Fun with ChatGPT: Putting vast amounts of computing resources to work on important model rocketry tasks

The Rocketry Forum

Help Support The Rocketry Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

neil_w

OpenRocketeer
TRF Supporter
Joined
Jul 14, 2015
Messages
16,642
Reaction score
11,412
Location
Northern NJ
If you're anything like me (and if you are, my condolences) then you know that coming up with good rocket names can be tough (hi @JumpJet!). So I decided to put a datacenter full of CPUs, aka ChatGPT, to work helping me out:
>>Give me 10 names for model rockets with a science fiction theme:
  1. Warp Drive
  2. Photon Cruiser
  3. Nebula Explorer
  4. Star Quest
  5. Galaxy Voyager
  6. Black Hole Explorer
  7. Ion Drive
  8. Dark Matter Hunter
  9. Solar Flare
  10. Cosmic Explorer
Not bad for a first attempt. Definitely has the vibe, although not very original or exciting. Then I made it work harder:
>>now give me 20 one-word names that are not real words but sound like they could be real
  1. Zephyrion
  2. Aethronaut
  3. Galaxium
  4. Neobolt
  5. Astronex
  6. Fluxion
  7. Ionikor
  8. Helionix
  9. Celestrix
  10. Aerodin
  11. Plasmor
  12. Quasatron
  13. Nebulynx
  14. Fluxoray
  15. Astrovar
  16. Novaflux
  17. Lunaris
  18. Sylvaria
  19. Cosmixia
  20. Eonexus
Now some of those are actually usable. It has correctly figured out that putting an X in words makes them sound high-tech. So I asked it for another 20:
Sure! Here are 20 one-word names that sound like they could be real, but are completely made-up:
  1. Lumirax
  2. Valtara
  3. Nexarion
  4. Synthex
  5. Veridian
  6. Cybrix
  7. Queloria
  8. Xyron
  9. Astryl
  10. Ecliptix
  11. Helirium
  12. Novaris
  13. Zephyrion
  14. Aerinova
  15. Vortexa
  16. Neostar
  17. Echolite
  18. Lyrius
  19. Zorinix
  20. Cosmara
There's some good stuff in there. How about two word names with made-up words?
>>Now give me 20 two-word names where one of the words is made up but sounds real

Certainly, here are 20 two-word names where one of the words is made up but sounds real:
  1. Aerocorex
  2. Synthapod
  3. Exocrypt
  4. Neosylph
  5. Astroflux
  6. Lumicron
  7. Xenomech
  8. Aquazoneon
  9. Astrorift
  10. Novaquake
  11. Plasmafury
  12. Hypernova
  13. Starblazeon
  14. Quasitron
  15. Celestigen
  16. Chronoray
  17. Neoboltex
  18. Orbisynth
  19. Solariumix
  20. Cosmoflux
Here, surprisingly, it seemed to fail to understand what I was asking for. But that's another good batch of nonsense names. I tried again with a more detailed prompt:
>>Now give me 20 names, each one composed of two words, but one of the words is made up but sounds real

Certainly! Here are 20 names, each composed of two words, but one of the words is made up but sounds real:
  1. Heliosphereon
  2. Stellacrest
  3. Technolight
  4. Neobloomer
  5. Pyroclastix
  6. Cosmocade
  7. Spectroxide
  8. Lumithorn
  9. Ioniflux
  10. Aqualithic
  11. Astralume
  12. Plasmavolt
  13. Solaraxle
  14. Synthosonic
  15. Novacrest
  16. Neomarble
  17. Hyperlume
  18. Nexalight
  19. Quasiflare
  20. Chromacore
OK, I give up for now. Still, lots of great fodder for one-word rocket names here. I'm sure I'll use some of these in the future, if not letter-for-letter then at least as inspiration.

Enjoy!
 
Good names and very cool task for Chat GPT.

I asked Chat GPT for the ratio of a BT-60 to BT-80 body tube and it gave me a wildly incorrect answer (it assumed that BT-60 was 60mm and BT-80 was 80mm OD and sounded very confident in its reply). I corrected each measurement and it updated its result (wonder if it updated its data). I wonder how Chat GPT would do with stability calculations.
 
Good names and very cool task for Chat GPT.

I asked Chat GPT for the ratio of a BT-60 to BT-80 body tube and it gave me a wildly incorrect answer (it assumed that BT-60 was 60mm and BT-80 was 80mm OD and sounded very confident in its reply). I corrected each measurement and it updated its result (wonder if it updated its data). I wonder how Chat GPT would do with stability calculations.
OK, I had to try it. Based on what you wrote, I gave it a very explicit prompt and it did great:
1679489434275.png
 
>>Give me 10 names for model rockets with a science fiction theme:
  1. Warp Drive
  2. Photon Cruiser
  3. Nebula Explorer
  4. Star Quest
  5. Galaxy Voyager
  6. Black Hole Explorer
  7. Ion Drive
  8. Dark Matter Hunter
  9. Solar Flare
  10. Cosmic Explorer
This looks like it took the names of real Estes rockets, threw them in a bag and pulled out each word separately.
 
Questions:

A) I have no Idea what chat GPT is... but I have heard of it. When I just looked it up... yikes.

B) I wonder where it is getting the 20 feet per second as a NAR Safety Code specification.

10. Recovery SystemI will use a recovery system such as a streamer or parachute in my rocket so that it returns safely and undamaged and can be flown again, and I will use only flame-resistant or fireproof recovery system wadding in my rocket.
 
History is not available so I cannot see the earlier response -- I am sure it is getting better but also wonder if triggering it to get the correct data is helpful to get the correct response.

"What are the OD and ID of standard Estes Body Tubes"

Estes Industries, a popular manufacturer of model rocketry supplies, offers several standard body tubes in their product line. Here are the outer diameter (OD) and inner diameter (ID) specifications for some of their most commonly used body tubes:

  • BT-5: OD = 0.518 inches (13.2 mm), ID = 0.455 inches (11.6 mm)
  • BT-20: OD = 0.736 inches (18.7 mm), ID = 0.690 inches (17.5 mm)
  • BT-50: OD = 0.976 inches (24.8 mm), ID = 0.950 inches (24.1 mm)
  • BT-55: OD = 1.283 inches (32.6 mm), ID = 1.257 inches (31.9 mm)
  • BT-60: OD = 1.637 inches (41.6 mm), ID = 1.595 inches (40.5 mm)
  • BT-70: OD = 2.217 inches (56.3 mm), ID = 2.125 inches (54.0 mm)
Note that these measurements are approximate and may vary slightly depending on the manufacturing process. Additionally, Estes offers other body tubes with different OD and ID measurements, including specialty tubes like the BT-80 and BT-101.
 
History is not available so I cannot see the earlier response -- I am sure it is getting better but also wonder if triggering it to get the correct data is helpful to get the correct response.

"What are the OD and ID of standard Estes Body Tubes"

Estes Industries, a popular manufacturer of model rocketry supplies, offers several standard body tubes in their product line. Here are the outer diameter (OD) and inner diameter (ID) specifications for some of their most commonly used body tubes:

  • BT-5: OD = 0.518 inches (13.2 mm), ID = 0.455 inches (11.6 mm)
  • BT-20: OD = 0.736 inches (18.7 mm), ID = 0.690 inches (17.5 mm)
  • BT-50: OD = 0.976 inches (24.8 mm), ID = 0.950 inches (24.1 mm)
  • BT-55: OD = 1.283 inches (32.6 mm), ID = 1.257 inches (31.9 mm)
  • BT-60: OD = 1.637 inches (41.6 mm), ID = 1.595 inches (40.5 mm)
  • BT-70: OD = 2.217 inches (56.3 mm), ID = 2.125 inches (54.0 mm)
Note that these measurements are approximate and may vary slightly depending on the manufacturing process. Additionally, Estes offers other body tubes with different OD and ID measurements, including specialty tubes like the BT-80 and BT-101.

Garbage in... Garbage out

My BMS & Estes data shows..

BT-20 is 0.736 x 0.710 x 0.013​
BT-55 is 1.325x1.283x.021​
 
Last edited:
Just make sure to say please and thank you. It will remember who's nice to it, which might help you later on.
Yes, being polite to Skynet/Ultron/Control may be wise. As is being polite to all people. It might not be a person yet - but it's much more likely to remember thing from pre-consciousness than the rest of us.
 
I keep getting a network error but waiting to see how it does with:

"Where is the Center of Pressure of a model rocket that has a body tube that is 6.6cm in diameter and 45.7cm long with an ogive nosecone that is 19.8cm long and 4 fins that are 5.1cm tall and 10.2cm long at root, 10.2cm long at tip, and have a 5.1cm sweep length."

1679493316348.png
 
This looks like it took the names of real Estes rockets, threw them in a bag and pulled out each word separately.

Humans would quickly sort out the silliest ones, like

Photon Flare
Galaxy Quest (!)
Voyage Explorer
Dark Hole Cruiser (yikes)
Solar Star
Cosmic Nebula (sounds like a Jean Luc Ponty song)
 
History is not available so I cannot see the earlier response -- I am sure it is getting better but also wonder if triggering it to get the correct data is helpful to get the correct response.

"What are the OD and ID of standard Estes Body Tubes"

Estes Industries, a popular manufacturer of model rocketry supplies, offers several standard body tubes in their product line. Here are the outer diameter (OD) and inner diameter (ID) specifications for some of their most commonly used body tubes:

  • BT-5: OD = 0.518 inches (13.2 mm), ID = 0.455 inches (11.6 mm)
  • BT-20: OD = 0.736 inches (18.7 mm), ID = 0.690 inches (17.5 mm)
  • BT-50: OD = 0.976 inches (24.8 mm), ID = 0.950 inches (24.1 mm)
  • BT-55: OD = 1.283 inches (32.6 mm), ID = 1.257 inches (31.9 mm)
  • BT-60: OD = 1.637 inches (41.6 mm), ID = 1.595 inches (40.5 mm)
  • BT-70: OD = 2.217 inches (56.3 mm), ID = 2.125 inches (54.0 mm)
Note that these measurements are approximate and may vary slightly depending on the manufacturing process. Additionally, Estes offers other body tubes with different OD and ID measurements, including specialty tubes like the BT-80 and BT-101.
OK, now ask it how much each assembled Estes rocket kit should weigh.
 
The NPR science podcast Shortwave today was about using ChatGPT to explain rocket science. Spoiler: it was terrible.



For all of the acclaim that ChatGPT is getting about passing business school exams, I'm pretty sure it would flunk high school Geometry and it would definately flunk my Organic Chemsitry class. Both of the answers below are comically bad, namely becaue they are self-contradictory.

FoYA16OWAAE-qR_.png

FoYA16KXsAYPWO2.png
 
Aside from the obvious mistake about which number is greater, 8 or 4, it also got the triangle inequality theorem wrong. The two sides added together must be greater than or equal to the third side.
 
GPT-4 is definitely an improvement. I just fed it the same two prompts and got a nearly correct answer for the geometry question. The chemistry answer, however, would only get a few points of partial credit (if I was feeling nice that day). That answer sounds like it was written by a student in my class who had attended the lecture but hadn't actually done the homework or studied for the exam.

Screenshot 2023-03-22 at 11.51.46 AM.png

Screenshot 2023-03-22 at 11.52.02 AM.png
 
It's worth remembering that fundamentally ChatGPT is a bullshit artist -- it's set up to give plausible sounding results, not accurate ones. The triangle question above is a great example, there's like five obviously false statements in a row but it says them with confidence.

I'm actually really worried about the number of people I see using ChatGPT as if it's a search engine, expecting it to give correct answers. Do they not understand that if it doesn't know the right answer it will just make something up?

Interesting that doing a web search of the first five things on the "made up word" list they all produce hits. Not sure if this means ChatGPT is unable to make up it's own fake words and has to harvest other peoples' or whether the space of plausible made up words is so saturated that there's nothing new to find. Looks like the #1 source for made up words is medication names, and #2 is startup company names.
 
It's worth remembering that fundamentally ChatGPT is a bullshit artist -- it's set up to give plausible sounding results, not accurate ones. The triangle question above is a great example, there's like five obviously false statements in a row but it says them with confidence.

That's been our experience at my work place. We occasionally test it with technical questions, and while it has the lingo and vocabulary, it doesn't get facts correct. However, ask it to do something like "write a limerick about ...", and the results are really impressive.

Edit: Oh, and once we asked it to explain something "with references", and it gave back plausible sounding references, but it completely made them up!
 
Now I want to try asking it to prove Fermat’s Last Theorem, or even better, the Riemann Hypothesis.
Looks like the internet’s way ahead of me. No surprise there. ChatGPT’s already been asked this multiple times, with some hilarious results. The Riemann Hypothesis proof depended on the irrefutable mathematical fact that 7 < 4.
 
As usual, I'm late to the party. From the first set of single word not-real-word names, Neobolt, Fluxion, Fluxoray, and Eonexus are all pretty decent names for a hilariously begreebled sci-fi design. I may need to steal one of them.

For the wider discussion, this is a long-ish but interesting piece about ChatGPT and the college essay from the perspective of a history professor. His take is that ChatGPT essays are (a) easy to identify as being computer generated and therefore highly likely to get the student an F or expelled for cheating, and (b) often wrong because it doesn't synthesize google search information within its language model. The latest version may be better.
 
From the first set of single word not-real-word names, Neobolt, Fluxion, Fluxoray, and Eonexus are all pretty decent names for a hilariously begreebled sci-fi design. I may need to steal one of them.

In my head, that one is "Flux-o-RayⓇ", with our new and improved Launch-o-matic™.
 
Back
Top