For possiblity of upgrading to a 808nm diode on steroids and have no failure, the quick answer is both maybe and no. The reason being, the compound DPSS crystals have the limits as to how well they handle the pump light. 400mW diode may pop the crystals, or instantly turn the glues to liquid from heating of the KTP crystal. 100mW DPSS laser usually use slightly large DPSS crystals, with Nd : YAG (or Nd : VYO4) disc on separate heatsink close to the diode window, so the thermal issue is mostly taken care of. It's the KTP that gets pretty warm in operation.
Switching out a diode in a DPSS wasn't a problem for me as I usually figure out where to line up from dot on the crystal in parallel of the diode facet before putting them completely together. However, I do agree it's a b*tch to line the DPSS up, the nonlinear / laser crystals are surprisingly sensitive to axis polarization of the light intended to pump them. Sometimes you have to slightly rotate the diode slightly out of phase to the crystal to hunt for the most efficient absorption of the 808nm light by Nd : VYO4 crystal so resultant 1,064 can be doubled into 532nm, which requires lot of patience and access to the laser power meter.
As for 400mW diode in a puny laser, it's not worth it.