What would Steve Jobs do?

220 1

This morning I woke up to the most fun news I have read in quite some time:

Pixar Pioneers win $1 Million Turing Award

I loved Toy Story when it came out. And Jurassic Park. And Avatar. (Ok, not Titanic.) Lord of the Rings and Terminator 2 were enjoyable. 

I immediately recalled:

After Jobs was ousted from Apple in 1985, he bought Pixar (at the time called Graphics Group) from Lucasfilm for $5 million. He became the company’s largest shareholder and CEO until Disney bought it for $7.4 billion in 2006.

At the time I read the Turing-prize news this morning, I was getting ready to exercise on a treadmill, to a music video (rotate Rihanna, Shakira, Adele, Rod Stewart, Eagles, The Bee Gees, Olivia Newton-John….you get the drift) on the home theater (108 inch screen, digital light projection, with excellent sound, of course), but decided instead to switch to watching Toy Story!

Olivia Newton-John likely did not have photonic computation in mind when she sang (in 1981): 

Let’s get physical, physical
I want to get physical

Duo Lipa (in 2019) similarly seems to have little interest in non-Boolean computational models and non von Neumann architectures, or looking at computing beyond what Turing did, as she sang:

Let’s get physical
Let’s get physical
Come on, physical 

As you know, my excursion – transgression? – into quantum so far has been mathematical, algorithmic. 

I now want to get physical.

My expedition to the physical – akin to how I conceived of OrganJet and Nudge in Re-imagining the US Transplant System — begins with my long-time trusted mantra of :

Maximally Inverse.

I cannot put into words the sheer repulsion I felt at the sight of a quantum machine (both of Google and IBM).

If not for the fact that I delight so much in the mathematics of it all, and the sheer joy that I experience in the newness of the subject, such ugliness may have been sufficient for me to look elsewhere for mental enjoyment and intellectual recreation.

Beyond the visual disgust, I wondered:

Is it physically possible to solve an Ising problem without supercooling?

Can it be done at room temperature?

Is it economically possible to do so without being a Trillion-dollar market cap company (before the Coronavirus-driven market meltdown) like Microsoft, Amazon, Google or even a (mid-cap?) $100B company like IBM?

What can be constructed using commercially available components?

Can it be miniaturized so it can fit into our (denim) pocket?😏

Just like the motivation for OrganJet, when I wanted to democratize Steve Jobs, where I wondered “how poor can you be and still avail of what Jobs had?”, I now asked the maximally inverse question:

How can we create a quantum-inspired machine with as little money as possible?

A Google search (oh, the irony) gets you this:

A poor man’s Ising machine based on opto-electronic feedback systems for solving optimization problems

One can use coherent light – a laser – and compute using phase (and frequency) superpositions.

The interactions of these superpositions lead to a collective state, and this state carries the result of the computation.

This is an example of photonic computation.

Oh, wouldn’t Newton be impressed how far we have come from his invention of the reflecting telescope in manipulating light? 

He would probably not be amused to read what Einstein – who seems not aware (or had forgotten) of the plague years, nor the fact that Newton’s childhood was traumatized because his (biological) father died before he was born and his mother abandoned him for the first few (3?) years of his life — wrote in the 1931 Foreword to Opticks:

Fortunate Newton, happy childhood of science! He who has time and tranquility can be reading this book live again the wonderful events which the great Newton experienced in his young days. 

Well, Einstein was a physicist, not a historian. Let us give him a pass, as among other contributions, he did explain the photo-electric effect.

Photonic computation is just one example of a much broader class (Thanks Davide Venturelli of NASA QuAIL!):

Oscillator-based computing.

This is not a new idea. It had been patented nearly 65 years ago, by, wait for it: 

John von Neumann.

Interestingly, the person looking well past Turing machine, Boolean logic (and von Neumann architecture), was von Neumann! 

Please enjoy:

US Patent 2815488A

which is currently held by, no kidding, IBM. 🤷🏽‍♂️

Last year, Peter McMahon (then at Stanford, now at Cornell, my PhD alma-mater), had contacted me as he has built one such device. His Science paper is:

A fully programmable 100-spin coherent Ising machine with all-to-all connections

Now his post-doc and other students are running experiments on the device to see if they can replicate the solutions the Tepper Quantum Computing Group (using D-Wave machine) reported in:

Graver Bases via Quantum Annealing with Application to Non-Linear Integer Programs

So, how can I help bring such physical devices to market?

How can I cultivate – refashion — the research agenda in quantum (and quantum-inspired) computing towards Ising solvers rather than be forced into a (ugly machine) groove – hegemony? – of Circuit (Gate) quantum framework?

Serendipitously, in my recent bi-monthly Skype call with my research collaborators Anil Prabhakar and Prabha Mandayam, at IIT-Madras, my undergraduate alma-mater, I found out that the Indian Government had approved a 4+1 Master’s program in Quantum Information:

Inter-disciplinary dual degree program on Quantum science and technologies.

The finest undergraduate students in the world (IIT-Madras is ranked #1 in India among the IITs, and as is now well recognized, it is harder to get into an IIT than Harvard, Stanford and Princeton combined) would be admitted into this program based on their Grade Point Average!

I thought I would seize this opportunity to nudge the new generation of researchers and engineers towards bringing non-classical computation (can be non-optical, indeed it is quite likely, they will be microwave or analog electronic versions of CIM, thanks Peter for your timely email!) into our expanding corpus of scientific knowledge and engineering innovation.

Upon approval from the appropriate authorities, we will be publicly announcing:

The 2020 Tayur Prize for prototyping and optimizing on a Coherent Ising Machine.

It is my sincere hope that a new avatar of Steve Jobs will emerge, create a beautiful quantum-inspired machine, that fits in our denim pockets, and build a multi-Trillion-dollar market cap company that changes the world.

PS: It has not escaped me that the current CEOs of both Google and IBM are alums of IIT.

1 comment

  1. It doesn’t have to be ugly. It should be beautiful, and use coherent light!
    Congrats on the Tayur Prize – brilliant.

Leave a Reply