What complicated problems does this introduce? It’s been a while since I ran a Tor node (I didn’t like my home IP getting blocked from some websites even for running a relay node), but unless they made it more confusing since then it takes some deliberate configuration to get it running as an exit node. As for getting the implementation correct so IPFS doesn’t leak info while using Tor, I believe that’s what the audit referred to in the thread posted before is meant to address.
IPFS is meant to be transport agnostic; so rather than trying to saddle the protocol itself with the performance overhead and weaknesses of one specific anonymous transport implementation, it can theoretically be made to work with any network (e.g., I2P, Tor) that attempts to provide anonymity if the user needs it.
Link to related thread: Does IPFS provide any guarantees about anonymity?