Writing to you from Los Angeles, CA. I spent the weekend in Mammoth making the most of spring riding conditions and an unusually warm weekend on the mountain. I also had an interesting experience during my drive home that I thought was worth sharing:
It happened about three and a half hours into the drive from Mammoth back to Los Angeles. At this point in the journey, we’re settled in. The radio station has been decided, the snacks are out, and everyone is ready to get back home.
Suddenly, the traffic slowed to a crawl. By the time we stopped, the line of tail lights in front of us stretched nearly a mile. A few people immediately turned off the road and crossed the median to start going the other way. Clearly, they weren’t going to stick around to see what the commotion was all about.
Of course, we checked Google Maps to tell us what to do. It reported a slight accident, no more than a ten-minute slowdown. Satisfied, we decided to wait it out.
Ten minutes stretched into twenty, which became another half hour. The traffic showed no signs of letting up. In fact, it seemed to be getting worse. As we inched our way along the highway, we thought about those cars that had turned off the road at the first sign of traffic. Did they know something we didn’t?
When we finally reached the front of the line we saw… nothing. The road was fine. No accident, just a line of cars all getting off at the same exit. The highway was closed for a quarter-mile strip, although there was nothing that should have caused such a monumental traffic jam.
Confused, I checked Google Maps again. With the same self-assured tone, it told me that the wait would be no more than ten minutes — the same estimate it had showed nearly forty minutes earlier.
Eventually, we made our way around the incident and continued our ride home. Over an hour had passed and, based on the honking, people weren’t happy about the slowdown. As we got back on track I realized what had happened.
What should have been a minor delay was exacerbated by the fact that everyone was relying on the same software to give them directions. The traffic jam was caused by hundreds of cars all trying to take the same algorithmically chosen route. Since the highway was closed, the new “optimal” route was a small dirt road that was never designed to handle a wave of cars all trying to use it at the same time.
Without the software, we would have taken a slightly less conventional, but ultimately faster route. Although the evidence was right in front of us, we couldn’t disobey the confident prediction that had been right so many times in the past.
It normally takes about twenty-five seconds to travel a quarter-mile on the highway. Instead, it took us nearly an hour to travel the same distance — an increase of 144 times.
Software is an excellent tool for optimizing ordinary situations. You can trust it to find you the fastest route on any given day. However, when the environment is irregular, you should ignore it. In fact, irregular situations are often the ones that matter the most. The “average” time it takes to travel a quarter-mile doesn’t matter much when you’ve been sitting in your car for an hour, wondering why the prediction was spectacularly wrong.
As software continues to optimize our lives, we need to recognize situations where our intuition is more valuable than algorithmic confidence.
What’s new from me
I’ve been enjoying writing book summaries lately. They help me internalize the main lessons from things I’m reading by putting them in my own words. Here are two of my recent favorites:
In this week’s edition
Occasionally, I stumble upon a corner of the internet that is filled with interesting and underrated ideas. More often than not, these pockets of fresh air are the digital homes of brilliant people who have been silently publishing their thoughts for years.
Bert Hubert is a technologist that writes about biology. He has written about COVID-19, DNA, and the process of invention. I first discovered his website earlier this year when he wrote a fantastic post explaining the technology behind mRNA vaccines.
His latest post is about the complicated relationship between money and innovation. He discusses why you can’t solve every problem by throwing money at it, and why some solutions are bred from the necessity of improvisation that comes from a lack of resources.
Lindy score: 2028
Satoshi Nakamoto, the pseudonymous creator of Bitcoin, holds over 1 million bitcoins across various wallets. The value of their fortune would be worth over $50 billion dollars, placing them just behind the founders of Google on the Forbes list.
However, these bitcoins have never been moved. They were originally mined on an ordinary laptop in the early days of the Bitcoin network before it was overtaken by professional operations with specialized hardware. The movement of these coins is even listed as a risk factor in Coinbase’s S-1 document:
the identification of Satoshi Nakamoto, the pseudonymous person or persons who developed Bitcoin, or the transfer of Satoshi’s Bitcoins;
This post from 2013 is one of the earliest public analyses of the wallets that supposedly belong to Satoshi. At the time of writing, the Bitcoin community was still relatively small, but the value of the coins was in the millions of dollars.
The author describes their method for identifying the wallets associated with Satoshi’s mining hardware. It involves identifying the specific hash rate associated with the mining account that sent the first Bitcoin transactions. At the time, it was an extreme amount of effort to uncover what seemed like a minor mystery.
The most surprising part of the whole story: we still have no idea who Satoshi is. When the post was written, nearly know who knew (or cared) that the creator of the Bitcoin software was anonymous. Now, their identity has risen to an almost religious status.
In a moment of optimism, the author finished their post with these words:
One of the consequences of these graph is that if the real name of the sender of a single transaction belonging to the entity is identified, then Satoshi mystery identity will be revealed. I bet that this will happen in the days following this post.
Lindy score: 2029
Worse Is Better is an essay written by software designer Richard Gabriel. It outlines why it’s better to start with a small piece of software that gains functionality over time (now called a “Minimum Viable Product” or MVP).
Gabriel gave a talk by the same name where he explained why software, like genes, selects for traits with high fitness, even if they contain parts that are worse than their competitors. This became the basis for the idea of “worse is better” which is now implicitly part of every technology startup’s culture.
He describes Unix and C as “the ultimate computer viruses” because they convince their human operators to install them on nearly every computer they use! This is despite serious flaws in both their completeness and security.
Much like a virus, software undergoes selective pressure to improve over time. So, it’s better to quickly release “bad” software with at least one very strong positive attribute and improve it over time, rather than waiting for it to be perfect.
From the essay:
It is important to remember that the initial virus has to be basically good. If so, the viral spread is assured as long as it is portable. Once the virus has spread, there will be pressure to improve it, possibly by increasing its functionality closer to 90%, but users have already been conditioned to accept worse than the right thing. Therefore, the worse-is-better software first will gain acceptance, second will condition its users to expect less, and third will be improved to a point that is almost the right thing.
Next time you’re worrying about polishing your Magnus Opus before releasing it, just remember: sometimes, worse is better.
Lindy score: 2053
Have a great week,