Can't edit Direct Lake Semantic Model to reduce size: Resource Governing error

We have a P1 capactiy with a Lakehouse.  I started building a custom Semantic Model, and then (when trying to add a Measure), got the following error:

Resource Governing: This operation was canceled because there wasn't enough memory to finish running it. Either reduce the memory footprint of your dataset by doing things such as limiting the amount of imported data, or if using Power BI Premium, increase the memory of the Premium capacity where this dataset is hosted.

I thought I can drop one of the tables in my model or hide some columns.  But now I can't make ANY changes to the model because I keep getting the same error.  If the model is already over size, how do I change it to reduce the size?

UPDATE:  The next day, I noticed (using Bravo for Power BI) that the size of my semantic model was significantly reduced.  I'm guessing it has to do with some data getting evicted from memory.  During my initial testing, I was using a report that pulled back a lot of records, so it chewed up all the available memory.  But I guess it reset somehow (through the magic of Direct Lake), and I was able to edit it today.

I also came up with another workaround by using Power BI ALM Toolkit to merge part of my "too big" model into a mostly blank model (after enabling writing via the XMLA endpoint).