Antilibrary

I used to be a bit embarrassed about the books I owned but had not yet read (or hadn't finished) until I came across the concept of the antilibrary in Nassim Taleb's book The Black Swan:

The writer Umberto Eco belongs to that small class of scholars who are encyclopedic, insightful, and nondull. He is the owner of a large personal library (containing thirty thousand books), and separates visitors into two categories: those who react with “Wow! Signore professore dottore Eco, what a library you have! How many of these books have you read?” and the others — a very small minority — who get the point that a private library is not an ego-boosting appendage but a research tool. Read books are far less valuable than unread ones. The library should contain as much of what you do not know as your financial means, mortgage rates, and the currently tight real-estate market allows you to put there. You will accumulate more knowledge and more books as you grow older, and the growing number of unread books on the shelves will look at you menacingly. Indeed, the more you know, the larger the rows of unread books. Let us call this collection of unread books an antilibrary.

Ursula Le Guin

I only discovered Ursula Le Guin’s writings after she passed recently, and I've found her views to be incredibly rich and inspiring—what an amazing person. 

This interview is worth a read, highlighting the breadth and depth of her intellect. 

If you need an introduction:

Named a Living Legend by the Library of Congress for her contributions to America’s cultural heritage—the author of more than sixty books of fiction, poetry, creative nonfiction, children’s literature, drama, criticism, and translation—she was one of only a select few writers (the others being Eudora Welty, Saul Bellow, and Philip Roth) to have their life’s work enshrined in the Library of America while still actively writing. She joined the likes of Toni Morrison, John Ashbery, and Joan Didion in receiving the Medal for Distinguished Contribution to American Letters by the National Book Foundation, and her work garnered countless awards: the National Book Award, the PEN/Malamud, six Nebulas, six Hugos, and twenty-one Locus awards among them. Her name regularly appeared on the Nobel Prize for Literature short list, and writers as varied as Neil Gaiman, Salman Rushdie, David Mitchell, and Zadie Smith herald her as an influence.    

On imagination and justice:

As Ursula once said in an essay accompanying the 500-year-anniversary edition of Thomas More’s Utopia: We will not know our own injustice if we cannot imagine justice. We will not be free if we do not imagine freedom. We cannot demand that anyone try to attain justice and freedom who has not had a chance to imagine them as attainable.”

On her distinctive approach to her craft:

I hear what I write. I started writing poetry when I was really young. I always heard it in my head. I realized that a lot of people who write about writing don’t seem to hear it, don’t listen to it, their perception is more theoretical and intellectual. But if it’s happening in your body, if you are hearing what you write, then you can listen for the right cadence, which will help the sentence run clear. And what young writers always talk about—“finding your voice”—well, you can’t find your own voice if you aren’t listening for it. The sound of your writing is an essential part of what it’s doing. Our teaching of writing tends to ignore it, except maybe in poetry. And so we get prose that goes clunk, clunk, clunk. And we don’t know what’s wrong with it.

On battle metaphors:

I do try to avoid saying “the fight” for such and such, “the war” against such and such. I resist putting everything into terms of conflict and immediate violent resolution. I don’t think that existence works that way. I’m trying to remember what Lao Tzu says about conflict. He limits it to the battlefield, where it belongs. To limit all human behavior to conflict is to leave out vast, rich areas of human experience. 

Software Programs the World

I got a lot out of this podcast! I initially listened to it driving, but the insights were so interesting that I had to listen to it twice afterwards, taking notes to absorb it all. It's an exciting brave new world we have coming.

The most interesting parts:

  • Foundations (jump to clip)
    • Foundational element one: Moore’s Law has "flipped" over last 7-8 year
      • Traditional Moore's Law: a new chip was release every 1.5 yrs that was 2x as fast at same price (this lasted for about 40-50 years)
      • This is the dynamic that drove mainframes, minicomputers, PCs, and smartphones
      • About 7-10 years ago, this ended; chips "topped out" at about 3 GHz
      • Now the dynamic has "flipped" so that every 1.5 years, chips are just as fast but half the price
      • This is a "massive deflationary force ... where computing is becoming essentially free"
      • In this business, we “chart out the graphs and assume we get to the end state," one where chips will be essentially free
      • Chips will be embedded in everything (this is a new world)
    • Foundational element two: all of those chips will be on the network
      • Wifi, mobile, wired, etc.
    • Foundational element three (continuation of "Software Is Eating The World"): software, then, will allow you to program the world
      • Cars, things in the sky, buildings, homes, businesses, factories, etc.
      • "This is just starting"
      • Entrepreneurs are more interesting, more aggressive than ever before because they assume "if there is something to be done in the world, software can be written to do it"
      • Consequences: "...investing in markets that 7-8 yrs ago we would have never anticipated operating"
  • New platforms (jump to clip)
    • Platforms will be different than what we've had until 5-10 years ago: the platform was a new chip (faster) and a new OS
    • Platforms today: distributed systems, scale out systems
    • These are not on a chip, rather built across a lot of chips (distributed systems)
    • Cloud was first example (AWS) - can now create a program that can run across 20k computers (run for 1h, cost $50)
    • Rise of Hadoop, Spark (distributed processing)
    • Financial technology: bitcoin, cryptocurrency
    • Now: AI (machine learning, deep learning) which is "inherently parallelizable" - can run across many chips and get very powerful as they do so
    • "Can do things in AI with distributed computing that you couldn’t imagine 5y ago"
  • The GPU
    • Initially developed for gaming for very high resolution graphical processing → unexpected uses
    • "New application of an old idea"
    • Thirty years ag in physics lab -- if you need a simulatio with large number of parallel calculations (e.g., black holes, biological simulations), write algorithms to parcel problem into pieces and run in parallel
    • In this days days: vector processors ("sidecar computers")
    • 30y later: GPU is basically a vector processor (sits alongside CPU)
    • Ben Horowitz at Silicon Graphics → physics applications, flight simulations, computational fluid dynamics
    • Simulating real world → need same capability (exact same processor)
    • HW platform emerging: NVIDIA
    • NVIDIA has become “seemingly overnight” → market leader in GPUs and chips for AI
    • All entrepreneurs in AI building on NVIDIA chips (in contrast to Intel in previous years)
  • In AI, what are the things that lend themselves well to startups versus larger companies (e.g., FB, Google, Apple)
    • Challenge: people think of AI as narrower than it really is; rather, it is an entirely new way to write a computer program (broadly applicable to problems)
    • Could use AI to analyze consumer data (hard to compete with Google)
    • BUT: many areas where no one has any data yet (HC, autonomy)
    • Big company advantage: lots of data
    • In reality: just a new way to write a program
  • Interfaces
    • The smartphone was an advance over WIMP
      • WIMP interface: windows, icons, menus, printer
    • That was an advance over text based interface of DOS
    • BUT -- life is different
    • Natural interfaces: natural language
    • AI can enable natural language and natural gestures
    • Opportunity to build interfaces for things you couldn’t before
    • One idea: what applications couldn’t you have before because there wasn’t a workable use interface for it
    • Amazon/Alexa
    • Not tied to old generations
    • No “strategy tax”
  • Ability to leapfrog 
    • For many major new advances, interfaces depend on platform
    • But - big companies also have strategy tax -- existing agenda, have to fit next thing into old platform
    • Example: Amazon has taken lead from Apple, Google -- even though it flopped with phone!
    • Lack of phone became an advantage! Clean breakthrough product
  • But where can startups play?
    • A year ago, would have said AI would be domain of big companies (can afford engineers, hardware; data sets)
    • All three have changed
    • AI technology is standardizing (open source → cloud)
    • AI as a service
    • “AWS for AI” (Google, Amazon, Microsoft, etc.)
  • TensorFlow ("This is a big deal")
    • A lot of students on TensorFlow (“trickling down very fast”)
    • Most teams at hackathon had AI and machine learning components
    • Hardware costs coming down across the board
    • In one year -- AI supercomputing chips with algorithms in the cloud (massive deflation)
    • Big data sets -- startups can assemble big data sets BUT...
    • Newest generation of experts -- focusing on small data sets
      • "They'll say, Primitive and crude machine learning required large data sets but not the newer algorithms (they can work on small data sets) - early but enticing (brings problems into small company realm)
    • With these GPUs — can create simulated versions of the real world using video game tools (can train AI)
    • Earthquakes, floods, thunderstorms, swarms of birds
    • Train AI - "AI actually has no idea it’s working in a simulated world and not the real world"
    • Potentially: run millions of hours of simulated training at very low cost
    • Google’s Deep Mind data set -- "game playing itself"
  • Why are simulations so important?
    • Ten years ago - AI, neural nets, and deep learning were frowned upon
    • Improbable (company) - can do large-scale scale out simulations using cloud computing technologies and new, proprietary technology
    • Can get a complete picture of the world, can generate a data set
    • "It’s expensive to make things happen in the world (physical changes...building roads, planes...are hard, expensive; have consequences)..."
    • In contrast: simulation, run experiment, introduce change -- easier, cheaper, no consequences
    • Can run millions, billions of simulations
    • Can make real world decisions with more foreknowledge
  • What are other areas where you can think of real world applications of this technology? (jump to clip)
    • New platforms: health + computer science
    • AI hardware for different type of programming; today, Google has a new chip for deep learning cloud
    • New breakthroughs for quantum computing (more powerful deep learning systems)
    • Chip in everything: platforms to run/manage those chips
  • Theme: tech reaching into new places; “tech is outgrowing the tech industry”
    • Thesis: software is eating the world BUT “hard investments” (Soylent, Oculus, Nutribox)
    • Oculus was actually software (breakthrough tech often needs new hardware)
    • Soylent and Nutribox -- same thing
    • Big believers: big breakthroughs in knowledge (Turing, Shannon) -- new model of the world, companies that build on that new knowledge
  • SaaS — acquisitions; what is left to do there? SFDC or vertical or totally new platforms
    • SaaS as old versions of things in the cloud (WDAY, SFDC, SFSF) -- big categories
    • Changed from on-premise to cloud: seeings sw applications for things that in the old days were cost prohibitive (“screwing it in and hiring army of Accenture consultants) (e.g., expense reporting: Concur) -- new things come into economic viability
    • What was unviable before?
    • Can also scale down to small companies as buyers (<1k employees) - Oracle Financials v. NetSuite
    • Verticals (real estate, construction)
    • Interesting trend
    • Historically: SAP, IBM, Oracle … accessible to top 500-1,000 companies in a handful countries
    • So previously big companies in big countries had an advantage (dominated by 2k-3k multinationals globally)
    • N. Am and Western Europe v ROW
    • Interesting conclusion: smaller company or not one in the Western world (leapfrog similar to what happened with telecom)
    • Larger companies may have harder time adapting
    • Maybe: power shift from larger to smaller companies
    • Companies in western world to those in ROW
    • “Leveling of playing field”
  • The macro view of the economy — “world is starved for innovation and growth”
    • $10t of capital held in gov’t bonds trading at negative yields
    • “Paying the bank interest”
    • “People cannot find enough productive places to put their capital”
    • Negative, conventional view: starved for growth
    • Positive view: $10t waiting for new opportunities (HC, education, consumer products, media, art, science, cars, housing, etc.)
    • “What needs to be done in the world?”
    • “World has never been more ripe for a VERY large wave of innovation that would be quite easy to finance”
    • More money than ideas and creative, effective people
  • Company building and founders — types of founders; what has changed?
    • “Gotten more risk tolerant”
    • “We’re much more interested in the magnitude of the strengths than the number of weaknesses”
    • Lack of experience is a strength: “Hard to rewrite the world if you’re too steeped in the world”
    • Financial terms: “buying volatility”
    • “World class strengths where we care about them”
  • One piece of advice
    • Management: “The most common mistake founders is making decisions based on very proximate perspectives without taking the time to think about how others in the company will see the decision...let’s look past the person I’m talking to.”
    • Strategic: “People need to raise prices.”
    • Most companies have sophisticated views on product, design, engineering and naive views on prosecuting a campaign
    • One dimensional view between price and volume (“pricing cheap, selling more”) 
    • Two dimensional view
    • Raise prices, and you can afford a bigger sales/marketing effort
    • Most companies have prices that are too low to get people to buy
    • Too hungry to eat problem
    • Vicious cycle
    • When you charge higher prices, people take the product more seriously, impute more value, make a serious decision, and when they buy it, they experience a greater sense of engagement, commitment, and stickiness

Tagore

 

Where the mind is without fear and the head is held high;
Where knowledge is free;
Where the world has not been broken into fragments by narrow domestic walls; ...
Where the clear stream of reason has not lost its way into the dreary desert sands of habit; ...
Into that heaven of freedom, my Father, let my country awake.

 Rabindranath Tagore, Gitanjali

 

Amazon

“This is still day one in such a big way.” – Jeff Bezos

The Everything Store by Brad Stone is a tremendously good book. Below are my notes and thoughts.

Bezos on what makes Amazon unique:

At Amazon...

  • We are genuinely customer-centric
  • We are genuinely long-term oriented
  • We genuinely like to invent

“Very few companies have all three of those elements.”

On point of view, or thinking differently:

Alan Kay: “Point of view is worth 80 IQ points.” Examples in the book of unique points of view:

  • When Amazon launched book reviews, he received an angry letter from a publisher telling him his business was to sell books, not trash them. “We saw it very differently,” Bezos said. “When I read that letter, I thought, we don’t make money when we sell things. We make money when we help customers make purchase decisions.”
  • D. E. Shaw, where Bezos worked before starting Amazon: “While the rest of Wall Street saw D. E. Shaw as a highly secretive hedge fund, the firm viewed itself differently … [as a] versatile technology laboratory full of innovators and talented engineers who could apply computer science to a variety of different problems. Investing was only the first domain where it would apply its skills.”

Bezos "stealing" ideas:

  • “I don’t think there was anybody Jeff knew that he didn’t walk away from with whatever lessons he could.”
  • “Good artists copy, great artists steal” (Picasso). 
  • "Stealing" ideas from Jim Sinegal, CEO of Costco
    • Sinegal “didn’t have an exit strategy” – “he was building the company for the long term.”
    • “It was all about customer loyalty.”
    • “Costco buys in bulk and marks up everything everything at a standard, across-the-board 14 percent, even when it could charge more. It doesn’t advertise at all, and earns most of its gross profit from the annual membership fees.” 
    • Sinegal doesn’t regret educating Bezos: “I’ve always had the opinion that we have shamelessly stolen any good ideas."
    • Stone doesn’t make the connection, but Prime looks a lot like Costco’s membership fee. 

Competition:

At the outset: “There was competition already. It wasn’t as if Jeff was coming up with something completely new.” At least not at first, but he was thinking about it very differently than the others. 

When Barnes & Noble launched a competing website and sued Amazon, there was a highly publicized Forrester Research report in which Amazon was referred to as “Amazon.Toast. "Jeff to employees: “Look, you should wake up worried, terrified every morning. But don’t worry about our competitors because they’re never going to send us money anyway. Let’s be worried about our customers and stay heads-down focused.”  

Bad news? Assemble the SWAT team:

In early 1998, Mark Breier, the VP of Marketing, showed Bezos a survey that the majority of consumers did not use Amazon.com and were unlikely to do so because they bought very few books. Bezos instructed Breier to assemble a “SWAT team” of recent hires from Harvard Business School to research categories that had high SKUs, were underrepresented in physical stores, and could be easily mailed. Breier: “I brought him very bad news, and for some reason he got excited.” Bezos had the playbook in his head from the beginning. Breier seemed to have nudged him to the next phase. 

On Marketing:

“Over the first decade at Amazon, marketing VPs were the equivalent of the doomed drummers in the satirical band Spinal Tap; Bezos plowed through them at a rapid clip, looking for someone with the same low regard for the usual way of doing things that Bezos himself had.”

“We spend only forty basis points on marketing.” 

Athletes: On hiring Harrison Miller to lead the rollout of a new category—toys: “Miller knew nothing about toy retailing, but in a pattern that would recur over and over, Bezos didn’t care. He was looking for versatile managers—he called them ‘athletes’—who could move fast and get big things done.”

Articulating culture in 1998: customer obsession, frugality, bias for action, ownership, and high bar for talent.

The flywheel: lower prices → more customer visits → increased volume (direct and third party) → increased efficiency from fulfillment and compute infrastructure → lower prices

Coordination: As Amazon grew, coordination became more difficult. At an offsite, a group presented ideas to improve communication between groups. Jeff stood up with a red face and the infamous blood vessel in his forehead pulsing and said, “I understand what you are saying, but you are completely wrong. Communication is a sign of dysfunction. It means people aren’t working together in a close, organic way. We should be trying to figure out ways for teams to communicate less with each other, not more.” The right question wasn’t, How do we communicate better? It was: How do we improve effectiveness. He later said, “A hierarchy isn’t responsive enough to change. I’m still trying to get people to do occasionally what I ask. If I was successful, maybe we wouldn’t have the right kind of company.”

The Innovator's Dilemma as a manual. The Innovator’s Dilemma had a significant impact on Jeff Bezos (and, being an avid reader, many other books did as well). 

  • Steve Kessel ran the book category for a few years until about 2004, when Bezos asked him to take over the emerging digital business. 
  • Bezos: “If you are running both businesses you will never go after the digital opportunity with tenacity.”
  • Bezos had learned that he needed to set up a new and independent business to pursue a disruptive technology properly
  • He told Kessel: “Your job is to kill your own business.”
  • Bezos was influenced by the book Creation by Steve Grand in which Grand described his approach to a 1990s video game called Creatures. Creatures gave players the ability to “guide and nurture a seemingly intelligent organism on their computer screens.” His approach was to allow complex, higher-level behaviors to emerge from simple computational blocks called primitives.
  • Bezos: “Developers are alchemists and our job is to do everything we can to get them to do their alchemy.”
  • Can this apply to businesses as well? What if within a business the functions—Product, Marketing, Sales, etc.—were primitives on top of which young, relatively untested leaders built new, experimental businesses?
  • How many large, successful businesses emerged unexpectedly as the result of solving an internal problem? Palantir is one example.

Gut calls, intuition, vision. One recurring theme in the book is Bezos’s “gut calls”—times when data wasn’t available, was inconclusive, or even pointed to a conclusion contrary to what Bezos believed and Bezos proceeded in line with his intuitions anyway. A few examples:

  • Prime. “In many ways, the introduction of Amazon Prime was an act of faith.” 
    • “The service was expensive to run, and there was no clear way to break even.” 
      • Diego Piacentini, a senior executive running international operations, said, “We made this decision even though every single financial analysis said we were completely crazy to give two-day shipping for free.”
      • Bezos, however, knew from “gut and experience” that it had the potential to change customer behavior—and the overall company—dramatically. He had seen Super Saver Shipping lead to bigger orders and purchases in new categories. He had seen the increase in spending due to lower friction from 1-Click ordering. 
      • And Bezos was right: “The service turned customers into Amazon addicts.” 
      • And costs did come into line. The fulfillment organization “got better at combining multiple items from a customer’s order into a single box, which saved money and helped drive down Amazon’s transportation costs by double digit percentages a year.”
        • This led me to recall this quote from the philosopher Albert Hirschman (source): “Creativity always comes as a surprise to us; therefore we can never count on it and we dare not believe in it until it has happened. In other words, we would not consciously engage upon tasks whose success clearly requires that creativity be forthcoming. Hence, the only way in which we can bring our creative resources fully into play is by misjudging the nature of the task, by presenting it to ourselves as more routine, simple, undemanding of genuine creativity than it will turn out to be.”
  • AWS and pricing. The AWS team, having some sense of Bezos’s philosophies, initially proposed EC2 pricing at $0.15 an hour at which they would breakeven. Bezos unilaterally changed that $0.10.
    • Bezos believed Amazon had a natural costs advantage and that at such pricing IBM, Microsoft, Google, etc. would hesitate to enter the market. 
    • Stone doesn’t mention this, but I wonder if it was also another platform vision, another flywheel. Perhaps Bezos saw that compute infrastructure would become another flywheel connected to the distribution infrastructure flywheel. Increased usage of the distribution infrastructure flywheel led to lower prices, whereas increased usage of the compute infrastructure flywheel led to product innovation. 
  • Kindle pricing. Bezos priced the books at $9.99. “There was no research behind that number—it was Bezos’s gut call.” The price for digital books was the same as that for physical books, typically $15, so it meant they would lose money, but Bezos believed that publishers would eventually lower their prices on digital books to reflect their lower costs. 
  • Kindle wireless. Wireless connectivity to a cellular connection had never been tried before, but Bezos believed that consumers should be able to download a book easily without having to connect to wifi. Bezos faced resistance on both the engineering and the economics but pushed them to do it anyway. 
  • Random customer anecdotes. “Random customer anecdotes, the opposite of cold, hard data, also carry tremendous weight and can change Amazon policy. If one customer had a bad experience, Bezos often assumes it reflects a larger problem and escalates the resolution of the matter inside his company with a question mark.” Wilke: “It’s an audit that is done for us by our customers. We treat them as precious sources of information.”
    • This is why Medallia is an incredible product and on track to be an incredible business.
  • Distribution centers. “[Amazon’s accounting group] fretted about opening seven costly distribution centers and even about having gotten so deeply immersed in the muck of distribution in the first place. Bezos insisted the company needed to master anything that touched the hallowed customer experience, and he resisted efforts to project profitability. ‘If you are planning for more than twenty minutes ahead in this kind of environment, you are wasting your time,' he said in meetings.”

Bezos on Blue Origins, his space travel venture: 

  • “Slow steady progress can erode any challenge over time.”
  • The group’s motto: Gradatim Ferociter, which means “Step by Step, Ferociously.”

Miscellaneous notes:

  • Ignore what other people think.
  • Think. Don’t use PowerPoint. Use six page written narratives.
  • Work backward from the outcome you want. Example: write the press release for a product before starting development.
  • Bezos is a big fan of Nassim Taleb’s book The Black Swan. A key lesson: avoid narrative fallacy. Favor experimentation and clinical knowledge over storytelling and memory. 
  • Nurture the idea. “When a company comes up with an idea, it’s a messy process. There’s no aha moment” 
  • At D. E. Shaw, Bezos was “constantly recording ideas in a notebook.”
  • Alan Kay: “It’s easier to invent the future than to predict it.”
  • “[Bezos] embraces the truth. A lot of people talk about the truth, but they don’t engage their decision-making around the best truth at the time.”
  • “Jeff almost always prefers to build it.” 

Amazon's early timeline:

  • July 5, 1994: founded
  • November 1, 1994: www.amazon.com registered
  • April 3, 1995: first order
  • July 16, 1995: site goes live
  • Early 1996: revenue growing 30-40 percent a month
  • 1996 revenue: $16m
  • 1997 revenue: $144m (which was than what Bezos had predicted as “best case” revenue for 2000 when he was raising capital in 1995)
  • 1998 revenue: 3x growth?

Funding:

  • Early 1994 – early 1995: $10k from Bezos, $5k from Shel Kaphan (first employee), $84k in interest free loans from Bezos
  • Early 1995: $100k from Bezos’s parents (Bezos told parents there was a 70 percent chance they could lose it all)
  • Mid 1995: another $145k from Bezos’s parents
  • Late 1995: $1m at $5m (post-money?) valuation from local investors investing about $50k each (20 of 60 approached)
  • Mid 1996: Projecting $16m in sales, Kleiner Perkins, with its $60m valuation, beats out General Atlantic’s initial $10m valuation, investing $8m (Bezos insisted that Doerr, not a junior team member, join the board)
  • May 5, 1997: IPO (raising $54 million)
  • 1998-2000: $2.2 billion in bonds
    • May 1998: $326 million in junk bonds
    • February 1999: $1.25 billion in convertible debt at 4.75 percent
    • Early 2000: $700 million bond offering