For years I've been subscribing to Jack Ganssle's excellent newsletter, The Embedded Muse. It's a must read for anyone working with embedded systems. It's also very good as a general resource for engineers who work with any type of hardware or software. If you're a software or hardware engineer, I urge you to take a look at a few back issues and then sign up for the free newsletter.
The issue which arrived today contains an interesting discussion about the number of hours engineers are asked to work. I fall heavily into the camp which questions the wisdom of working more than 40 hours. Lest you think I'm a slacker, consider that I worked for 8 straight years at startup companies, and I probably averaged 55-60 hours per week during most of that period. I can speak from personal experience that the weeks with extremely long hours were not nearly as productive as those where I worked 40-45 hours.
Of particular significance are the stories where upper management gets upset when they find no one in the lab after hours. Whenever you find upper management in the lab after hours, that's also a result of poor management. Good managers will be able to communicate with line managers without checking up on them. Good upper managers will also be sufficiently removed from day to day operations that they realize they may not have the entire picture. Employees connected to lab equipment from home are awfully hard for anyone to detect. If you think that you can only achieve results by having a negative impact on the engineer's personal life, I can safely say I don't want to work for you and I suspect most experienced engineers would say the same.
I would suggest that any project where management asks for a long term commitment of long hours is a direct result of inexperienced or poor management. Any manager with a reasonable amount of time in the industry knows that while you may get a delivery out with a short burst of concentrated effort but if you ask for it over a period of months (or more), you're going to get a product which requires more and quicker maintenance releases thanks to the mistakes even the best engineers make when they're overtired. You're also asking for a large turnaround in your engineering team unless you're able to offer some form of extreme compensation (large bonuses and/or stock options).
For young engineers, I would advise that you consider what you're getting out of any company which is asking for extreme time commitments. If you're getting some form of financial compensation or some form of experience which will prove extremely valuable in the future, I'd say go for it although I'd get it in writing if at all possible. Otherwise, I would suggest that you update your resume and start looking around for a company whose management might not be so out of touch with reality.