AWS fired 40% of DevOps team and a couple days later everything collapsed

Tell me, what can go wrong when you fire 40% of the DevOps team and replace them with AI? Amazon Web Services decided to find out in practice. And literally a couple days after the cuts, a major outage occurred that brought down Snapchat, Roblox, Fortnite, McDonald’s and dozens of other mass resources.

According to information from an internal letter that was quickly published on Amazon’s wiki and just as quickly deleted, the new system supposedly can independently fix outages. As well as restore virtual networks and roll back failed Lambda deployments without human participation. Sounds impressive, right?

The system promised to instantly detect and fix permission errors, rebuild broken subnet configurations and all that. But something went wrong because AWS crashed and dragged thousands of platforms from different industries with it. The servers were restored of course, but a residue remained.

Amazon hasn’t yet confirmed information about cuts and replacement of engineers with AI. But the coincidence is too eloquent: fired people, implemented automation, and a couple days later everything collapsed. It’s the classic story about how personnel savings turn into catastrophe. I’ve been running this channel for almost a year and see many stories about layoffs and replacement of people with AI. And rarely has this succeeded anywhere.