Y2K + 1
I did a lot of Y2K remediations, starting in late 1998. It was fun work. I was working in PCAO at the time, and there were a lot of small embedded computers in plants and industrial facilities. These had 2 digit decimal year counters that would overflow on the change of year. Predictable, but bad behavior.
A common scheme in the industries I coded in was to add a firmware patch that would change the year to 1972 and add an offset (28 in this case) as the time was output. The offset was, in most cases, separate from the year counter and was added on every output of dates.
At one company, one of the critical control machines created a canonical date for all of the other machines in the shop. So they installed the firmware, the year changed, the firmware changed the date to 1972 and added 28 to the offset. Of course, there was code to mitigate the Feb 29 date.
Fast forward to the end of 2000. The firmware had not been updated. So, the year changed, the firmware set the year to 1972 and added 28 to the offset register. The clocks told all of the other computers it was now January 01 2028 (1972 + 56). A few minutes after midnight, the machines started running their cleanup code; one step was to delete any database records not created in the most recent ten years. By the time anybody figured out what had happened, the databases were empty and starting to collect data for 2028. It took many hours of hard work to fix that particular issue.
I was not responsible for anything in that domain, so I got to just sit back and watch the pandemonium.
I first ran into the 2038 problem when I was writing device driver code for handling expiration dates on magnetic tapes. That one scares me. I knew from my work that the Y2K issues were largely hype.
If I am alive in 2038, I will be in a bunker somewhere with food, water, and a way to adequately protect me and my family.