Technological issues
Stack Exchange works hard for its communities, but there is a limit. If thingiverse changes its API (which it can, and does, do at any time without notice) then anything we build on it has to be modified, and users will be very whiny until it's fixed.
If we choose to integrate model viewing or any sort of API integration with an outside service, we will have to make sure the service provides some sort of legacy guarantee minimal API that won't change, and then we will only be able to use that.
Political issues
I would be very uncomfortable endorsing a closed source, commercial service on this site. There are many communities within the world of 3D printing that are very unhappy with makerbot, thingiverse, and their parent company, and encouraging users to use this service may result in pushing away some very valuable experts within the reprap world.
Looking to the future
If we integrate any sort of model library or model viewer, I strongly suggest we choose an open source project the community can heartily endorse. We may link to outside models, but the core API should be servable by stack exchange without license or much maintenance so 1) we don't have to deal with API changes messing us up, 2) we don't rely on outside services that may change their terms of use or API, 3) we don't have to form complicated contract bound relationships with 3rd parties just to get the guarantees we'd need to convince Stack Exchange to invest in a new feature, and 4) we don't become too politically involved such that we disenfranchise users who may have very strong feelings about this subject.
Wait and see
This is the early beta. Let's find out what people link to in questions before dedicating resources to improving the links/plug-ins/oneboxes. It may be that thingiverse objects are linked to frequently, and that it would add value to the site to improve that experience, but it's also just as likely that thingiverse turns out to be infrequently linked and the effort spent now would essentially have been wasted. Let's wait until we have tens of thousands of questions so we can properly evaluate what the community actually uses and needs.