I would say that that would probably be safe, but holding wire down to multiple tiny contacts on a board with tape can be more work than just doing the soldering, especially if you're dealing with 4 separate contacts/pins and 2 separate wires like that, and debri like flux that might prevent good conduction.
Have you got any flux? Those solder blobs look a little dry. It does look like the pads are long gone, but you could try reflowing with a little flux quick too and seeing if that changes anything, as if there's anything left of the pad, still connected to the trace, the pins may just not be making a good connection. I imagine it wouldn't hurt anything at this point to try it.
I believe that he's saying to connect the pin that he's connected to the diode with a blue line to that end of the diode (with a wire), as that trace leads directly back to it at that location (as if his blue line is the wire, terminating at the contacts you should solder it to on either end), and test. If that doesn't fix it, then short any of the pads highlighted in red to the contact or pad you previously bridged with the blue wire (this you could do easy enough with a scrap bit of wire, just be careful what you bridge, not that it would be easy to screw up, even with really shaky hands). Once you find one that outputs a space when shorted to your blue wire, solder a wire from the switch leg highlighted in orange to the red pad/leg that outputted a space when it was shorted to the blue wire.
A keyboard switch is a momentary switch, so what he described is just finding another place that same trace leads to (since the one that leads directly to the pad you lifted is not on the bottom layer of the pcb), and then once you've found it, jump it to the orange switch leg by soldering a wire between them so that your space switch will once again be what shorts the two together, instead of the wire that you used to find the right trace.
Does that make sense?