Train Your Brain

Read time 1m50s

Unless you’ve been in an isolation tank for the last week, you have heard of — or been directly affected by — the Microsoft system failure that caused outages from airlines to banks to hospitals. What a mess.

How do these things happen? We have been trained to believe that sending our data and systems to the “cloud,” where big companies like Microsoft, Google, Apple, Meta, and others have the expertise and the sophisticated technologies to keep our information safer than if we each had an extra computer in the house that stored everything for us.

Microsoft outsourced a significant piece of its security infrastructure to another technology behemoth called CrowdStrike.

All hell broke lose as a result of a defective software update that was pushed by CrowdStrike to millions of computers running Microsoft’s Windows.

There was no malicious hacker from an obscure country demanding a billion Bitcoin to fix the problem. There were no tense negotiations to get the bad guys to unlock the data. There was no begging to avoid the release of personal information to the public. This was a self-inflicted wound at CrowdStrike.

The following is from technology writer Ed Zitron and it efficiently summarizes how we have slid into stupidity when it comes to a single point of failure:

So, there it is. The problem with a single point of failure: when it goes, everything goes. There’s a reason you jump with two parachutes. There’s a reason for the net under the trapeze act. There’s a reason that cars have seatbelts and airbags.

Things fail. Materials, mechanical devices, software, people….everything fails at some point.

My favorite example is Theranos, a company with a brash, young founder that claimed it had a device that could run hundreds of medical tests from a pinprick of blood. Almost a billion dollars was invested in a company that had no audited financial statements, no technologies that could be tested, and no external verification of any sort — so people relied on the word of its founder, Elizabeth Holmes. She was the single point of failure and she failed miserably. All the investors’ money was lost and Holmes is currently serving an 11-year prison sentence.

For example, Nvidia is a company that makes the specialized computer chips that power AI systems. Its stock has increased 8X over the last two years and the outlook for AI is even brighter today. Guess what? About 20% of Nvidia’s revenues are from one customer. Lose that customer and Nvidia is a much different business. That’s a single point of failure.

Does your favorite sports team rely on a single player for a disproportionate amount of its scoring? Think about LeBron James being involved in 35% of his team’s points each game. if he fails, the team fails.

Key Takeaways

1. We must design our decision making to look ahead and identify single points of failure and not let FOMO, the push for completion, or competitive forces distort our thinking and enable a single point of failure to be overlooked. As we have seen this week with the CrowdStrike debacle, the consequences can be catastrophic.

2. Study any investment, new project, or decision and find the points of failure. Are new sales reliant on the founder? Will a competitor copy our technology and sell it cheaper? Dig hard until you can identify the points of failure and then decide if you can live with the risks.

3. As much as anything, recognize that identifying a single point of failure is something within your control before you start down the path of an investment or project. Once things are moving, groupthink will take over, the push to finish the project will be overwhelming, and you will be fighting a losing battle if you are flashing warning signs.

For more on Theranos from us, click here.

Ed Zitron writes a cynical and accurate blog on the technology industry. I read his posts as soon as they arrive. For more from Ed, click here.

You might also like: