So today I was following up on a question I got from the Chief Engineer during a presentation I gave him last week, about unstable aerodynamics on a vehicle we're doing insight work for. I had come to some conclusions about its uniqueness regarding the question he was asking and ran them by one of the department's subject matter experts:

Aero: So I found the behavior the Chief engineer wanted me to investigate in the flight test data. I was hoping I could look at some heritage programs and see similar behavior, but I'm not having much luck. Here's what I want to run by him before I take it to the chief engineer:

1) Vehicles A, B, C, and D are 'most similar' to the current program but C never flew and none of them are similar enough to be particularly useful. (He reminded me that C flew an unpowered partial-scale drop test) I later remembered vehicle E which was another similar vehicle, and vehicle F which is a VERY dissimilar vehicle but flies in a very relevant flight regime and almost certainly experiences unstable aerodynamic flowfields similar to our

2) There are no other relevant vehicles to our problem in the open literature (except a Soviet model that never flew, is very much like 'A' - since they copied A to build it, and we probably couldn't get the info for it anyway)

3) If I'm correct about 1 and 2, could he point me to a couple of good references that I can pull numbers from (our vehicle maxes out at Mach 2.5, say, and I can pull the relevant numbers on the others from to show 'see, not the same'

He suggests I use ChatGPT for 3. Just ask it my question and ask for links to relevant papers. It will summarize the papers and I can verify directly in the papers if the numbers it spits out are alright.

So I first open up the NASA-managed, NASA-provided AI interface I've been told to use. I ask it my first question about vehicle A, with the intention of working my way down the list. It said it couldn't answer my question, and had basically 3 suggestions for me:

1) If I shared some relevant papers with it, it could read through them and see if they answered my questions.

2) I could visit my local research library and find some papers, if I hadn't yet done so.

3) I could talk to relevant experts in the field and see what they think.

Oh! Thank you soooo much! That's very helpful.

So I next followed the SME's advice and used ChatGPT. It linked me to several relevant papers and websites, gave good summaries of each paper to tell me their relevance, and I worked through my vehicles one at a time until I proved to myself that I was correct about all of them and had numbers to go in a table to summarize the tale.