Hmm, why not? For the astable configuration, you could use a 100F capacitor with R1 = 10 Meg and R2 = 7.5 Meg, for a 55 year period. Base current for the Threshold NPN will come from the Trigger PNP (and hopefully temperature drift matches OK). Other than maybe the 100F capacitor might have some variation in capacitance and leakage current over the course of 55 years ;-)

https://www.digikey.com/en/products/detail/tecate-group/PBLH...

Easier with a calendar reminder

"Trivial" But then you realize you forgot to account for a 32 bit counter wrapping up. Or potential failures in the power supply or other capacitors

Every EEPROM is basically that, and they're designed for data retention of around 100 years. I imagine it wouldn't be hard - embed two metal plates in glass?