Only if you ignore all other facts.
You have to ignore the FACT that four other letters were mailed at the same time, and ALL went to media outlets.
You have to ignore the FACT that the trail of anthrax through the post offices indicates that the letter first went to Lantana where The National Enquirer was previously located.
You have to ignore the FACT that the letter was almost certainly opened by Stephanie Dailey whose job was to open letters addressed to The National Enquirer.
You have to ignore the FACT that if the letter was addressed to AIM instead of The National Enquirer, it would almost certainly have been returned to the Post Office instead of being opened by Stephanie Dailey.
You cannot simply say it could have been addressed to AIM instead of AMI without looking at ALL the facts. Maybe it is still possible, but it is so unlikely that it is in the range between preposterous and ridiculous.
Companys with large mail volumes are handled differently than individuals ~ always have been ~ their carriers don't select and mark up their pieces for forwarding after a move ~ rather, that's done at a secondary distribution that used to be known as a "city case".
Think of it this way, the piece is mailed at Boca Raton post office. At the same time the terrorists buy a new pack of embossed postage envelopes. They load them up the next morning out at the park, then, that evening, drop them in several street collection boxes where they end up getting "lost in the mail" (due to a variety of quite normal operations) and travel in a stack of trays all the way to a major mailer somewhere in central New Jersey.
Those letters are then fed into the regular mailstream over a period of time as those trays are taken off the stack for use.
This explains the origin of the mail, the identity of the terrorists who mailed it, how some of it came to be canceled in New Jersey, and why it was canceled on different dates, and, most importantly, 100% of the cross-contamination that resulted in downstream infections, and in a case in Connecticut, death.
Might have been the other way around too.
Such errors in distribution are common (to the degree errors occur).