Archive for October, 2011

Exponential Technologies. Genomic Computing. Energy Innovation. AI. Quantum Supercomputers.



Quantum Supercomputer:

A classical computer has a memory made up of bits, where each bit represents either a one or a zero. A quantum computer maintains a sequence of qubits. A single qubit can represent a one, a zero, or, crucially, any quantum superposition of these; moreover, a pair of qubits can be in any quantum superposition of 4 states, and three qubits in any superposition of 8. In general a quantum computer with n qubits can be in an arbitrary superposition of up to 2n different states simultaneously (this compares to a normal computer that can only be in one of these 2n states at any one time). A quantum computer operates by manipulating those qubits with a fixed sequence of quantum logic gates. The sequence of gates to be applied is called a quantum algorithm.

An example of an implementation of qubits for a quantum computer could start with the use of particles with two spin states: “down” and “up” (typically written and , or and ). But in fact any system possessing an observable quantity A which is conserved under time evolution and such that A has at least two discrete and sufficiently spaced consecutive eigenvalues, is a suitable candidate for implementing a qubit. This is true because any such system can be mapped onto an effective spin-1/2 system.

A quantum computer with a given number of qubits is fundamentally different than a classical computer composed of the same number of classical bits. For example, to represent the state of an n-qubit system on a classical computer would require the storage of 2n complex coefficients. Although this fact may seem to indicate that qubits can hold exponentially more information than their classical counterparts, care must be taken not to overlook the fact that the qubits are only in a probabilistic superposition of all of their states. This means that when the final state of the qubits are measured, they will only be found in one of the possible configurations they were in before measurement. However, it is incorrect to think of the qubits as only being in one particular state before measurement since the fact that they were in a superposition of states before the measurement was made directly affects the possible outcomes of the computation.

Qubits are made up of controlled particles and the means of control (e.g. devices that trap particles and switch them from one state to another).[8]
For example: Consider first a classical computer that operates on a three-bit register. The state of the computer at any time is a probability distribution over the 23 = 8 different three-bit strings 000, 001, 010, 011, 100, 101, 110, 111. If it is a deterministic computer, then it is in exactly one of these states with probability 1. However, if it is a probabilistic computer, then there is a possibility of it being in any one of a number of different states. We can describe this probabilistic state by eight nonnegative numbers a,b,c,d,e,f,g,h (where a = probability computer is in state 000, b = probability computer is in state 001, etc.). There is a restriction that these probabilities sum to 1.



– Right now people episodically sequence their genome (say every year) to find new problems. Every year new discoveries are found as the number of genomes sequenced, grows.

BUT it would be better to store every person’s genome and as the intel inside grows, identify problems in individual genomes and if problems are found, email those individuals to alert them to their genomic defects.

The upside to having an aggregate global genomic database, is that random error (genetic mutations) are factored out. Thus, one can use the global genomic database as the benchmark to compare individual genomes.

Intel Inside on Global Human Genome
– Comparing your genome against the global database of everyone genomic data, to find problems in your genome

Unmet need: Relational databases for genomics?

Clinical Genomics and Research Genomics:
– Clinical genomics will need 10^-7 accuracy, or 600 errors out of 6 billion. Or less than one per gene. Right now it is at 10^-5.
– Research genomics can make do with more errors because researchers can merge data from other datasources to statistically average out the errors


Intel Inside on Genomics as Platform as Service:
– Imagine a PaaS model where companies build solutions (listed below) using the global genomics database

Intel Inside on Genomics as Software as Service:
– Imagine a SaaS model where users log in to see trends in the personal genome and use apps provided by PaaS companies to better understand the data, visualize it and translate it into actionable information that can be shared with physicians.


Storage and data processing problem:
• Storing petabytes of data is huge hardware problem. The disk makers are not able to meet the storage demand of genomics and chip manufacturers are not able to sequence genomes in under 15 minutes. Companies need to innovate to meet both the storage demand and data processing demand, so the genomes can be sequenced quickly with a 10^-7 error rate under 15 minutes a pop.

Search problem:
• Being able to index all of it   —- Google of genomics?

Reads vs. variants: GitHub, Collaboration and Open Innovation Model
– Currently researchers share terabytes of genomics data and ship it around in harddrives. To implement a cloud service, one needs to invent a system like GitHub that uploads one read origin and then subsequently changes/differences to the origin. Furthermore, research should be conducted by building apps on top of the genomic data to analyze it. An open source community can flourish if the research community adopts the open innovation model, rather than the publication-grant model.

RSA security – EMC owns it
VMware – EMC owns it – virtualization for cloud

Data compression:
– No company has come up with data compression software
– No JPG, MPEG, MP4 style compression which is status quo in PC industry
– Right now scientists are mailing hard disk drives. Noobs.
– Basespace from Illumina is basically Dropbox for Genomics. Collaboration and file sharing for genomics.

User Experience, Data Visualization:
• With all this data, UX designers need to reduce it to what humans can understand. E.x. Flight cockpits show a horizontal line to show if you are going up or down, so the pilot can maintain altitude. The same thing needs to happen with genomics. Are you navigating the right area – navigating scientists to right area in the code base in an unmet need?
• Any sufficiently advanced technology is indistinguishable from magic. GPS reduces its triangulation technology to simple, turn by turn graphical user experience.

– Patterns, measurable trends in genomes from a particular race, age group, family tree etc

The Whole Earth Catalog

Landscapes: Volume Two from Dustin Farrell on Vimeo.

When I was young, there was an amazing publication called The Whole Earth Catalog, which was one of the bibles of my generation. It was created by a fellow named Stewart Brand not far from here in Menlo Park, and he brought it to life with his poetic touch. This was in the late 1960’s, before personal computers and desktop publishing, so it was all made with typewriters, scissors, and polaroid cameras. It was sort of like Google in paperback form, 35 years before Google came along: it was idealistic, and overflowing with neat tools and great notions.

Stewart and his team put out several issues of The Whole Earth Catalog, and then when it had run its course, they put out a final issue. It was the mid-1970s, and I was your age. On the back cover of their final issue was a photograph of an early morning country road, the kind you might find yourself hitchhiking on if you were so adventurous. Beneath it were the words: “Stay Hungry. Stay Foolish.” It was their farewell message as they signed off. Stay Hungry. Stay Foolish. And I have always wished that for myself. And now, as you graduate to begin anew, I wish that for you.

Stay Hungry. Stay Foolish.