I still wouldn't do it, you aren't going to turn a miyata into an Eisentraut. And I wouldn't count on the construction to be perfect either. On the subject of stress risers, the lug is going to be a stress riser no matter what you do. The material under the lug is constrained, and its sitting right next to material that isn't constrained the same way. It's an obvious stress riser. You could thin to zero right at the edge, but what's the fun in that? Even then there would be a stress riser.

I may post a picture of the frame with very thin lugs that cracked. If any frame was going to be saved by thinning the lugs, that is the one, the lugs are very thin. But logically, even a radically thinned lug is going to more than double the thickness of the tubing, and lugs almost always have a couple of sharp points. This frame cracked on the head tube starting at a point on the lug. I figure there were a multitude of sins in the construction of that frame, but having a relatively flexible lug didn't help. Paint job was awesome though. I respect Tom Kellogg as a framebuilder, but a marketing spiel is not exactly proof of anything. I bet I can go on the wayback machine and find any number of defunct web sites of newbies that say they are the best framebuilder ever -- maybe including the guy that made the frame that I fixed that cracked. I have a Ph.D. and my thesis involved studying fatigue. But I wouldn't write such a thing because I am not going to do the testing required to validate the conclusions made. That's the danger of education, you become more cautious about publishing poorly informed opinion

Someone cut apart a batch of older frames to look at the workmanship, it was uniformly bad even on some fairly well regarded brands. I wish I could find the pictures. The old guys considered the lugs to be the structural element that carried the loads. The idea that there should be an internal fillet that connects the tubes independent of the lug is a fairly recent development.