via martinshkreli
Hi.
Glad to be back after a few years. I was banned because someone used my account to advertise something. Be careful who you trust!
Quantum computing is something I have followed closely since D-Wave’s founding. There is no, or in the best case, next-to-no intrinsic value for any of the stocks I mentioned above. This means they are worthless. That means they are very bad investments, and their prices will fall soon. I will explain why.
The key thing to understand about quantum computing is a field called “computational complexity”. The reason no one understands this field is it is difficult for many to comprehend. Using complexity and common sense, you will soon learn that quantum computing is extremely slow. “But, I thought it was super fast?!” You were lied to. Quantum computing can’t do anything at all right now, and even in the best case scenario, it will do very little in the future.
What is an algorithm? It’s just a pre-specified series of steps to calculate something. Just like a recipe. But are there different kinds of algorithms? Classes of them? You bet. This is what complexity studies. Let’s take sorting. Sort 52 cards (or 5200 cards) in order. Pretend the cards are just numbered 1 to n.
The running-time of various algorithms ranges from just [n], to [n * logarithm (n)] to [n2]. These are all polynomial time algorithms or faster. This means that the running time of the algorithm scales with the input size.
If you’re still reading, you’re probably asking “what the hell does this have to do with my momentum stock?” Keep reading, because this is really important.
There are some algorithms that run in exponential time. For example, if you needed to brute-force search over multiple nested binary outcomes [2^n], this gets very difficult very fast, thanks to the exponent. These problems are very hard for computers to solve: even all of NVDA’s chips in the world combined couldn’t solve these problems if n is even a modest size (256).
Quantum computing is only superfast for one problem: Shor’s algorithm. Peter Shor’s invention moves factoring from exponential time to polynomial time. This is a scientific miracle. It started the “race” to create a quantum computer (there was no reason otherwise). IBM began doing this in the early 90s IIRC. We haven’t gotten very far since. Nevertheless, the dream of shifting a wholly intractable problem into polynomial time is a scientific computing achievement like no other.
There’s just one problem: it’s useless. Unless your business is prime numbers or discrete logarithms, no one needs this magic trick. What good is a quantum computer? Can it do other problems?
Here is the dark secret of quantum: quantum computers are extremely slow if they are not running Shor’s. Quantum computers have gate speeds ranging from 1kHz to 1MHz. Any modern computer runs in the GHz regime with multiple cores (only one “core” on a quantum computer). Quantum computers have very bad fidelity (99.99% vs 18 9’s for transistors), so they will fail at running any large program. Right now they cannot even run Shor’s at all. So, if you wanted to do something AI-related or even simple calculator problems, quantum computers will always be slower than transistors.
Try to multiply or add two five-digit numbers on quantum.ibm.com — don’t hold your breath, it can’t be done. Quantum computers have a tiny amount of space. At the end of the program, the quantum computer has to collapse the superposition and “read out” the answer. Unfortunately, it reads out the answer in classical bits. The biggest working QC is 153 qubits, so the biggest string you could read out would be your full name (unless you are Vivek). There is no RAM for quantum. You need to calibrate it each time to run a program. It’s a disaster and always will be.
FAQ
So, why the hype?
The word quantum sounds cool. That’s really it. The companies lie. The CEOs are going to jail (at least a few of them)
Any specific stock details?
QBTS is not even a quantum computer. It’s a sad “annealer” which doesn’t even pretend to provide a quantum advantage. Their computers have been around forever and don’t work. I have a lot more info on this, but it is a total mess.
QUBT is also not a quantum computing company.
RGTI is the most honest company in the space, I think because the CEO wants to avoid jail time. He will tell you he has no idea why his stock is up and there is no revenue for this.
IONQ is the most aggressive in saying they will be faster than NVDA, which I hope I have proved to you above is impossible.
The rest are fake me-toos.
Why are the big companies interested?
Sundar has said that “we will figure out what to do with them when we build them”. Corporate research has existed forever. IBM made huge advances in nanotechnology and never monetized them. Bell Labs is what Google, Microsoft and IBM want to be. Quantum computing is one of the futuristic cool science projects left to do. But even people at Google have told me that they don’t think there is revenue possible here. Microsoft has researched hundreds of random projects. Being first is mostly a Nobel prize opportunity.
But the stocks will keep going up!
So buy them!
Isn’t cryptography a business opportunity?
No. Chris Monroe (founder of IONQ) said on a podcast that the NSA does not want a quantum computer. He said they want to know when other countries will have them. We have post-quantum cryptography already. It takes two seconds and is free to install (pip install oqs). America has the best hackers and spies–we don’t need these slow-ass computers for anything. BUT, even if you assume the ENTIRE NSA budget is spent on quantum computers, the sector is still insanely expensive. Remember, your dogshit little quantum stock is probably not going to be the winner in a sector of 40 companies (most of whom didn’t have to SPAC because they were running out of money).
Aren’t there other algorithms?
Yes, there is Grover’s which is simply a quadratic speed up: n –> sqrt(n). That sounds fast until you realize the gate speed and RAM issues, and the big secret: Grover’s is very case-specific and extremely hard to implement. Ask chatGPT if it thinks Grover’s will supplant Nvidia or x86 anytime soon.
There are a lot of other fancy named algorithms: but they are all isomorphic to Grover’s! They reduce to the same thing. HHH is the only other unique algorithm and it fails for similar reasons.
Anyway, I am short all of these stocks. This reminds me of the dot-com bubble. They will go back to the penny stocks they once were.
Short Quantum – IONQ, QBTS, RGTI, QUBT, BTQ, LAES, CCCX
byu/martinshkreli inwallstreetbets
Quantum Computing Inc $QUBT fraud
byu/aizver_muti inwallstreetbets