Sorry for the late reply, didn't have time to read the whole PDF till today.
The safe distance formula on the PDF is slightly different (and gives different results) than other NOHD formulas out there. I'm not sure how much we can actually trust those formulas :/
It also has an interesting formula to compute "The flux of energy through a point at a given distance and angle from a diffusely reflecting surface":
H = Pλ*Q*cos(Φv) / π*r^2
where Pλ is the reflectance at the wavelength in question, Q is the energy of the incident beam, Φv is the angle of observation with respect to the normal, and r is the distance from the point of reflection to the point of observation. For safety calculations this equation can be greatly simplified by assuming Pλ = cos Φv = 1.
According to their example it's safe to view a diffuse reflection of a 2W visible laser from 40cm away for the duration of the blink reflex (0.25s).
If we do the calculation in W using the MPE for continuous exposure (1mW/cm^2) the minimum safe distance for a visible 5W laser would be 40cm. If this is true looking at the diffuse eflection of even the highest power visible lasers from at least that distance should be safe. I'd take the calculation with a grain of salt because the equation is for energy (J), not power (W). It would make sense to assume it holds for power but since this is a safety calculation it would be better to confirm. Also beware that most surfaces also have at least a bit of specular reflection.
Edit: that also means using lasers to burn stuff from short distances without goggles isn't safe even if the surface is completely matte.