As you know, last week a ship got stuck in the Suez Canal. And by the time you read this, it will most probably remain stuck (UPDATE: It was finally re-floated again). Whatever the circumstances that led to it, the story quickly received an Internet meme status.
Some among us bearing the title "software engineers" were quick to react:
Some were even brave enough to "suggest ways" for saving the day:
Unfortunately, the story stopped being funny once you realized that the world of trade is currently losing around $9.6 Billion daily from this incident.
This situation made me think about the real world and how we programmers see it from our filter bubble. We had long been taught that software mistakes are inevitable but fundamentally easy to go around. A simple restart or getting a spare instance up, and our system is almost as good as new. That is until we either fix the underlying issue or having to "un-clog" our system once again.
Back in the early days, software was very much tied to the hardware it ran onto. That is no longer the case. Virtualization, containerization, and lately, serverless have turned hardware into vapor. In a large portion of cases, where and how software gets executed has become irrelevant, as long as it runs. Slow CPUs, network hiccups, a failing SSD drive? That's what you have the Cloud for. A faster, beefier alternative and redundant backups are one click away.
We learned to take it for granted that most physical objects are robust enough to consider their failure as something akin to a rounding error. Of course, I am vastly exaggerating here - there is an entire segment of our industry ensuring systems reliability. However, that's usually the last thing you or I think about when launching a new website or a mobile app.
And yet, machines fail.
What if the purpose of making software wasn't writing clean and clever code? A Twitter thread I wrote on the last day of my job last week. I will soon repost it on this blog.
Tech reviewer Marques Brownlee (a.k.a, MKBHD) sat down with Beeple to discuss his career as an artist and how things led to him selling an NFT for $69 million a few weeks ago. It's easy to look at this sale as an isolated event and attribute it to the current hype without looking at how much methodical effort Beeople has put into his work over the last close to 15 years. Every single day. It's a super insightful interview, and I warmly recommend it to everyone.
The insider/outsider problem: One of the biggest barriers to good financial regulation is the fact that the specialists who are in the best position to understand Big Finance (and therefore regulate it) are all insiders with conflicted relationships where a lot of money is at stake. The same is true in machine learning.
Related to the cover image of this newsletter, I found this funny generator of in-game death/start messages. You can see it in action below 👇🏼