The chocolate is mentioned because some time ago people discovered that you can just use a piece of diffraction grating or holographic stickers as mold for molten chocolate and it will transfer the diffraction grating/hologram to chocolate. Now you can buy commercial silicon molds for creating chocolate with holograms, you can also get 3d printer build plates with similar idea. Just reproducing a hologram in DIY environment with easily available household items is unusual, doing it with food items is more amazing. Applied science channel has a video on that as well from few years ago although he wasn't first one to come up with similar idea.
This technique with laser seems to produces the diffraction grating by varying oxide layer thickness not by creating 3d texture so resulting surface is still flat and attempting to use it as mold will not transfer the pattern to chocolate.
The reason many commercially available diffraction gratings have 3d texture (and thus suitable for copying with chocolate) is because stamping a hot piece of metal into plastic is a very cheap way of doing it.
Again (see my other comment in the thread), I thought the oxides were just for color variation and there was depth changes that could be used for a mold.
Anyway, there are still ways of moving forward with the idea. For example, chemically removing the oxide layer to a desired thickness sounds feasible. If I were him, I would try it (but maybe in another video, as the whole process would be a whole different rollercoaster).
Maybe, but stainless may not be the best material for that, because the oxide layer formed is largely chromia, and chromia is a motherfucker, which is why stainless doesn't rust. Etching chromium off the chromia sounds practically difficult but probably feasible; etching chromia while leaving metals sounds hard. Maybe molten sodium hydroxide?
Instead, you could choose a different metal whose oxides are easy to etch. Magnesium is probably the extreme case here, with an oxide that instantly vanishes in the weakest of acids, but if someone gave me a US$7000 fiber laser, I would try to keep the laser beam away from thin pieces of magnesium. But mild steel, for example, forms oxides that etch pretty easily with acids. I think copper oxides also etch easily with either acids or bases, too, and the copper itself is more resistant to etching.
Really, though, if you're molding silicone or chocolate, you don't need the high strength, flexibility, conductivity, etc., of metals. Maybe etch your grating into a material chosen for other properties. Glass, for example, is perfectly isotropic and has no grain structure to introduce into your cuts, and it has a low TCE. It sticks to silicone, but not to chocolate. Fused quartz is a glass with a near-zero TCE. I assume but don't know that the MOPA laser can ablate the glass surface.
Other amorphous solids might be more amenable to easy laser shaping and not stick to silicone. Sugar glass, for example.
Speaking of unreasonably dangerous things, though, it occurs to me that, if you laser-marked metal that was submerged in perchloroethylene (or carbon tet), the oxidation you'd get would be metal chlorides rather than metal oxides, and for almost all metals the chloride is easily etched with just plain water. You probably don't want to try that with aluminum.
Are you worried that the glass might overheat and break, or that it might produce a more dangerous specular reflection than Krasnow's polished stainless foil?
I think the reflections from glass will be about 5% of the reflections from polished stainless steel. So, while in general I strongly endorse your point, I don't think it applies in this context.
Is there any risk of the laser going through the glass and damaging parts of the machine? Or things such as refraction?
As I understand, those machines have a sheet of materials they can work with. This is probably the safest way to go: just stick with the materials the laser is made to safely operate on.
