Sunday, September 30, 2012

Physics/Quantum physics


Leonard Susskind's The Theoretical Minimum: Quantum Mechanics

This series of posts is an effort to summarize my understanding of basic  quantum physics. I've been studying modern physics since graduate school, on and off in my free time, reading both well known texts (e.g. Feynman's lectures) as well as popular books (e.g. books by Hawking and Greene). A few months ago, I attended Susskind's lectures on quantum physics (part of his Theoretical Minimum series) at Stanford. While I did have a basic (though somewhat flawed) understanding of some of the basic pieces of quantum physics, Susskind's lectures seemed to translate the principles in a way which no other work I had read before did.

The lecture series was divided into 7 topics:

  • System/state definition
  • Mathematical foundation
  • Basic principles of quantum mechanics
  • System evolution with time/change
  • Uncertainty
  • Entanglement

Topic 1. System/state definition

What made Susskind's lectures so understandable is his approach which focuses around the definition of a system and its state. He describes a system whose state behaves according to certain rules (described below). The system behavior under these rules are studied using mathematics and the behavior is derived. The behavior that emerges is non intuitive with effects such as uncertainity and entanglement emerging from the math.

The basic rules have experimentally been shown to be the rules under which physical systems really behave at a quantum scale, so the emergent behavior described by the math must be true, no matter how non intuitive it is. Susskind doesn't discuss the reasons why the rules are true. It's a somewhat philosophical discussion, though he does say that you either
  • accept the rules as ground truth, and therefore accept the non intuitive consequences as the way our universe works, or,
  • believe that there exist hidden variables, behaving deterministically, responsible for the visible non deterministic behavior and you're free to continue to search for them, though almost a hundred years of experimental science has failed to find them
Here is the description of system/state and it's rules. Consider a system whose behavior is governed by two rules:
  • Measurement of state is probabilistic: Consider a system with two states (+1, -1) whose state can be measured by an apparatus. Any attempt to measure the state of the system using the apparatus gives a result of either +1 or -1. If the apparatus is used to try to measure the system by forcing the apparatus to measure possible intermediate states (between +1 or -1), the result of the measurement is still either +1 or -1, with the average of a large number of measurement samples converging to the value of an imaginary intermediate state E.g. Let say that +1 and -1 represented two orientations of the system at 180 degrees from each other. If the apparatus is aligned to the axis, the result is either +1 or -1 and will continue to be so, ad infinitum if nothing else changes. However if  the apparatus is not aligned to the axis, the result will still be either +1 or -1 and will change from measurement to measurement. However, the average of a large number of measurements taken using the apparatus not aligned to the axis will converge to the cosine of the angle between the apparatus and the axis. 
  • Measurement causes state change: Consider a system whose state (+1, -1) can be measured using an apparatus. Measure the state of the system, by aligning the apparatus and the system. The state is measured to be either +1 or -1 and will continue to be so ad infinitum, if nothing else is done to the state, except for the same measurements. However, if the apparatus is rotated through an angle less than 180 degrees, and a measurement is taken, then it may be either +1 or -1. If the apparatus is once again aligned to the system axis, and a measurement is taken, it will no longer necessarily be equal to the first set of measurements.
Its important to note that the first rule is not a limitation of the resolution of the apparatus, but a property of the system. The second rules implies that logical propositions (AND/OR) do not carry the same implications for this system as they do for classical physics. Temporal ordering of the operands will  influence the result of the operation. The uncertainty principle will follow from this rule.

This concludes the description of the system, which along with a mathematical foundation (Topic 2) will be used to derive the principles of  quantum physics (Topic 3).

Posts on the other topics will follow in the next few weeks/months.

The Startup Game: Inside the partnership between Venture Capitalists and Entrepreneurs

"The Startup Game: Inside the partnership between Venture Capitalists and Entrepreneurs" - William Draper, 2012

This book is an interesting read that is part biography, part history lesson and part tutorial on the Silicon Valley venture capital industry. The book covers three topics:
  • The origins and evolution of the early venture capital industry in Silicon Valley (Chap 1)
  • The venture capital process, both in the early days and today (Chapters 2,3,7, 9)
  • Draper's career in areas other than venture capital: his time in China/India, in Washington, and his philanthropic work. (Chapters 4,5,6 and 8)
The first two topics were what I found most useful and are covered in the synopsis below. Draper describes the functioning of the VC industry along with a number of general guiding principles, some well known, others unique. He illustrates them by interspersing them with anecdotes from his experiences with other VCs, entrepreneurs, economists, politicians and investors.

Chapter 1 is a a history of the origins of VC in Silicon Valley.
The first venture capital firm in silicon valley was Draper Gaithner and Anderson (DGA), started in Palo Alto with $6 million in investment in the early 60s by William Draper's father. None of its initial investments could be described as technology - they ranged from defibrillators to aborted attempts to invest in real estate in Hawaii to dental floss dispensers(?). Its first technology ventures included Diablo, Century Data (disk drive technology) and Kasper instruments (semiconductor mask alignment). Interestingly, for a venture capitalist, Draper credits an academic, Fred Terman, as responsible for the creation of the valley ecosystem, with an interesting observation on how Stanford made Terman its president around the same time that Yale's president, a history professor, canceled it's engineering program because Yale did not consider itself a "trade school".

Chapter 2 discusses the five areas (Funding, Teams, Pitch/Product/Market, Deals, Relationships) which shape how venture financing works.
  • Funders: Today's model for a VC firm is a small firm comprised of limited partners (who put in most of the funds) and general partners (who do most of the management). Draper describes how this model started with DGA, where they started with approx $6million from three general partners (including the Rockefeller family), three limited partners (D, G and A). The profit sharing agreement was 60% of profits to the limited partners, 40% (the carry) to general partners, who split the 40% equally. Limited partners were charged an annual fee of 2.5% for managing the fund by the managing partners. Today this model with a few tweaks to the numbers is still the prevalent model with a few additions: For tax reasons, general partners must invest 1% of money in a fund (tax reasons). Investments are made in a series of funds, with a termination date for each fund, in contrast to Sutter Hill ventures (Draper's second VC firm) fund which was an evergreen fund (and still is), where limited partners can opt out at specified dates. Today, this model is used to manage VC funds totaling approx. $200 billion (circa 2011). He also talks briefly about other sources Angels, Corporations, Venture debt, but not in depth.
  • Team/team evaluation: Draper's methods for evaluating founders/teams are somewhat loosely defined, intentionally, and he talks about why it varies from VC to VC. In general, he looks for references, leaders who know the field intimately, have run another company, and have ideas for new features/technologies/products/markets.
  • Pitch/product/market : The pitch is as much about selling the team/leader as it is about selling the product/market, but must cover at least the following:
    • How the product differentiates itself from competition
    • Market size
    • Capital requirements
  • Deal: The deal is what brings the funders and the team together. The goal is to value the company and decide on how much funding to provide and at what what cost.
    • In the early days (70s-80s) the method he used, was:
      • For a company with no profits, but good steady revenue, good prospects, the price was 1 year annual sales.
      • For a company making profit, the rule was 20 times profit (which is equal to 1 years sales if the profit is 5% after taxes).
    • The benchmarks for the first (commonly called A) round is to sustain the company for 1-2 years only. It should result in a 50-50 ownership split between VCs and entrepreneurs. Any dilution for employees is shared equally among VCs/entrepreneurs.
  • Relationship: The 3 stages of the early company are
    • Trial and Error: Mistakes are made, but can be corrected. This is when the technical team executes.
    • Reality: Reality sets in about market size and customer acquisition.
    • Rev it up/Close it down: Company survives or dies.
He ends the chapter, with his list of the top ten errors an entrepreneur makes, a useful list, which I've reproduced below, classified into categories:
  • Business development:
    • Overestimation of market size/customer acquisition rate
    • Unclear marketing plan
  • Communication:
    • Unclear elevator pitch
    • Approaching VCs incorrectly
  • Execution:
    • Underestimation of schedule/timelines
    • Over utilization of the entrepreneur
    • Inflexibility
    • Failing to cut costs when needed
    • No action during a recession
  • Team:
    • Board lacks diversity
Chapter 3 discusses Drapers' views on the characteristics he looks for in an entrepreneur: intelligence, education, energy, passion, expertise, integrity, with a anecdote illustrating the cost of vision with bad execution.

Chapter 7 discusses the exit of a venture. The goal of a venture is value creation followed by the exit. Here Draper discussed the IPO as an exit. His opinion is that an IPO should happen iff the company can deliver increasing profits and revenue for the next several years, so the public can make a gain on its investment. Any other consideration should be second to this criteria. He follows up by a short description of how the business of IPOs has changed in the last 10-20 years, starting with the four horsemen of investment banking and their take over by commercial banks as the Glass-Steagall act was repealed in 1999. He describes the general concepts of an IPO which at a high level is that companies sell a chunk of stock to underwriters (investment banks). Usually underwriters buy all of stock at a price set the day before IPO with an additional fees, typically 7%.

Overall, a good introduction to the venture capital process, with history, a number of interesting insights, and anecdotes mixed in.