The magnetic particles on the cards probably have similar characteristics to those used in ferrite permanent magnets. The demagnetization of the latter has been stated that 1% of the field decays within milliseconds of being magnetized. The next 1% decays in another second. The next 1% decays around a century later(!).
The limit to the lifetime of the data on these cards is likely to be the physical limits - that the oxide layer doesn't scrape off, etc. Keeping them in a cool, dry and intert place should keep them far longer than we will last.
The actual recording technique is simple. There are two tracks per side. A pulse on one track is a binary one; on the other is a binary zero. As such, it is self clocking. The differential nature makes it very noise and level immune. While terribly crude compared to modern magnetic data recording, it is simple, inexpensive and robust.
In the early 1980s, the PPC-UCI Newsletter published a spoof about an HP-41 module with 3141592654 bytes of storage using an incredibly small hard disk drive (thanks for a great article, Joe!). Amazing that micro-SD cards are more than an order of magnitude smaller in size and larger in storage. The old magnetic cards were crude but they were an important step to where we are now.
