Discover more from Shanti’s Newsletter
Chapter 2: Power Laws and Disruption
Think faster to seize the initiative
Rethink the org chart
Saw this on Twitter long ago: Optimize your startup for some metric that’s the responsibility of two vice presidents at an incumbent, and not only will they not see you coming, they won’t do anything to stop you.
The next big thing
Something really useful I learned at a conference
Applying that to coaching Colin’s soccer team
Subscribe to get more like this in your inbox.
The CFO of a space startup company cold-called me this week. After an hour on the phone, we’d outlined the capital strategy for a business spinoff, and identified the key risks that need to be retired to allow it to move forward. After two hours on Zoom with the founders of a two-sided marketplace, we had a Wardley map showing how to prioritize work, and a new catchphrase. Call me (+1626-354-9919) and I’ll figure out an unexpected way for you to improve your business. If it helps, pay me what you think the advice is worth to you. You can also hire me to bring order to your requirements flowdown (that system is for another newsletter) or for technical due diligence on potential investments.
The Next Big Thing
Carbice Corp. will make computation a little bit cheaper for everyone, by making it simpler to dispose of waste heat.
Carbice is a spinoff of Georgia Tech, where Prof. Baratunde “Bara” Cola and his students developed a new application for carbon nanotubes. I first met Bara several years ago at a Starburst event, and immediately thought, “I need this yesterday.” The product, Carbice Carbon, changes the design, engineering, and manufacturing, and packaging of electronics. I introduced Bara to space friends at Northrop Grumman, Planet, and Capella, and lobbied NASA* to look at Carbice. No luck on the last one, so I repeatedly begged Bara to let me invest.
I recently saw, in a mission kickoff meeting, “Use Carbice as much as possible.”
Follow Carbice on Linkedin and ❤️ their big reveals this week.
*At least 3 different section managers are involved in putting a circuit board into a box at JPL. No one of them could improve the process.
What are you trying to do? Carbice makes a thermally conductive, electrically insulating, mechanically compliant tape. They call it SpacePad. It’s a thin metal foil coated with forests of carbon nanotubes on both sides. You stick it between heat-generating integrated circuit and a piece of metal structure that can suck away the waste heat.
How is it done today, and what are the limits of current practice? Heat from space electronics is conducted through the component pins into the circuit board, then through a copper layer in the board, to the edges, through wedge-locks, into the chassis body, through a bolted interface, and finally into the structure. There are scores of piece parts, the chassis is heavy, and because the conduction path is so convoluted, circuit boards tend to overheat. There are gap-filler materials, which you can squish between a circuit board and a metal panel, but these are not particularly conductive, they degrade after a year or two, and you can’t open the box to do rework. Similarly, thermally conductive greases are single-use and not stretchy.
What's new in your approach and why do you think it will be successful? Carbice Carbon is just a molecule, so it’s good for an infinite number of compression / expansion cycles. It doesn’t lose elasticity as it outgases over time. You can use it as a gap filler to remove heat directly from the hot spots of a circuit board. With fewer thermal interfaces, heat rejection systems can be less overdesigned.
What difference will it make? Spacecraft projects that design around Carbice Carbon will save hundreds of thousands of dollars in total engineering costs, and reduce delays during system integration. Between doing less thermal engineering, and being able to simply unscrew a panel to access a circuit board, a whole lot of labor and uncertainty go away. Autonomous cars can accommodate more computation with the same heat rejection systems. Fewer lines in the delta-T budget mean that mobile processors can run just a smidge faster.
What are the risks and the payoffs? Carbice is building its first factory in Atlanta for large-scale production to move beyond the high-value applications (space, chip manufacturing) and serve high volume industries. To fully take advantage of Carbice Carbon, you’d want to rethink how circuit boards are laid out and packaged (see my fine powerpoint art), which also means rethinking the engineering teams.
At the first Smallsat Symposium, where we wondered why we all paid $800 to wear a suit in the parking lot of Hogan Lovells, Steve Jurvetson explained his theory of change: the cost of computation has been exponentially decreasing for 3,000 years (Moore’s Law), which means that you can now calculate optimal solutions, reasonably quickly, to problems that were previously handled through heuristics. See, for example, Moneyball. Change is the only constant.
Andy Grove’s book High Output Management is about sustaining functional teams at Intel. Grove describes spending his days being entertained in meetings that ran on time, and smacking down anyone who stepped outside their stovepipe. He made sure the machine kept itself well-oiled. Grove stresses that while integrated product teams move faster, only functional organizations can scale. Must be nice to have a monopoly on stolen IP. What the book doesn’t say is that relying on size to suppress competition leads you to forget about disruptive innovation. Oops.
In general, thanks to computation continuing to get cheaper, a fast-moving, customer-focused team can run circles around a dinosaur before the dinosaur even knows it’s being competed with. This is all just Moneyball (computation > heuristics) and Moore’s Law (tomorrow > today).
Winning the OODA loop
When I started refereeing my kids’ soccer teams, about 8 years ago, I noticed that most of the advice kids got from coaches and parents wasn’t particularly useful. Of course, I’ve been taking notes all along. Now, I don’t know the “right way” to coach soccer, but I do know a bit about neuroscience, so I came up with a theory from first principles. You can read all about it if you have a few hours. To summarize, I teach the kids to create surprises, so the opponents are so busy adapting that they don’t have time to compute a plan.
Colin allows me to coach his 10-year-old soccer team, The Unicorns. This is an AYSO All-Star team, meaning they’re the strongest players of their age from the town rec league, and they play in a circuit of similar skill. Nobody practices toe taps on their own if they could play Minecraft instead, so it’s about teamwork, not virtuosity. The plan must be working, because the Unicorns have seriously outperformed expectations. Here’s Colin in goal, looking unconcerned about Arcadia’s striker, who will get mugged by the defense before he can shift his weight to the other foot to take a shot on goal, if he could even do that with his left foot.
You will often hear a frustrated soccer coach yell, “talk to each other!” Well, how are they supposed to know how to do that? Did the team spend time practicing how to talk? Simon Wardley defines doctrine as how you “avoid hurting yourself,” and his #1 recommendation is to use a common language. So, as the Unicorns learn each new concept, we decide on a new vocabulary word, and drill on using it.
Sparkle! - "man on"
Drop - call for a drop pass
Rainbow - high looping pass, usually crossing to the side
Garbage - sweep wide to collect missed crosses
Magic - give and go
Glitter - offside trap
Crash - run to the goal
Dice Four - corner kick set piece
Zing - low fast cross to center
Cupcake! - I screwed up, so everybody readjust
That last one is probably the most important – to normalize mistakes. We follow Bob Ross’s rule: there are no mistakes, only happy accidents. So, when the goalkeeper punts to start a counterattack, only the ball shanks wide and out of bounds, he yells, “Cupcake,” and everyone looks around for threats.
Soccer is big in Diamond Bar, a modern bedroom community in the hills east of LA. They even let the sports teams use the parks, and they even maintain the fields so there’s actual grass. It’s a delightful place to play.
In one of the games, we heard the opposing coach instruct his team, “everyone remember to play your position.” The boys grinned, because they know how to punish people who rely on that heuristic. While the purple team spread out to their positions for the kickoff, the Unicorns all huddle near the bottom of the center circle. The purple team kicked off with a pass to a hapless winger, who was immediately run down by 5 Unicorns. While the purple coach was trying to joystick his players to counter our herd, the Unicorns dynamically swapped positions to create localized imbalances, and their coach (me) chewed contentedly on a kale salad. At one point, the Unicorns’ left defender ended up at the right post of the other goal, so our attacking midfielder dropped back to the sweeper position. All the while, the purple team players divided their attention among their opponents and the apoplectic commands from their coach.
What does that have to do with this letter’s theme? The 9 kids on the field can outcompute 1 coach.
I was cleaning out my Google Drive, because I’m too
cheap frugal to pay for multiple cloud storage services. Sure enough, I discovered a few days later that I’d deleted the wrong folder. After a few moments of panic and terror, I realized that this must happen all the time. Turns out, Google can recover deleted files through this obscure URL: https://support.google.com/drive?p=file_recovery.
Suggestion: Download your most important Google Docs to local storage.
More episodes: Start from Chapter 0