ENTIRELY FORESEEABLE AWS OUTAGES

Summary from the TLDR newsletter

Amazon Web Services experienced at least two minor outages at the end of last year caused by malfunctions in its internal AI tooling. These outages were entirely foreseeable, even by employees. Large language models are extremely complex systems, and this complexity creates real risk. Amazon never misses a chance to point to 'AI' when it is useful, but when there's an outage, suddenly it's just a 'coincidence' that AI tools were involved.