Yes, some sounds are harder than others and take longer for babies to learn. (I imagine the R is difficult for Russian children.) Regionally, many adult native English speakers do not pronounce a standard TH; substituting a stop (like T, D — traditional US East Coast accents, Ireland) or a labial (like F, V — popular in Estuary English). In America, if a school child pronounces TH as F/V, then a good school may send the child to a remedial pronunciation class.
I’d recommend that the learner stick to the standard variant. Not only does that sound “proper” (nonstandard is associated with low social class), but it will allow the learner to adapt to nonstandard variations if necessary. Starting with a nonstandard pronunciation would be a problem, as the learner wouldn’t have a concept of the TH as a separate phoneme (or phonemes, counting the voiced and unvoiced TH).