s and 80s didn't make (or at least successfully make at scale) the transition to any of a number of successive waves of computing.

And it's difficult to see the path to Google (though DEC had AltaVista) and certainly to Facebook or Amazon for any of those companies.

One "path" to a good, new business is just to see the opportunity and take it.

E.g., what was the "path" taken by each of FedEx, Google, Facebook, or Amazon?

FedEx: Nationwide overnight delivery of small packages, pickup as late as 7 PM and delivery before 10 AM. "Path": Founder F. Smith had a family fortune from Southern Greyhound, had flown in Viet Nam, and was running Little Rock Airmotive, executive jet lube, tune, interior decorating. He took a rock solid executive jet, the Dassault DA-20 Fanjet Falcon, 28,660 pounds gross take off weight, 6000 pounds or so max cargo weight, did modifications to make it haul freight, got the FAA approvals, had a few such jets, set up in some old WWII hanger space on the Memphis airport, started with service to just a few cities, and grew from there. I wrote some software to schedule the full planned fleet, pleased the BoD, enabled crucial funding, and saved the business. I did some calculus for the differential equation

     y'(t) = k y(t) ( b - y(t) )
pleased the BoD, and saved the company a second time. I couldn't hope to do better than that, and the promised stock was very late. I went for a Ph.D., something that might help my career and that no one could take away from me.

Google: Two guys in a computerized library project at Stanford. People liked their search results better than those of Alta Vista.

Facebook: Zuck's Hot or Not at Harvard, somehow much more attractive than MySpace, and some VC money.

Amazon: Bezos had a, IRCC, Princeton or Harvard computer science major, worked on Wall Street, and rounded up some VC money. Early on it was just books and records. As the commercial internet caught on, lots of people had Web browsers; secure means of on-line payments were implemented; Internet data rates were fast enough to send lots of really clear pictures of products; and Bezos was smart enough to build one heck of a server farm. Now IIRC he is selling ~500,000 different products.

Some big companies did try: A standard view of IBM inside IBM was that it was "not a computer company or an electronics company but a marketing company and would be in grocery stores tomorrow if they saw an opportunity." That is, IBM was not tied down to just computers or just their 360/370 mainframes. Hmm, grocery stores? Wal-Mart, Whole Foods, and Amazon? Now with the Internet, computing for taxi companies, etc. there's a shot at grocery stores! But groceries aside, early on GE did try transistors and computers and were ambitious enough to play a big role in MULTICS. Then GE was about as well positioned to win in the future of computing as IBM, DEC, Univac, Burroughs, etc. IBM's Watson lab was into "wearable" computing in the 1980s -- compare with Apple's "wearable" computing now. Same for AI. IBM was doing data science way back, e.g., the IBM Scientific Subroutine Package and their Optimization Subroutine Library. Now IBM owns SPSS, the old statistical package for the social sciences, that is, good coverage of linear multi-variate statistics. Early efforts? Yes. Good, winning results now? Mostly no.

IBM and the Internet: Early on IBM had TCP/IP on a chip. Long IBM was supplying core chips to both Cisco and Juniper. And shockingly at one time IBM was running all the Internet on an NSF contact. Amazing: At one time IBM had all of the Internet and now, as far as I can see, they have nearly none of it. And for social computing (e.g., Facebook), IBM had Prodigy. Since Prodigy was jointly with Sears, they had a path to Amazon.

A lucky horse can find water or be led to water, but it takes still more to have him drink. For both GE and IBM, I put much of the blame of "the myth of the generalist manager".

I don’t really disagree with a lot of what you say. There’s also a lot to be said with respect to the innovator’s dilemma. That said, I wouldn’t put a bet on Facebook outlasting IBM.

Yup. But the flip side of the innovator's dilemma is "Adapt or die."!

How did you end up working at both GE and IBM? I hope you've found something better to do by now.

Yes Dad's father ran the village general store, and his stepfather ran the village feed and grain mill. So Dad had a background in running a business. Still he got a Master's and a job for a salary. It paid the bills, paid for both my brother and I to get Ph.D. degrees, etc. But it wasn't real business.

I was at GE because they gave me a big raise from working the Navier-Stokes equations for the US Navy. Soon I was making in annual salary six times what a new high end Camaro cost. That was enough for some Kennedy Center, Wolf Trap Farm Park, Shenandoah times, good French cheese and grape juice, big times on Turkey Day, XMAS, etc. for my wife and I. She went for her Ph.D.

Later I wanted to do applied math on Wall Street -- learned of Renaissance Technologies and James Simons only much later, too late. And by then my wife was seriously ill so to have a stable job to help take care of her, I took a job at IBM's Watson lab in an AI project. I saved money even in grad school; at IBM for the first time in my life, I lost money -- cost of living in NYS was outrageously high.

For now, right, I'm doing a startup, am a sole, solo founder, do all the core applied math and the computing. So far the computing is 100,000 lines of typing and appears to run as intended. I'm rushing to gather more input data and go live. Latest problem: My first server has 4 memory sticks, DIMMs, DDR3, ECC, each stick 4 GB for 16 GB in total. Last night I discovered that Windows is seeing only 3 of the sticks and is using only 2 of them. So, later today will remove and replace the DIMMS and see if that gets me back to the 16 GB of main memory (when I first plugged the computer together, Windows saw and used all 16 GB -- last night the Windows memory test feature looked at the 8 GB and found no errors). Will be installing two new hard disk drives, 2 TB each. Just fixed a hard error on a hard disk -- with careful use of ROBOCOPY options, moved the data off, did a long form format, moved the data back, and I no longer get requests from Windows to run CHKDSK, check disk. That's how the work goes.

I hear you about the DIMMs and ROBOCOPY. Tip: I spend less time on that hardware minutia and backup reliability and restore operations now that I rent servers on AWS. You might want to try it. Focus your time where your value add is. Also looks good to investors/customers. Also helps avoid overprovisioning (capital expense that doesn't get recovered when business falls through).

I hope you're just being an unfunny smartass and not being condescending and insulting on purpose. It's not welcome.

I'm merely expressing regret for his bad luck in life, the same as you would for anyone who accidentally worked at Enron or Facebook.

Having worked at one of the two, I thought it was hilarious rather than insulting.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact