Dear Vlad,
.......
The answer to many such questions as posed by the referenced remarks is that it is not that simple. E.g., the Bohren-type experiment (negative resonance absorption of the medium) has been around at least since the late 1950s, and when done in the Bohren fashion, easily outputs quite a bit more energy than one inputs. Every time, anywhere. It is routinely done in nonlinear optical departments every year, in many universities. You don’t see Ford and the big companies jumping onto that, however! And one doesn’t even see it stir a wrinkle in the scientific community. See Craig F. Bohren, “How can a particle absorb more than the light incident on it?” American Journal of Physics, 51(4), Apr. 1983, p. 323-327. Under nonlinear conditions, a particle can and does absorb more energy than is in the light incident on it. Metallic particles at ultraviolet frequencies are one class of such particles and insulating particles at infrared frequencies are another. See also H. Paul and R. Fischer, {Comment on “How can a particle absorb more than the light incident on it?’},” Am. J. Phys., 51(4), Apr. 1983, p. 327 for validation of the experiment. The Bohren experiment is readily repeatable and produces COP = 18. How many of you have bothered to even read the paper?

Every charge and every dipole in the universe also continuously outputs real EM energy in all directions at light speed, thereby exhibiting a COP = infinity. A solar cell exhibits COP = infinity also. So does a windmill. Note that we are not speaking of efficiency, which is ALWAYS less than 100% in a real system, and never exceeds 100% in nature. Else it implies that the system is creating energy from nothing.
So COP>1.0 EM systems are in fact already known. That alone doesn't cause any great vibration of large companies and investment groups. And it hasn't changed the electrical engineering departments, or the big power companies. Hey! Infrared is heat, and putting out 18 times as much heat as you input should be attractive. Consider all those steam boilers in all those power plants, for example.
So I would turn the question around. Overunity (COP>1.0) EM systems for power are definitely not simple, contrary to the "expert" and confident pronouncements of those who have never even seen an overunity circuit. If it were so simple, then those hundreds of thousands of sharp young graduate students we have produced in our universities over the last many decades would have done it long ago. They haven't.

You also cannot do overunity systems with a standard circuit or standard textbook approach; otherwise the grad students would have long ago done it. If you cannot answer the question, "What specifically must be done, in clear principle, to produce a COP>1.0 EM circuit or system?", then you need to think much deeper. That is the point at which one begins.
There isn't even a legitimate theory of COP>1.0 circuits; my book is intended to be the first real cut at initiating one, in the hopes that the young grad students and post docs will pick it up and finish the job.
The classical electrodynamics model (also used in electrical engineering) specifically prohibits COP>1.0 circuits. Anything built in compliance with it will definitely be COP1.0 systems arbitrarily from the Maxwell-Heaviside equations in the 1880s when he symmetrically regauged Maxwell's equations (actually Lorenz did it in 1867, but was not given credit and when H.A. Lorentz did it much later, he was credited with it).
The symmetrically regauged EM equations select only that subset of Maxwellian systems that are in equilibrium in their exchange with their external environment. All COP>1.0 systems must be in disequilibrium with their active environment a priori.

There is some quite different phenomenology involved with COP>1.0 EM systems, that does not appear in the textbooks. Even most successful COP>1.0 inventors do not understand these phenomena, and just eventually find or stumble a way through to handling them. COP>1.0 systems, e.g., output a mix of both negative and positive energy. Try finding a text on what negative EM energy is (the physicists are still heatedly arguing over it) and how to handle it (or even what causes it to appear). Except for my book, you will not find it in any form that is usable by the COP>1.0 researcher. That alone required some real years of work to decipher.
Every COP>1.0 system is an excited system, being far from equilibrium. All excited systems have decay mechanisms; nature sees to that. There is no literature at all -- except for my book --- that deals with nature's unsuspected decay mechanism for slapping a COP>1.0 system back into COP1.0 system going, even when you get one.
Indeed, a COP>1.0 system cannot even be modeled in the electrical engineering model(s). You have to have higher group symmetry electrodynamics, non-Abelian, imbedded in at least quaternions or higher algebras. As an example, one cannot even see what Tesla's patented circuits do unless one uses a higher algebra than tensors. Barrett (an excellent first rate electrodynamicists and one of the pioneers of ultrawideband radar) already proved that, and used a quaternion analysis to show what Tesla actually did. For overunity systems, you have to model (and think in terms of) the "supersystem", consisting of (1) the system and its dynamics, (2) the local vacuum and its dynamics, and (3) the local spacetime curvatures and their dynamics. All three components of the supersystem interact continuously with each other, and the "active environment" of the system that must furnish it the extra energy is those two additional components: the active vacuum and the curvatures of spacetime. That is not even in conventional classical electrodynamics at all. Vectors and tensors just will not cut it; vectors particularly. But the supersystem can be modeled; it just takes a lot of hard work and some real specialists to do it. That has been started, fortunately, but is still embryonic. It still has a long way to go. The best papers along those lines are by the AIAS (Alpha Foundation's Institute for Advanced Study) authors, and are published in Foundations of Physics, Physical Scripta, Optik, etc.
Most of the fellows complaining that they cannot understand why these COP>1.0 systems are just not being produced like pretzels after one has a successful lab experiment device, do not read the physics literature and have never even been exposed to higher group symmetry electrodynamics, unfortunately. Most do not even know what broken symmetry is, or what the asymmetry of opposite charges means. Yet they mostly think it is as simple as grabbing some Radio Shack parts and whipping it together. By analogy, they are unwittingly making brain surgery seem as simple as splinting a broken legbone. It isn't. Indeed, most do not know the precise difference between efficiency and COP.
Most COP>1.0 systems are also highly nonlinear. That means they are subject to chaotic oscillation, self-oscillation, and sensitive to initial conditions including sometimes memory in materials. Scale-up is highly complex and not simple at all, contrary to prevailing opinion. There are ways to control chaotic oscillation, and even random oscillation, but unless one wishes to chase strange attractors all over the map for a few years, one has to use the Russian control system. It works, and since the latter 1990s books are finally available on the Russian method.

Most inventors thus having successful COP>1.0 systems actually have working laboratory experiments and experimental little buildups. That is sufficient to prove the principle and demonstrate a working model. They DO NOT have a robust power system, ready to go into production easily, with just a little simple scale-up. Most such present COP>1.0 systems require from one to two years of very expensive and arduous final research and development, before they can be properly scaled up into production. Several areas of physics outside electrodynamics are involved, in addition to the higher group symmetry electrodynamics. That effort, including building and fitting a math model, is certainly doable, but it is also certainly not easy, and it is certainly not cheap. It's a minimum of a 9 to 10 million dollar per year effort, and that's on the "bare bones" program.

Meanwhile, the only thing inventors have is their intellectual property rights. So the inventor also has to be quite careful, including with information release, else he can jeopardize his patent rights easily. The inventor did not make the laws. Other folks did, and the inventor doesn't really get a fair shake (it gets worse every year). E.g., the best he can get is protection for 20 years. Yet if I write a book, the copyright is good for the rest of my life plus 50 years. That's a whale of a difference in the treatment of intellectual property rights between "authors" and "inventors". And I can copyright a written work for about ten bucks. A good patent, on the other hand, is thousands of dollars, many thousands.

Another thing is something called "cluster patenting". Unless a fantastic job has been done on the patent claims, others (the Japanese are often very skilled at this) will "cluster patent" all variations that fall just outside the claims in the original patent. The inventor may well turn around and find that, for any practical utilization, he will have to get a license from the "cluster patent" pirates, to use his own patent! As an example, the Japanese used to place perhaps 200 engineers into a cluster patent effort on what was regarded as a potentially significant patent.

So the COP>1.0 inventor is usually in a situation where he must successfully obtain a major financial partner to be able to finish the research and development to get to the production stage. And we won't even go into the suppression efforts that get unleashed against the successful COP>1.0 electrical power system inventor. He has to watch his backside, or he can easily "meet with a sudden suicide" on his way to the supermarket. Or a fatal auto accident. Or have drugs planted on his premises, with the narcotics agents then coming in and knocking the door down in the middle of the night. The sniffer dogs go straight to the heroin or such, and he has just been "caught redhanded with the goods". No way can he convince any jury that he isn't guilty. He winds up behind bars for 20 years, wondering what on earth happened to him. Lots more interesting things also are likely to happen to the overunity inventor who gets to a practical and large machine. If he is lucky, he may get the "offer he cannot refuse". If he's unlucky, he may get an ice dart dipped in curare. Or get hit with a neat little Venus technique shooter (one that warps the wave front) that throws the heart into giant fibrillation. As a last resort, a sniper may get you with a silenced rifle from about 200 yards --- such an attempt was made against Sweet, e.g. Having been involved with all those little goodies and some others, let me assure you that it's real.

Most successful COP>1.0 systems that I know of and have worked with, are in that condition where the additional finishing research and development has to be done before the unit can be scaled up for production and marketing. Forget the normal sources of funding. The scientific community states it cannot be done and that it is against the laws of nature. Most "normal" research financing channels are closed to you. That leaves you open to a steady stream of really incredible folk, some good folks, and some that crawled out of the local rocks where they stay with the other reptiles.
Anyway, that's the story on most of the present COP>1.0 systems that are legitimate. To even work in the present system, the COP>1.0 inventor thus has a really difficult job of it, often a near-impossible job. But I'm hopeful that it will yet happen. My purpose in writing the book (over two years in preparation, many rewrites) was to try to turn loose the sharp young grad students on the problem, with the necessary information already in hand. The book is intended to give them most everything I've learned in 30 years of very hard work, which has cost me now some $300,000 out of my own pocket. That's a lot of money for a working fellow, now an old retired dog. I've been lucky to be able to work with several inventors who did have legitimate COP>1.0 systems. So I've been able to see the novel phenomenology, and struggle with it, and puzzle over it very deeply, as to what could be its real nature. Also, I had 30 years to study and try to read the physics literature, to see what the devil the odd phenomenology really were. We did make some progress. The intent is for the young grad student and post doc not to have to spend 30 years of his life just rediscovering the same things. He can start with where I am, correct any errors I may have inadvertently made, and go on from there, much further.
Now a word to the would-be overunity community. You need to quit assuming that COP>1.0 systems are easy to build. To the contrary, they are very difficult to build! Else those young grad students would have built them long ago. You need to recognize that the standard EM texts you studied and the EM you know, will not do the job and have nothing to say about legitimate COP>1.0 systems. It isn't in there, and you err gravely in assuming that it is and just using standard garden variety electricity and electrical engineering. Let me put it this way. An overunity circuit has to work exactly backwards from a normal circuit, at least for part of the time and in some part of the circuit. It is after all a form of true negative resistor, or must contain an element that acts as such. . A normal resistor receives ordered energy from the circuit and disorders it, dissipating it back to the outside environment. A negative resistor receives disordered energy from the active environment, orders it, and furnishes it out into the circuit as extra ordered energy to be used by the system. You must also clearly understand the difference between a "differential" negative resistor (like certain tunneling semiconductors) and a true negative resistor.