My view is that while a single slip gauge should be within the tolerance one can work too, umpteen together have a cumulative error, so the stack may be different to a single slip of that thickness.
Obviously one uses the minimum number of slips for any particular thickness, but one or two missing slips in the set can increase the number required, thus increasing the total error. Even with errors like that, temperature and any other faults, mine are better than I can (generally) manage – or need.
My set is imperial, old and includes a mix of more than a single name (although by the same manufacturer). Still more than good enough for me! Not yet used them for more than checking my other measuring kit, so far.
To be honest, I think think they are a necessity for a machinist doing work for third parties, but for most ‘internal’ work for ‘one offs’ us hobbyists undertake, making the parts to fit each other is as easy as, or easier than, making the individual parts to a tight size specification.
An example is using the key, as a check for slot width being as good as using a slip gauge. The important part, for us, is to ensure both key slots (internal and external) are the same. Even making a stepped key might be good enough for us non-perfectionists – but not for someone buying in parts from several sources, and expecting them to fit with minimum fettling, would be unacceptable.