I thought it was a simple 8 pointing to a perpendicular direction. Apparently the model is a bit more complex.

I don't think complexity in the predictive model is (necessarily) attributable to a more complex bee language. One possible hypothesis is that the language doesn't need to be 100% precise, and that it needs to change over time to prioritize targetted communication rather than expressive communication. I'm not an expert in bee genetics, let's just say it's complex, so it's certainly possible that the bees may be changing or making the dance complex to increase the competition of their peers while not increasing the fitness of bees that share, this is excluding the possibility of solitary spy bees or spy bees from another hive.

Another possibility is that the dance is a simple 8 pointing into a simple direction(possibly with a magnitude) , but the bee's memory and execution isn't perfect and the model is overfitting that imperfection.

If I had to guess, I don't think bees are developing complex languages for complex directions, they don't have streets to follow, are not complicated by blockages or rivers, nor do they have to avoid predators, they can just fly in a single direction.

The spy theory is weak as well, bees are protected by spys by scent, soldiers attack intruders, so inside their hive perimeter there's not much sense in encrpytion as a protection mechanism.

Intra-hive competition is possible? I think a hive is more of a single organism that competes and mates with other hives in most senses. But of course there might be individual or even gene competition. There's a single breeding female, but she can be overturned, which presents a vector for competition. But there's also resource attribution. There's of course no paternity certainty as the sperm is pooled and the eggs are pooled, so there's pooled anonimity that mixes up paternity, but perhaps there are some ways for bees to sniff out kin(even if non deterministically). It's kind of far fetched as it would require that sense of kinship to travel across the chain of worker bees, such that the resource can travel all the way down to the larva, but I can see how there would be an incentive for gene diversity to increase intra-hive competition. There's trophollaxys which might have even more communication and kin-affinity/loyalty, but I figure it wouldn't be trivial to establish whether a pair of bees share a male father or not, or any in betweens.