Command prompts and scripting is fine for IT-professional working with servers and database-systems but there is a reson why we got graphical user interfaces like the one on Mac, Windows and there were others too like for example Atari ST TOS (The Operating System) that was based on Digital Research GEM (Graphical Environment Manager) ) and Amiga. For a while Atari was big in the gaming and music industry (It had built in MIDI-ports )
Started in the eighties too with Commodore VIC 64 and even gave cources in schools for some years using CPM 86-based special school computers and character based interfaces and programs. Later PC:s and MS-DOS like everybody else. I really hated those black screens with green or amber text.
I liked the Mac-interface with its black on white interface but never became a Mac-user because in the industry there were only MS-DOS, MS-DOS/Windows that finally transformed into Windows in version 3.
Started with SQL Server 1.0 (a version made by Ashton Tate and not Miccrosoft) and worked with Windows Servers (Windows NT) over SQL Server-databases in more than 30 years. I have used windows from its very ugly days of version 1.03 I think it was. It was truely terrible until version 3 came.
Today I just want to use my photo-applications on Windows 11 to work without any problems and finally after all these Photolab 9 problems they do but it took an upgrade to a new tailormade PC especially designed to cop e with all the new AI-models I want to use.
It is a bit fun to see that this mentality that computer software and services has to be free even has a strong grip in the AI-world.
It’s a bit funny to see stinginess deceive wisdom so often even in this new beautiful AI world just like the free altruistic Linux on desktop world. When I run the AI-powered iMatch Autotagger in my DAM system with Open AI’s huge API over the internet, it gives the best quality for Description metadata and that at a very reasonable cost. There is no hint of a problem doing it with 6GB or 8 on the graphics card.
While the stingy instead runs significantly less powerful and worse free models locally on LM Studio and/or Ollama and then discovers that the graphics card suddenly, just like with AI masking in Photolab 9, has become a hopeless bottleneck that requires a new graphics card with at least 16 GB for about 1000 U$ to work. On top of that, 32 GB of RAM is needed, which has just increased in price from 120 U$ to 500 U$. When the graphics card then requires more power than your old power supply can handle, it too has to be replaced or even the entire computer 
Sometimes I actually wonder why it is so unusual for RAW converters to offer the option to use commercial APIs as an option. Then people would at least have a choice. With iMatch DAM, you have a lot of choice when it comes to AI support. You can run both locally for free and over the internet with commercial APIs. You can even opt out of American Open AI or Google and choose, for example, Mistral, which is developed in France instead. Maybe something for Alec, who seems to have a bit of a hard time with American companies.
I haven’t really gotten there yet, but I’m thinking of trying Mistral too. My most important photo programs like Photolab (French), Capture One (Danish), iMatch (German) and XnView (French) are all European.