Topic: A moratorium on all discussions about AI (Read 86 times)

"I do not have the tech knowledge to know the best between reversible and conventional irreversible computation . . . "-I would appreciate it if you learned the basics about reversible computation before commenting. In any case, it does not take much to get enough of an understanding of reversible computation to be convinced that reversible computation is the only path forward. The energy efficiency of irreversible computation is limited to an expenditure of k*T*ln(2) per bit deleted where k is Boltzmann's constant. In practice, in order to overcome thermal noise, one should expect to expend k*T*100 per bit deleted. The only way to overcome this energy expenditure limit is to delete as little information as reasonably possible using reversible computing. With reversible computing, there is no limit to the energy efficiency per logic gate operation. And there have been scientific papers explaining that the energy expenditure of reversible computing can be millions or billions of times less than k*T. And since the computational complexity theoretic overhead (time/space) incurred by reversible computing is surprisingly small, one must expect for reversible computers to replace all conventional irreversible computers.

The only reason that reversible computing is not a buzzword is because the phrase does not sound as cool as "artificial intelligence","quantum computing","GPT", "nuclear fusion", "SARS", or even "SpaceSex" and because the media did not play techno music and use bright catchy colors in order to rile people up. Maybe we should say "isentropic computing" or "injective computing" or "adiabatic computing" or "universe compatible computing" when referring to reversible computing. We can also say that reversible computing is "superposition free quantum computing" to make it a buzzword.

". . .but if it's an innovative invention some experienced users will welcome and sooner or later it will have the buzzword to make it the talk of the town."-And this has not happened yet because most people are really slow at understanding anything about anything. People instead only talk about reversible computing when the media talks about reversible computing because most people are chlurmcks. To make things worse, whenever I have mentioned reversible computing, people tend to turn into the most heinous insufferable assholes. That is not right!

You have a valid OP, because the hype surrounding the ChatGPT bot and AI are more than the capacity of the bot. I believe the reason behind the over-hype of it is caused by the influencers that are making marketing videos about the ChatGPT like it is the answer to everything whereas the bot is just a piece of plagiarism equipment.
 I do not have the tech knowledge to know the best between reversible and conventional irreversible computation but if it's an innovative invention some experienced users will welcome and sooner or later it will have the buzzword to make it the talk of the town.

Fusion is that thing that has always been very far off, and it is very far off for good reason. The only places where fusion is really successful are places like the cores of stars and thermonuclear explosions, and both of those instances require extremely high pressures and temperatures. The sun produces 3.8 x 10^26 W of power but its mass is 2*10^30 kg which means it has a power output of 1.9*10^(-4) W/kg which is not a lot of power when you compare it to its mass. This means that even in sun like conditions with extreme heat and pressure, you only get 1.9*10^(-4) W/kg.  On the other hand, Betelgeuse has a luminosity of about 100,000 times that of the sun while it has about ~19 times that of the sun, so Betelgeuse has a power output of 1 W/kg. It will be hard to get power from nuclear fusion when one needs Betelgeuse conditions just to get 1 W/kg. It is difficult for the nuclei of atoms to fuse because the positively charged protons in the nucleus are very good at repelling each other and not fusing. I would be more skeptical about nuclear fusion, but I do not know enough about the technology to judge, and I am not willing to read very much on nuclear fusion since I have other priorities. In any case, I wish the best for power from nuclear fusion.

Even if energy becomes more widely available and free from carbon emissions and pollution, we should all agree the efficiency of computing will still be an issue. For example, if we consumed 20 times as much energy, we still have the problem of energy directly heating up the planet and causing global warming (if we consumed 20 times as much completely clean energy, we would still have as much global warming from just the waste heat). The energy efficiency of computation is also a problem because microchips get hot, and we will need to dispose of the heat before the chips get too hot. The heat generation limits the performance of microchips.

Reversible computing was never about making computing perfectly efficient. Reversible computing is about making the computation more energy efficient than irreversible computing.

Just a personal opinion of mine, but I believe that the development of nuclear fusion is not just  buzzword, it actually has the potential to change the availability of energy for humanity.
Actually, assuming that we indeed managed to create fusion reactors, the capacity of turning hydrogen into energy by nuclear means, it could imply that people may change the approach from efficiency models in favor of availability.

Reversible thermodynamics processes are about a literal perfect efficiency between states, but why some people seek  for it when the energy becomes so widely available ?

I am in favor of efficiency, just mentioning what could happen in the future, though.

"Yeah, reversible computing is definitely not as popular as some other buzzworthy technologies. Nevertheless, people are not completely unaware of the concept. For instance, quantum computing is one practical application that builds upon the concept of reversible computing. Is it not?"-Quantum computing is based on unitary transformations, and unitary transformations are invertible by definition. But you can also formulate quantum computation and quantum information based on quantum channels. By a quantum channel, I mean a completely positive trace preserving map. And quantum channels are not invertible. In fact, the quantum channels are precisely the thingies you get when you compose isometries (which are one-to-one and can be extended to unitary maps) with the partial trace (the partial trace is what happens when you delete information).

"Reversible computing certainly has the potential to play an important role in the future of computing. But, there are always multiple paths to the future and no one can say for certain which direction technology will take us."-Reversible computing is the only path to the future. We currently do not know what kind of hardware the computers of the future will have, but I can guarantee that it will be reversible as long as humanity makes it that far. Landauer's principle gives a limit to the energy efficiency of irreversible computation, and reversible computation is the only way to avoid this limit.

"That being said, I do not think you are in a position to tell others what they can and cannot talk about and no one should have that kind of power. After all, forums like this are all about sharing ideas and engaging in conversation with one another, right?"-A moratorium on all discussions on AI except for reversible computing is not practical. But people have been talking about an equally impractical moratorium on AI research which is a bad and impractical idea. I am only asking for people to educate themselves before exposing me to their word vomit. If I wanted a word vomit, I would ask GPT 4 for one.

"Well, I am not sure I agree with that statement. It comes across as a bit arrogant and presumptuous, don't you think? People might use recommender systems to make their lives easier, but that does not mean they lack the intelligence to think for themselves."-This is neither arrogant nor presumptuous. If people want me to respect them, they need to act respectable.

"Different people have different preferences and priorities, so it's difficult to make sweeping generalizations about what people tend to care about. In my opinion, most people prioritize technologies that have practical applications and can enhance their quality of life, over theoretical concepts that only appeal to a narrow group of experts in that field."-Except that reversible computing is the only path forward. If people do not know about reversible computing, then I really do not care what they have to say about our current inefficient AI.

Yeah, reversible computing is definitely not as popular as some other buzzworthy technologies. Nevertheless, people are not completely unaware of the concept. For instance, quantum computing is one practical application that builds upon the concept of reversible computing. Is it not?


Different people have different preferences and priorities, so it's difficult to make sweeping generalizations about what people tend to care about. In my opinion, most people prioritize technologies that have practical applications and can enhance their quality of life, over theoretical concepts that only appeal to a narrow group of experts in that field.


Well, I am not sure I agree with that statement. It comes across as a bit arrogant and presumptuous, don't you think? People might use recommender systems to make their lives easier, but that does not mean they lack the intelligence to think for themselves.


Reversible computing certainly has the potential to play an important role in the future of computing. But, there are always multiple paths to the future and no one can say for certain which direction technology will take us.

That being said, I do not think you are in a position to tell others what they can and cannot talk about and no one should have that kind of power. After all, forums like this are all about sharing ideas and engaging in conversation with one another, right?

These days there are plenty of people talking about AI and using AI systems, and they really need to stop. They are really annoying. People should instead educate themselves about reversible computing. Reversible computation will likely be billions of times more energy efficient than conventional irreversible computation (if people can figure out how to manufacture the things and if the things don't break down, so this is a challenge; read the ~2019 paper by Ralph Merkle on mechanical computing systems if you don't believe me). This means that the hardware that our present day AI systems run on is currently horribly inefficient and is better at heating homes in the winter than it is at computing. The computers of the future will be reversible, but people seem to be incredibly oblivious to reversible computing because it is not a buzzword like GPT, AI, quantum computing, or nuclear fusion. Yes. People tend to only be interested in technologies if it is a buzzword, while people tend not to care whether a technology can actually be revolutionary. This is because most people lack the intelligence to think for themselves, but they instead let the recommender systems think for them. Reversible computing is the future. This means stop talking about AI until you know the basics of reversible computing. This includes Landauer's principle, injective-bijective-surjective functions, reversible gates and circuits, uncomputation, Bennett's pebble game and complexity theory, reversible programming languages like Janus, and reversible hardware (adiabatic computing).

Please cease all communication about AI and use of AI (except when needed to learn about reversible computation) until you understand the basics of reversible computing.