Podcast studio transcription seems less capable now.

Did the podcast studio transcription also receive a version upgrade like the Enhance tool?  I edit a wildlife statistics podcast and Studio usually does a great job transcribing common names for species and statistical terms.  It also captured the "umms" very well, where I could easily delete those "umms" out of the audio using the transcription delete.  

It is the same hosts talking the way they usually do, same mics, etc.  but the Studio transcription is not on point.  For example, the host says "birding community" and the audio clearly states "birding community" but Studio is returning "British community" or "Swainsons warbler" returns as "Somewhat what are".  Again, Studio has not returned these type of errors before just seeing a big increase.  Are others seeing this as well?