What's the difference between automatic and manual LDA hyperparameter tuning?

I'm trying to use Gensim implementation of LDA which suggests using automatic learning of hyperparameters alpha and eta:

We set alpha = 'auto' and eta = 'auto'. Again this is somewhat technical, but essentially we are automatically learning two parameters in the model that we usually would have to specify explicitly.

However, after seeing this article about LDA hyperparameter tuning, I can see that it is also possible to tune these parameters as black-box: train the model with different fixed values of parameters, and then select the best one:

Let’s call the function, and iterate it over the range of topics, alpha, and beta parameter values

Is there any essential difference between these two methods? Is there any special case when the second method is better than the first one?