The 4 Craziest Computer Bugs In History

Instagram logo for Matt Bristow's blog LinkedIn logo for Matt Bristow's blog Logo to click to give feedback on Matt Bristows blog.
Brain icon to indicate ability to summarise blog with AI.

Summarise with AI

AI summary

Attention to detail is the difference between good and great.

It’s also the difference between okay and really really bad, and nowhere is that more evident than the world of computer programming.

With the exception of the AI that will eventually bring about the end of life on Earth, most computer programs just diligently fill out whatever the programmer wrote, even if that includes a very slight mistake.

Missing as little as one character in your code can lead to a crazy amount of disruption and destruction as the examples below show.

The Mariner 1 satellite disaster

The Mariner 1 was a satellite that was destroyed shortly after launch due to a software error in its guidance system, leading to a misinterpretation of flight commands and loss of control.

The “one character is all that stands between you and a sateillite dropping on you” example above wasn’t an exaggeration.

In 1962, the very first of the American Mariner program, aptly named “Mariner 1” was launched with the lofty goal of completing a flyby of Venus.

As 38 million miles is way too far to conscientiously send a human (it's a bit cold) this aircraft was unmanned and mainly ran by software, which contributed to the lavish-for-the-time overall price of $18.5 million, which adjusted for inflation stands at a healthy $370 million.

You can tell where this is going.

It’s subject to debate exactly what caused the error but it’s posited that a single hyphen was omitted in the guidance system caused the spacecraft to become uncontrollable. The decision was subsequently made to use a controlled self-destruct (this still worked luckily) to complete the most expensive light show of all time in the skies above Cape Canaveral after 294.5 seconds.

If that isn’t scary enough for you, even the systems designed to protect us have unrecognised errors that are slowly ticking towards disaster like the next example.

The Dharan Missile attack

The Dhahran missile attack occurred during the Gulf War when an Iraqi Scud missile struck a U.S. Army barracks in Dhahran, Saudi Arabia, killing 28 soldiers and injuring 100.

War is increasingly being fought with drones and software rather than humans and bullets.

If you are familiar with the Israel-Palestine conflict that is escalating at the moment, you may have witnessed the Israeli “Iron Dome” defense system, which is a hi-tech way to blow up incoming rockets by shooting a rocket of your own at them.

As you can imagine, it’s an incredibly complex task to hit a rocket travelling at lightning fast speeds with another rocket. The software that runs the Iron Dome detects incoming rockets, measures their speed, calculates their trajectory and then launches a rocket of it’s own, aiming at the place where the rocket will be, all in split seconds.

The Iron Dome is but an example of this kind of defense system, and a similar system was used to defend the US military base in Dharan, Saudi Arabia during the Gulf War.

In 1991, an Iraqi short-range ballistic missile was launched at the barracks, and the system's radar successfully picked it up as it was incoming.

The next step was to calculate where it would be next, and to intercept it. The system however had been operating at that point for 100 hours, during which a software bug affecting the systems internal clock meant it had drifted by 1/3rd of a second, causing the system to “look” in the wrong place, and no defensive missile was fired.

28 soldiers from the U.S. Army's 14th Quartermaster Detachment lost their lives on impact.

Increasingly, our safety and the safety of our loved ones relies on computer code, and no where was it more apparent than February 25th 1991 in Dharan.

A digital pandemic within World of Warcraft

On to a slightly less depressing example, but one that absolutely blew my mind when I found out about it.

The Corrupted Blood incident was a virtual pandemic in World of Warcraft, caused by a glitched disease that unexpectedly spread from players to others, mimicking real-world epidemic behaviors.

If you are unfamiliar, World of Warcraft is a massive multiplayer online role-playing game where users complete quests in a huge shared world.

To keep things interesting, the creators of World of Warcraft, Blizzard Entertainment, release new quests periodically for players to complete. In September 2005, they released a new boss called Hakkar the Soulflayer into the game, and yes I will be changing my name by the time the week is out.

Hakkar, as a video game boss tends to do, had his own special attack called “Corrupted Blood” that would slowly drain the players life.

Here’s the catch.

“Corrupted Blood” could be passed from one player to another, so if you were standing too close to your buddy when it happened, you would also be affected by this health sucking digital 'pathogen'.

Now, Hakkar was confined to his dungeon in the game, so the devs thought it would only affect players who entered the dungeon.

How wrong they were.

Despite players not being able to carry “Corrupted Blood” out of the dungeon, their in-game pets could, due to a developer oversight.

This meant that “Corrupted Blood” started spreading throughout the game like a real world pandemic, killing players all over the map.

Players eventually abandoned densely populated cities, enforced social distancing and some even volunteered to heal those affected with in game spells and potions.

It was so eerily reminiscent of an actual outbreak that epidemiologists study the in-game reactions of players to gather insight into potential reactions to an outbreak.

Knight Capital lose $7 billion in one hour on the stock market

Knight Capital lost $440 million in 2012 due to a software glitch in its trading algorithms, which executed unintended trades in 45 minutes, devastating the company's finances.

Ever messed up at work and broken an expensive product, dropped a pint or ordered too much of something with company money? Ever wondered what the worst possible version of that looks like? Here it is!

In 2012, American global financial services firm Knight Capital adopted a new software to help them conduct stock market trades at scale, as at that point they were handling more than 3.3 billion trades in a single day.

This software, as you will have come to have expected, contained a bug.

In the first hour of trading on August 1st 2012, this software bought an eye watering $7 billion worth of stocks on the open market.

Knight Capital tried to get the trades cancelled, but were unable to as in the court’s eyes, it would unfairly penalise the traders who sold Knight Capital the stocks.

This meant that Knight Capital had to sell these shares and absorbed a brutal $450 million loss, all due to an errant few lines of code that ran for about 45 minutes.

Logo to click to leave a comment on this blog.

Load comments


No comments yet, be the first!



Post comment