TL:DR - yes, it's a bit out there, but no more so than any other of the big attempts.

I've talked with Jesper and Johannes at length whilst I was a PhD student - their ideas are based on applying the techniques of loop quantum gravity to non-commutative geometry. To give a brief summary of each:

LQG regards the basic variables of geometry to be holonomies and fluxes - a holonomy is the transport of a vector around a small loop, coming back to the start to find the vector isn't pointing the same way (think about carrying an arrow around the a triangle from north pole to equator). This measures the curvature of the underlying manifold. The fluxes are like field lines in electromagnetism. It is these variables that are quantized (discretized) on a spin-network in LQG.

Non-commutative geometry is the idea that geometrical operators care about the order in which they are applied - area(A) length(B) != length(B) area(A) (very loosely). Non-commutativity is at the heart of quantum mechanics, and is the root of Heisenberg's Uncertainty Principle.

What they're hoping to do is build on the work of Connes and Chamseddine who have shown that the spectral action (special type of object in a non-commutative geometry, coming from application to the standard model) naturally reproduces the Einstein-Hilbert action (Basis of General Relatvity) in certain conditions. They hope that by applying LQG techniques here they'll get a full quantum theory of everything.

It's a long shot, of course, but all such things are - non commutative geometry is a strange beast, and no-one has shown that LQG is the right way to quantize gravity (though they have had some theoretical success in cosmology and black holes). It's a personal aesthetic as to whether you think this is more or less plausible than extra dimensions, or symmetries, or some altogether new principle. It's not something I choose to spend my time on as I don't think it's the right way to go (I don't like non-commutativity, and LQG involves fundamental discreteness in a way that I think doesn't work) but I would say it's as good an idea as any other on the market and deserves to be explored.