I suspect you just see that as bringing additional challenges that you don't know how to overcome.
At least for the challenges I'm immediately concerned with, I know how to overcome them.1 However atomic splits done this way require each intermediate recipient have:
incoming liquidity equivalent to
their share of the payment
the shares of recipients later in the route
outgoing liquidity equivalent to
the payment shares of recipients later in the route
In addition to being reliably online.
Footnotes
I think about this problem in the context of taking an SN-like sybil fee, ie a third party verifying a payment happened and taking a fee without additional interaction from the sender and receiver. The third party can do this by "wrapping" a bolt11 from the destination in a hold invoice paid to itself. This would work for multiple third parties too with each wrapping the bolt11 of one of the other third parties. ↩
Footnotes