Comment by mikewarot

I'm fairly certain that wouldn't happen. Unless you were to overfit the models until the error were to drop to zero, which would likely take almost infinite time. If you did get that point, you've managed to achieve lossless compression of the training data into the weights of the model.

Given that AI models are randomly initialized with noise, and the goal of training is to avoid overfit, there will always be variance between the weights of models, even if trained from the same data, due to those initial conditions, and chaos theory.

And all of the above, is for the same model architecture. I expect you could do some principle component analysis and come up with a transform to work between models, again if they were overfit to zero error. (After all, that would be a compression engine instead of an AI at that point)

Upon reflection, it seems to me that free Stanford AI course I took a decade ago actually stuck. 8)