Sure, but predictions don't occur in a vacuum. I have no trouble believing that within a certain metaphysical framework, multiverse models are predicted, but that doesn't make it relevant outside of that particular metaphysical framework. If the metaphysics are wrong, so is the model, and if you need as convoluted a model as this to make sense of modern physics, perhaps the problem lies in the initial metaphysics.
That's true; all multiverse models come with caveats and assumptions; but this is also true of scientific theories. The reasoning is, roughly, 'if the theory is correct - and it all its testable predictions have been verified - then there's a reasonable expectation that its predictions that we can't currently test are also likely to be correct. This is how, for example, the Higgs boson could be predicted 40 years before we had the technology to observe it.
But the models are not always 'convoluted'. The Schrodinger equation that describes how a quantum system evolves over time is relatively simple and elegant, and the behaviour of particles at quantum scales that it describes is demonstrably counter-intuitive and unlike what happens at the familiar macro-scale; but matches our observations. That's just how it is. The Everettian multiverse is the result of applying the Schrodinger equation to both observer and observed without ad-hoc additions. Maybe additional considerations should apply, but applying them just to avoid the multiverse is hard to justify.
My concern here would be how goal-seeking behavior can be accounted for at all on a materialistic ontology. How does non-teleological behavior give rise to teleological behavior?
As I said previously, concepts like 'goal-seeking' and 'teleological' are just high-level abstractions to describe certain kinds of behaviour. Preciely how you define it, and where you draw the line between what is goal-seeking and what is not goal-seeking behaviour is a matter of interpretation.
For example, in Aristotelian physics, gravity was described in terms of the tendency of objects to seek their 'natural' position. These days we generally reserve it for living things, things that have evolved.
A hypothetical example: when an amoeba moves towards a food source, we can trace the causal chain from the chemical gradient of food molecules that diffuse from the food source to the amoeba's cell membrane and are transported through it, and follow the chain of internal chemical reactions that result in the amoeba moving toward the densest concentration of food molecules. When the amoeba has absorbed and metabolised a certain amount of food, the concentration of internal metabolic by-products 'switches off' the chemical pathways that enable detection of food molecules, or disables movement along food molecule gradients, and suppresses food absorption. [note: this is hypothetical - the actual mechanisms may be different]
It is possible to describe the amoeba's behaviour in high-level teleological terms - it was hungry, so it sought out food and ate until it was sated. In those terms, it would appear to be obvious goal-seeking behaviour, yet, at a lower level, it could be fully described in terms of causal activity (chemistry). The goal would be satisfying hunger, and 'hunger' would be the state where there are insufficient metabolic by-products to disable the pathways that enable the sensing, movement towards, and absorption of food.
Perhaps 'goal-seeking' should only apply to creatures that have a nervous system, or even a brain - an ant, maybe... But even in an ant, it's possible to see how an internal physiological state or an external stimulus (e.g. pheromone) can lead to a certain pattern of behaviour, and how changes in those internal states (e.g. resulting from previous behaviour) or stimuli can change that pattern of behaviour or trigger a new pattern.
The more complex the behaviours and the more complex the brain that generates them, the more difficult it is to trace the causal sequences involved, but I see no reason to assume that, at some level of complexity, behaviour - emergent or otherwise - no longer has causal origins.
If you can propose an example of some human behaviour for which you are unable to imagine a causal history, I'd be interested to hear it - and to see if I can provide a plausible causal history.
We've already hit a serious snag, as we need to account for the existence of pleasure and reward circuitry at all. Pleasure and pain are phenomena; why does the former supervene on one set of circuitry and the latter on a separate set? Why is one good and one bad? We are immediately into the realm of value judgments once positive and negative reinforcement is being invoked, so it seems that hardwiring via natural selection presupposes the prior existence of values.
As I described previously, the values are ultimately those associated with evolutionary selective advantage, i.e. successful reproduction. Pleasure and reward are associated with behaviours that have a selective advantage (are 'good' for us in evolutionary terms), pain, disgust, and other negative sensations are associated with selective disadvantage ('bad' for us in evolutionary terms). From our distant evolutionary history, individuals with behavioural tendencies to avoid damage, infection, extreme risk, etc., were more likely to survive to reproduce and pass on the genetic components of those tendencies; those that tended to seek out high-quality food, mate frequently, protect their family, etc. were more likely to successfully pass on the genetic components of those tendencies.
These behavioural tendencies are typically closely associated with physiological responses - the threat of a predator would trigger a flood of adrenaline and stress hormones, because those responses were, historically, the most successful in supporting fight or flight and for surviving any damage that might occur. These physiological responses are the sensations of fear. They're usually triggered by external stimuli, and in turn trigger emotional states - so the physiological effects associated with fear will make someone feel fearful, etc. There are similar explanations for other physiological/emotional/behavioural states.
It is to be expected that the activities and physiological responses associated with things you have evolved to avoid are aversive, i.e. seen as unpleasant, and the activities and physiological responses associated with things you have evolved to seek out and engage in are attractive, i.e. seen as pleasant.
Social, cultural, & religious indoctrination provides a whole new set of values and associations for what are 'good' and 'bad' behaviours, which may conflict with more basic sensations ("I know it's wrong, but it feels good"), but the brain is flexible and adaptable and can learn to suppress, modify, redirect, or sublimate behaviours considered undesirable.
. You can certainly conceptualize these as ideas, but it just doesn't have the same ring of urgency than substantial reality. I hope you see where I'm going with that. 
