Thanks a lot for the extensive feedback, especially if it didn't hit with you the first time. As these are based on real twitter threads, some of them might be distasteful or unfitting for some. Sadly, I didn't have the time to write anything myself :-/
It's a good feedback that it made you feel like nothing mattered. In general, each choice adds a different data point to your chart and your "AI" is trained on these datapoints:
This is the function graph after the first two trainings based on different choices (for speed sake the upper is always picked number 2, the lower always picked number 3). So you can see that the behavior and reaction to the input (x Axis) is completely different between the two. So the development of the patient's values will also be different and based on that also the conversation they have with you.
The waiting time with the patients can be annoying which was the reason I added the fast-forward button. I planned for a more graphical display but given that I had only half the time due to real life, just finishing the core functionality was done ~12 hours before the end...
In general, the patient's input is based on the three values (it picks only ones that have values close to the current ones) and as your answers change those values, the conversation should differ. Of course, depending on your training they might be similar as well. But what values you get and how they change is based on your decision.
I tried to make this visible with the different text, but I think it shows that I should probably should have put less work into writing text that is 80% similar and rather take a more graphical approach to show that the numbers based on your choices are what matter. In your case, it kinda became like a text autobattler with the feeling that the strategy doesn't matter - which is the opposite of what it should be.
Thanks for your feedback, it really helps a lot! :)