It mostly has to do with accuracy of the simulations. The raytracing engine within Radiance, which is what Daysim is derived from, has been thoroughly validated over the last three decades. Example 1 (1989): https://eetd.lbl.gov/sites/all/files/publications/lbid-1575.pdf, Example 2 (2018): https://escholarship.org/uc/item/24h966pp . You can find nearly three dozen such validation studies.
As far as using artificial lighting in Honeybee is concerned, it really depends on what your end goals are. If you are interested in just doing an energy simulation, then E+ will suffice. If you are interested in doing an accurate lighting simulation, then use the daylighting module in Honeybee with the IES components.
I think this video is still relevant:
You should be able to find a relevant Hydra example in the description.