We should be paying attention to Quantum Computing.

I have to admit. Quantum physics and mechanics were never my forte. It was always a “get it done” thing at the university so I can move on to the next semester. Unfortunately I should have been paying more attention.

I clearly remember asking a professor the question of “Why is quantum mechanics relevant to a computer engineer”. His answer was not the most inspiring (in a classical Greek way), but he was right. “Because there is a limit on how small we can get with cmos”. I literally forgot about it, got a mediocre grade and moved on to finish my degree.

Fast forward, about 20 years later, here we are now discussing about the progress we have done in quantum computing and how massively more powerful they are from traditional computers. With their first performance numbers coming out, showing 500 million times speed increase compared to the fastest computer we have available now, it really feels that this is totally fiction or that dooms day is around the corner.

But let’s see how this all started and where we should be paying attention to.

Transistor, Size, physics and a liitle bit of imagination

Transistors are the basic building block of any binary computer. It’s a tiny electronic circuit that has the ability to either block or allow current to flow through it. When current flows through it, it’s in the “closed” state. This state is represented as “1”. When it blocks current to flow through it, it’s in the “open” state and this state is represented as “0”. And this is how we have “bits”: Each bit is represented by a transistor in classical computers. You can read more here

A transistor from 1947 built at Bell Labs.

When we started building computers in 1954, transistors where about the size of a light bulb. Currently, an Apple M2 Ultra processor fits 137 Billion transistors. And all this progress was actually well predicted from Mr. Gordon Moore in 1965 while working at Fairchild Semiconductor. This law (more an Axiom to be precise), states that every two years the number of transistors we can fit in a specific area will double. And more transistors mean more speed, so this is the way we scale processing power.

Now pickup your jaw from the floor and read further. In terms of actual size, we are currently doing transistors at 7nm. DNA molecules are about 2.5nm wide and atoms are about 0.5nm wide. As you see we are getting really close to making transistors the size of DNA. But is this really possible? Actually no. It’s not. We are already at our limits and the reason is a phenomenon called Quantum Tunneling. Without getting into much details, this is a phenomenon in which electrons can “escape” physical barriers and move to another state while it shouldn’t. And all this is because of the size we are working on. So, if you make a transistor small enough, then it technically looses it’s function since it can not control the passage of the current anymore.

So are we done with chip speed? More details follow.

Not just bits. qubits

So, now that we have reached the theoretical limit of how small can we make a transistor, it’s a logical question to ask whether we are done speeding up computers by making transistors smaller. Scientists and engineers have predicted that many many years ago (remember my opening about my university years). So they have been already working in finding a solution to this problem. And the solution has nothing to do with making things smaller. It’s actually about changing the way we think about computers all together. It all starts with scaling the bits from 0 and 1s to p(x)=1 and p(x)=0 and anything in between. Enter qubits.

A Qubit is a two-state quantum mechanical system that represents the smallest piece of information in a quantum system. It’s the quantum equivalent to a bit in a classical computer system. The major difference between qubits and bits is this: While in classical computer systems, bits can have only two states, 0 or 1, in quantum computer systems, qubits can be in any state between 0 and 1, with a certain probability. This phenomenon is named superposition and it makes all the difference between the two.

Although each qubit can be in any state in between 0..1 with a certain probability, we can only “observe” one of this states at a given time. It sounds like this is way too complex for the same result of a single bit, but it's really not. Since it’s qubit can be in any state with a certain probability, this means that the computations are done at the same time (in parallel), just like having n different bits each of it representing the potential state of the qubit. But all using a single qbit. The difference is that we can only “observe” one state at a time. But in reality, the states exist all at the same time. It does sound complicated but this is mainly from the fact that we are used in thinking stuff in a discrete way and not in a quantum way.

what does this practically mean

There is no better way to display the massive difference in computation power between quantum computers and classical computers, from raw numbers

A system with 500 qubits is way more powerful than a system with 2^500 bits.
A Quantum computer with 11 qubits can solve a prime factor encryption algorithm (RSA) of 2048 bits, in just a few minutes. The most powerful classical computer we have right now needs millions of years.

Pretty much this means that a simple first generation quantum computer will be able to break every encryption algorithm we currently use (and the internet runs on), in a matter of days.

SO ARE WE SCREWED?

We are definitely not there yet, butwe need to start paying attention. One of the most prominent and well-established quantum algorithms is the famous Shor's algorithm for computing prime factors. This is an algorithm specifically designed to run in quantum computers since 1994. We just don’t have a stable enough physical implementation of a quantum hardware to run it on.

Companies like google, IBM, Rigetti and a bunch more, have been working on creating stable quantum hardware. Currently there are systems with 400+ qubits (from IBM) but each company is playing around with different topologies and methods to keep the system stable. After they have a stable enough system, they will work on bringing the necessary tools to start actually using them.

An actual Quantum Computer built by Google

Currently, we have public and free platforms from IBM and Google that anyone who wants to get their hands in can play around with. Note that this is not an easy course. You will need to create your own cirquits, logic gates, observers etc in order to actually see what a qbit can do. But it’s fascinating so I strongly encourage you to do so.