I have this problem. When I fill a USB stick with music files, and stick it in the USB port in my new Honda CR-V, it plays all the tracks on an album in order of Title, i.e. alphabetically - its pretty annoying. So, what was I going to do. I could use MP3 tag to prepend the track number to all the titles, which I have already done with most of my 'non-commercial' recordings. Or I could only put my non-commercial recordings (already processed) on the stick.
I'm using an ancient PC to load the stick and am working with AI chatbots to optimize the writing process, which can be painfully slow. I thought to myself, how about asking the Chatbot to find all MP3 files having metadata where the Title does not have the Track number prepended and prepend it. It occurred to me that would be usurping the role of MP3Tag! Maybe this is something I could someday do in MP3 tag, i.e. why select a directory, just go find everything on the drive with a set of metadata patterns and update them. Can we do that?
You can already now load a lot of files from all sorts of folders (recursively) and then filter for files that match a certain pattern, e.g.
Filter: NOT %title% MATCHES ^\d+
I'm sure I can. But I just gave Co-Pilot this prompt: "Can you write a script that updates .mp3 file (and .flac file) metadata? specifically, if the title does not start with the track number, prepend the track number to the title. If the track number is less than 10, prepend a '0'. put a dash between the text of the track number and the original title. So if the third track title is 'How High the Moon', change it to '03-How High the Moon'"
Now I have a powershell script. Probably very similar to the one MP3Tag would write under the covers, without my having to revisit my knowledge of MP3Tag expressions and macros.
This is why the stock market has been very hard on software companies lately. I'm not going to stop using MP3Tag completely, but natural language is a very nice interface.
Such a prompt would require already specialised knowledge to create.
New Mp3tag users don't know anything about "metadata". Most new Mp3tag users don't differentiate between TITLE and _filename, and don't know how to handle tracks with three or four numbers in the TRACK metadata field.
IMHO: I doubt that an "AI mode" for entering tasks in natural language would be helpful for a tool like Mp3tag. Such an "AI" would need to ask many questions to ensure that the desired fields and data are manipulated correctly. Many new users lack patience, unfortunately.
Considering the enormous amount of energy that is required to get an AI answer, plus the effort to write a 74 word prompt I prefer a solution like the filter
plus Convert>Tag-Tag for TITLE
Format string: $num(%track%,2) - %title%
Much more efficient IMHO.
The AI's tendency to hallucinate is also a problem, see e.g. this thread:
The big problem with such a PowerShell script is this:
They don't adhere to any existing standards!
If you, as a PS-developer, do not adhere to the specifications for every music format such as .mp3 or .flac, the result are fields that cannot be read by most software and hardware players.
PowerShell does exactly what you tell it to. It does not "know" about any restrictions or limits unless you code them for each field and music format and test it with various tools for compatibility.
Your bottleneck, therefore, is the USB interface and the connected USB drive - not a lack of features (like AI) that might speed up the process. It strikes me as illogical to upgrade software with AI capabilities simply because one is using a storage medium and interface that are completely unsuitable for mass bulk processing.
Incidentally, for most people, crafting a well-defined prompt for the AI that yields the desired result is likely a greater challenge than entering a somewhat cryptic format string into a field for the converter or an action.