Remains of the Day: Issue 03
All about time: circadian rhythms of Silicon Valley, OODA loops, Little Women
|Eugene Wei||Jan 18|| 9|
Sorry, I realized I forgot to append this header to the last newsletter. As a reminder, I’m Eugene Wei, a former product exec at places like Amazon, Hulu, Flipboard, and I’ve kept a personal blog Remains of the Day since 2001, offering what I like to think of as “Bespoke observations, 80% fat-free”, though the perceived fat ratio will vary depending on your tastes in technology, media, and all the other random topics I write about.
I can’t remember where I saw this, likely in some meme, but for a long time it was true; despite the ubiquity of smartphones in everyday life, I never had a cell phone in my dreams.
It was a striking anomaly, but perhaps awareness of that oddity was all that was needed to resolve it, because recently I’ve started to have my iPhone with me in my dreams. Not only that, the last two times I’ve had my phone in a dream, I’ve been using it to film events inside that dream.
Immediately after waking up, for a brief moment, as my brain is still in that foggy ether between dream and reality, my first impulse is to reach for my phone to review the footage. It’s always with some regret, when waking, that I recall my dreams for that last time before I know they’ll fade from my memory forever, but not having the dream footage in my camera roll is some new form of FOMO.
Looking forward to the smartphone that ships with an added new camera: the dream lens. Specially designed to capture footage from your dreams. Sounds like a premise from some future Chris Nolan film or science fiction novel.
Last week I published a short piece on my blog on the John Wick universe as an allegory for cancel culture. In particular, I was struck, in John Wick 3: Parabellum, by the distinctive image of various citizens all over the streets of Manhattan, glancing at their cell phones, just before they grabbed the nearest weapon at hand and attacked Wick. These days, the phone screen is the dominant theater of war; it’s from that glowing rectangle that we receive so many of our cues, including those to, as Beyoncé said, get in (online mob) formation.
A related worthwhile read is Venkatesh Rao’s The Internet of Beefs, the latest piece to try to survey the war-torn landscape of the internet and decode the structure and mechanisms behind the neverending squabbles. I tend to agree that we’ve passed some tipping point, that the arrival of the internet and then of social media fundamentally altered the world’s ability to construct shared global narratives of the scale that united us in the past. That Trump and other populist strongmen have risen to power is not because they've achieved some new level of political genius but rather that the conditions of the world shifted to reward an unwavering commitment to tribal esprit de corps.
I’m constantly reminded that Yeats could be writing about today:
Turning and turning in the widening gyre
The falcon cannot hear the falconer;
Things fall apart; the centre cannot hold;
Mere anarchy is loosed upon the world,
The blood-dimmed tide is loosed, and everywhere
The ceremony of innocence is drowned;
The best lack all conviction, while the worst
Are full of passionate intensity.
The centre that held in the past may have been an anomaly, a symptom of a media world with a few centralized gatekeepers who could, with relative impunity, author mass illusions of similarity harmony. The modern age, in contrast, offers asymmetric emotional rewards for the narcissism of small differences.
The three most momentous media inventions of my lifetime are 1) the internet 2) the launch of Fox News channels around the world 3) the Great Firewall of China. The one thing they share in common: each enabled humans to fork huge swaths of humanity off from the previous branch of reality everyone lived in. Add in social media with its algorithmic incentives to amplify those differences and you arrive at the depressing state of affairs today. It’s still unclear what, if anything, can create new shared realities able to transcend these gaping fault lines.
I have more thoughts on this which I’ll try to tie up in a future post.
[Incidentally, this John Wick essay is the first time I’ve used Tufte-style sidenotes on my blog (thanks to the inimitable Gwern for sidenotes.js. I’m excited to play with them more in future pieces on my blog (email, unfortunately, doesn’t support this). I have a naturally discursive mind when writing and often find myself pruning thoughts that don’t even deserve to be parentheticals, but opening up the margins of my website gives them a home.]
The Circadian Rhythms of Silicon Valley
After you spend enough time in Silicon Valley, working in technology, you start to recognize in the cadence of its news stories the workings of a series of interlocked clocks, each on its own cycle time. The pace of action is governed by a series of explicit and tacit agreements between different classes of people here, all with varying reserves of patience.
We can begin our journey with any of these groups, but perhaps it’s easiest to start far upstream. There, you have limited partners (LPs), who provide money to the venture capitalists for their funds. LP’s expect that the fund returns their investment in something like 5 years, and pays out profit for, say, the subsequent 5 years.
VC’s, in turn, funnel the money from the fund to entrepreneurs. Like their LP’s, VC’s pass on a time horizon to entrepreneurs. VC’s generally expect you to swing for the fences because the economics of their portfolio are built around the occasional massive outliers, and they’ll be patient, but only to a point. Because their LP’s expect a certain contour of returns over that 5-10 year timeframe, VC’s can only wait on portfolio companies for so long.
As an entrepreneur, no matter how friendly your investors, at your back you always hear/time’s wingèd chariot hurrying near. Maybe you’ll get around 7 years, if you’re fortunate enough to survive through multiple rounds of funding and the Hobbesian competition of the open market, to reach some liquidity event for your investor. In the first few years, the investor will encourage you to keep taking swings, even if you topple into a pit of despair and want to return their money. Getting some portion of their money back isn’t all that interesting to them if there’s still some chance you can pivot and achieve some hockey-stick growth. If they still have faith in you, they’d rather you be taking more shots on goal.
Entrepreneurs, in turn, hire people with compensation packages composed of a salary and some grant of stock options. Typically those options vest over four years, yet another clock. In the best case, the startup achieves product-market fit within the first few years and the employee stays around for at least those four years to help the company grow into a stable, ongoing concern, maybe even a unicorn.
If things go poorly, on the other hand, the options probably aren’t worth much if anything and so the vesting period becomes largely irrelevant. Still, given the long and winding journey many companies take towards product-market fit, the four-year options clock always teases the promise of obscene riches. It’s a dream that creates enough gravity to bond many a constellation of employees together to ship a few moonshots.
Every employee has a personal financial situation that affects the urgency of their internal clocks, but in general, there are two primary competing temporal considerations. One is whether the employee believes that staying longer and vesting more shares is worth their time. That depends on how the company’s prospects at any moment compare to those of other companies spamming the employee with offers through LinkedIn and headhunters. Ultimately every employee with desirable skills is continually gauging whether they’re maximizing the marginal return on their time, however sophisticated their heuristics.
A complicating factor is deciding how much diversification is optimal. Holding options from a few companies can spread your risk across different sectors within tech, but if you think you’re already working for the next Amazon or Facebook or Google, you’re better off staying put as long as possible. Of course, if it were so easy to tell which companies were going to be the big winners, the decision wouldn’t be so difficult, and surefire winners of companies could reduce their comp packages. The meta point is that for any individual case there will be uncertainty, which is why investors diversify across many companies. That friend you know who ended up with a 9 or 10 figure net worth in the Valley? There was some luck involved.
The other clock individuals must consider is when they can afford a mortgage on a condo or house in the Bay Area. The cost of laying such roots has risen dramatically in the past decade. It’s increasingly difficult to imagine being able to afford a mortgage here without having at least one liquidity event of at least moderate size in one’s rearview mirror. How fast that clock counts down depends on your life circumstances, whether you have kids, how much you and your spouse want a bit more room, and so on.
The average tenure of an employee at the average tech startup is just under two years. I suspect if more companies do as Pinterest did and make it possible for employees to wait up to seven years after leaving the company before they have to decide whether to pay to buy out their vested options that the average tenure might shrink even more. It used to be many employees were financially handcuffed to the startup they were working at because they couldn’t afford to buy out the options they’d already vested. And even if they could afford to pay for those options within 90 days of leaving the company, there was a non-trivial chance those options might never be worth what they cost that employee. It’s quite a gamble on the part of the employee, one they often can’t afford to take.
Once you’ve spent enough time in the Bay Area, you begin to internalize all these cycles. When one of those timers starts to run out, you can anticipate certain repercussions or phase changes. The Board representative from your VC might start to ask tougher questions about an IPO or profitability. A firm might start raising prices or focusing more on cost management. Employees may start eyeing other job offers more closely.
Understanding the Circadian rhythms like these can also help you to understand why investors and companies and employees are willing to make some investments and not others.
For example, VC’s tend to shy away from industries like some types of green technology where the payout periods and capital required exceed the patience and pocketbooks of their LP’s. One of the most critical questions about hot tech sectors like VR, general AI, autonomous vehicles, and cryptocurrency is whether the breakthroughs will come in the next 5 to 10 years or longer than that. If we’ve been overoptimistic on the time horizon, then some VC’s are going to have some tough conversations ahead with both their LP’s and some of their portfolio companies.
Reporters like to complain that Silicon Valley is in a bubble. The thing is, when things are going well, Silicon Valley will almost invariably be in a bubble. Most startups need several rounds of funding before they can achieve some profitability. No investor will write the first check in if they don’t believe someone else will come in for the second and third and even fourth checks that the company will need before they climb out of the trough of the J-Curve.
Another way to define a bubble is as an excess of belief. But if your goal is to fund a bunch of long shots, 9 out of 10 of which will go bankrupt, a surfeit of faith is almost a necessary precondition for the whole thing to exist in the first place. Silicon Valley is powered by bubbles; that’s the region’s secret sauce.
First-time entrepreneurs are often surprised and dismayed to find that an investor they sent a deck to then shared that deck out with a bunch of other investors, but that’s just what investors do to see if there will be enough other believers to join them in providing the needed capital to get this company to the promised land. The idea that investors don’t care what others think is not true; they would love to lead a round if they have conviction, but part of their conviction is predicated on their belief that more investors will buy in down the road.
As an employee, it’s important to understand that when a company is under pressure to please its investors when the chances it doesn’t even exist in 5 years are greater than the chances it does, and when the average employee tenure is sub 2 years, the company isn’t incentivized to invest a lot on training and development. It just isn’t financially prudent. The singular goal is to survive and achieve product-market fit.
In bygone days of lifetime employment with one company, the ROI on employee training was a different story. When younger employees ask me for advice as they enter the tech world today, I always counsel them to take control of their continued career development. It’s unlikely anyone else will.
This does mean that a lot of managers and executives in tech companies that manage to survive are wartime promotions, with little qualification other than being high performing soldiers. Tech is plagued with a lot of unprepared, and frankly under-qualified managers, but that’s to be expected in such a high extinction, short cycle time environment. Silicon Valley treats its startups, in their totality, as an r-selection species; that is, the Valley births many startups but invests relatively little in each. The rare outlier that achieves traction is treated like a K-selected species, with investors pouring a ton of resources into them.
Getting Outside Your Competitor’s OODA Loop
Speaking of cycles of time, last year I read Boyd: The Fighter Pilot Who Changed the Art of War, a popular book among a certain niche of the tech cognoscenti. It’s the book most responsible for popularizing Boyd’s concept of the OODA loop in the business vernacular, especially among those who like to drop military metaphors from Sun Tzu and von Clausewitz and the like.
The most common application of Boyd in business is the principle of getting inside your competitor’s OODA loop. OODA stands for Observe - Orient - Decide - Act. Being inside your competitor’s OODA loop means the cycle time for you to complete one such loop is shorter than that of your enemy. For example, if the OODA cycle time of your product team is faster than that of your competitor you should theoretically beat them to product-market fit, as well as to every critical product update. Given the increasing returns and winner-take-all dynamic common in tech today, speed is often a powerful, if not decisive, competitive edge.
I come not to bury that advice, but to look at a variant, almost its opposite. Sometimes you can win by having a slower OODA loop if it just endures longer.
There can be many reasons for this. One, and it’s difficult to stress this enough, timing is critical to so much of Silicon Valley success. The same ideas are tried over and over again, failing time and again, until one day, for a variety of reasons, the stars align, the environment is conducive, and suddenly an idea takes off. It happened for online commerce and streaming video, and someday it will happen for VR and self-driving cars and cryptocurrency. The problem is that if you’re too early on something, having a faster OODA loop just means you burn through your investor’s money and implode more quickly.
Another reason a longer time horizon may be optimal is that many competitors may not have the heart or stomach for the long haul. Startups may give up because don’t have the capital reserves to stick it out through the lean years (see the previous section). On the other hand, established companies, especially public ones, may have the resources but not the commitment (or a long enough leash from their investors) to fail repeatedly in some category that isn’t their core business. If Google had persisted with Google Wave, maybe they’d have invented Hipchat or Slack long ago. If Twitter hadn’t let Vine wither on the, uh, vine, maybe they’d have built TikTok. Institutional trauma from repeated failure can weaken a company’s resolve.
When I was at Amazon, Jeff Bezos repeatedly stressed the importance of being willing to fail again and again on some effort as long as the team still believed in the overarching thesis behind it. Think about Amazon and grocery delivery. Way back in 1999, Amazon acquired a stake in Homegrocer, which eventually IPO’d then ran out of runway and got bought by Webvan, which itself left a crater in the earth when it imploded during the first dot com crash.
More than a decade later, Amazon is back at it with the Whole Foods acquisition, which itself is an evolution of Amazon’s own grocery delivery efforts like Amazon Fresh. Few tech companies are willing to fail as often over as long a period of time as Amazon. People forget that Amazon Web Services (AWS) as it’s known today is not even the first incarnation of Amazon’s web services offering. The first version was for affiliate sites to more efficiently scale their websites hawking books on various subjects.
Amazon tried having merchants like Toys R Us sell directly on Amazon.com; Toys R Us sued to get out of that deal later. Amazon then tried using its tech to power standalone websites for those same retailers, like Target. That deal later died as well. Today’s AWS isn’t specifically retail-focused, but it is built in many ways on lessons from those earlier shuttered efforts.
Of course, it’s not possible to have a long-term time horizon if you aren’t resourced for it. One of the first massive projects I worked on at Amazon was what was, at the time, the largest convertible debt offering in history with Morgan Stanley. This was before the first dot com crash, and the thesis was that the stock market was overheated, and if it crashed, a lot of our competition would run out of money and then time.
To ensure we could survive and also that we wouldn’t have to alter our strategy one bit if such a crash occurred, we needed to bulk up our balance sheet, even with our negative operating cycle and its ability to spin out cash as we grew. Thus we raised a billion-dollar-plus convertible debt offering, a just-in-case treasure chest.
That dot com crash did come, not shortly thereafter, causing what was a mass extinction event for tech companies, including many of our online commerce competitors. Volatility of faith is deadly when it is matched with volatility of funding. If you know others will be killed off by high variance in the negative direction while you can power through such a crisis, then by setting a long enough time horizon you can consider the competition to be extinct.
It’s a bit like the parable of the two men chased by a bear. One guy starts running, and the other guy asks, “Why would you do that? You can’t outrun a bear.”
The other guy retorts, “No, but I just have to outrun you.”
Or perhaps it’s a business form of persistence hunting. I don’t have to believe more than you, I just have to believe longer.
Movie of the week: Little Women
If we live in the internet of beefs, then it’s no surprise that film has been swept up in the culture wars. Like clockwork, January is now the month to lament the Golden Globe award winners and the Oscar nominees, even if the selections of a small group of obscure international film critics and a largely white male Academy just reflect, as they always have, sampling bias.
That said, count me among those who feel Greta Gerwig deserved a best director Oscar nomination for her adaptation of the Louise May Alcott novel. In keeping with this issue’s theme of time, I’ll begin by highlighting Gerwig’s remix of the novel’s timeline. The novel has a generational resonance, and many a copy has been passed on from mothers to daughters. The nonlinear structure of the film conveys that timelessness in a new way; it unmoors the March sisters from the strict bounds of chronology.
1917 is a more showy film with its stitched one-shot structure, but while that’s undoubtedly a hell of a technical feat (and Roger Deakins a deity of a cinematographer), Little Women packs the greater emotional truth. Many scenes in Gerwig’s film have multiple characters enter and jump right into the ongoing action and conversation, and anyone who has tried directing knows how complicated a problem it is to choreograph that type of staging, even if it isn’t as showy as having a single soldier sprinting across a field as an entire army runs past him.
Saoirse Ronan is a force, as always, and her scenes with Laura Dern epitomize a mother seeing some of herself in her daughter and being perhaps the only person who can ever empathize at such a deep level with her daughter’s pain. “But I’m so lonely!” has never landed harder.
It’s difficult, unless you’ve spent much time in narrative filmmaking, to understand what a rare gift it is to be an actor at the highest levels. It isn’t until you toil through several film school shorts and audition seemingly every aspiring actor in Los Angeles that you realize that most people are unable to make an emotional through-line legible with their face and body.
Little Women is filled with actresses like Ronan and Dern and Florence Pugh who are world-class at conveying, with their faces, a coherent and precise internal narrative, to mix two conflicting emotions and have each register distinctly. That type of expressive craft will never cease to stun me.
The ending of Gerwig’s adaptation also offers one final mystery. Does that concluding marriage occur as it’s shown on screen, or is it a concession to the demands of the marketplace? In some ways, the film ends with a test of the viewer: have you been watching carefully, have you understood the meaning of this tale? Depending on the answer to that question, you’ll interpret the ending one of two ways.
You’ll see what I mean when you see the film, which I recommend you do.