a 90 million dollar mistake

I hope your week isn’t going as badly as the startup who accidentally gave 90 million dollars to its users, and resorted to threats in an effort to get it back.

Or as badly as Facebook, which took itself offline for hours on Monday to the point where its staff couldn’t even enter buildings to help solve the problem.

Both of these mega-messups were due to simple errors in software, but they had big implications. They’re a reminder that while tech wows us with its capabilities, it’s often incredibly fragile and can be brought to its knees by nothing more than a typo or a coder having an off-day. More on this below.

Meanwhile over the past week, I have been:

Okay, let’s dive into the week’s most consequential tech news, shall we?

— Martin.

When the New York Times published an opinion piece on Monday morning saying the recent leaked documents showed Facebook was fragile, it didn’t know just how fragile it would prove to be a few hours later, in quite a different way.

If you’re not up-to-date with how it happened, the plain-English explanation for the huge Facebook outage was that a simple mistake in a configuration update led to the internet not being able to find Facebook services whenever anyone tried to access them. And because Facebook uses its own services for much of its internal communication, and even as authentication for opening physical doors in its buildings, the response took a lot longer than it might otherwise have done.

The modern world is dizzyingly complex, but at the heart of it is humans. Humans who make mistakes. Humans who assume other humans have thought about the problems they spot but aren’t responsible for fixing. Humans who—due to time pressures, budget restraints, or laziness—hold crucial software together with the coding equivalent of string and sticky tape.

It’s easy to shrug off a Facebook outage as not very important, but if you live in a country where Facebook or WhatsApp is the main communications infrastructure, you probably had a grim time on Monday.

The headlines shouting about how Mark Zuckerberg ‘lost’ $6bn of net worth in hours during the outage due to stock price drops highlight another problem. Mark Zuckerberg wasn’t meaningfully impacted by that drop. No-one was. The people who obsess over ‘world’s richest person’ charts will have got excited, but that’s about it.

So we had a problem caused by weak infrastructure that underpins a company often presented as far too strong. And that led to a fake problem with fake money, which was really nothing more than numbers on a spreadsheet triggering a narrative about wealth.

Much of the world as we know it today is actually layer upon layer of narrative about how the world really works. When we see the cracks in those layers, it can be disarming. But if all those layers were all stripped away, would really have much of a world left?

Another example of ‘everything is narrative’ came late yesterday, when Mark Zuckerberg responded to the Facebook whistleblower’s hearing in front of Congress.

Facebook’s PR approach to the hearing has been to not directly address any of the accusations and instead present former employee Frances Haugen as an unreliable witness. Zuckerberg’s own post reads quite persuasively until you realise that it doesn’t really say that much about recent controversies - it just deflects from them.

But yes, everything is narrative. I suspect Haugen’s testimony as a clear, credible communicator on a topic both US parties believe needs addressing, could be the catalyst politicians needed to do something, no matter how accurate her evidence may or may not be.

What they actually do to curb Facebook’s power is the big question. It could be toothless legislation or it could reform free speech in a way that hurts US citizens and people far beyond its borders. That remains to be seen, but for now Facebook is distinctly losing the communications war.

Last week I discussed how Amazon staff had expressed reservations about its new in-home robot, but the company had launched it anyway. If any staff flagged up obvious problems with its new gifting feature, they must have been similarly ignored.

As the Verge explains:

I thought of several ways this could be abused within seconds of reading about it, and Amazon is staffed by smart people so they must have thought of them too… and yet other priorities seemingly won out over user safety.

User safety should be a primary concern before a big tech company launches a new product. The days of ‘launch quick and figure it out as we go along’ are mostly over. User behaviour and expectations around tech have matured, and big companies launching problematic products shouldn’t be excused anymore.

Any product like this should have to go through a user safety gatekeeper at Amazon who signs it off as not having the potential to make people’s lives less safe. Surely Amazon owes the world that at a minimum? I mean, they didn’t even build in a way for someone to opt out of being sent gifts by strangers! It’s maddening, it really is.

More big reads:

…and that’s everything I’ve got for you this fine Wednesday. If you enjoyed this newsletter, please share it with someone you think will like it.

See you in in your inbox exactly one week from now, give or take a few minutes.

— Martin