I believe @gdt is correct, AI seems to be a lot of malarky……
Prompt was - What is the distance from (PT31) 35.685188890 -84.566496030 to (PT33) 35.683247975 -84.566380047 to (PT36) 35.682953718 -84.565805490 to (PT15) 35.676522981 -84.570140749
In SW MAPS the Measure Tool reports (All points were RTK FIX LAT/LON)
PT31 to PT33 707.369ft
PT33 to PT36 201.365ft (BTW I measured this with Tape and attempting to compensate for elevation changes using up more tape, I am pretty sure this is 201.x ft.
PT36 to PT15 2,671.689ft
Here’s what AI’s had to say
Grok said (Elon for the win ?!?!?) Segment distances:
PT31 → PT33: ≈ 216 m (709 ft) PT33 → PT36: ≈ 61 m (201 ft) PT36 → PT15: ≈ 815 m (2,675 ft)
Google Web search AI Mode
The segment-by-segment breakdown is as follows: PT31 to PT33: ~706 feet (0.13 miles) PT33 to PT36: ~201 feet (0.04 miles) PT36 to PT15: ~2,693 feet (0.51 miles)
ChatGPT said PT31 → PT33 Meters: ≈ 216.0 m Feet:≈ 708.6 ft
PT33 → PT36 Meters: ≈ 57.6 m Feet:≈ 189.0 ft
PT36 → PT15 Meters: ≈ 825.6 m Feet:≈ 2,708.6 ft
CLAUDE AI said PT31 to PT33: Distance: approximately 715 feet
PT33 to PT36: Distance: approximately 215 feet
PT36 to PT15: Distance: approximately 2,850 feet
Perplexity said
Using great‑circle (haversine) distance on WGS‑84, the segment distances are approximately:
PT31 about 0.13 miles ≈ 710 ft.
PT33 about 0.04 miles ≈ 200 ft.
PT36 about 0.46 miles ≈ 2,420 ft.
I will need to try out @Andrew74 ‘s new Tool he’s testing at the link…. I am getting great RTK FIX with new PointPerfect RTCM/NTRIP compared to PointPerfect MQTT (eoL Fall 2026), but now I am pondering what is right for my validation of Survey Pins to field measurements. Maybe GROK is pretty good….??
And in going with GROK for the win I asked below…..
FYI the deed measurements I am working on validating are 708.22ft, 201.47ft, 2,677.76ft I am trying to set flags to be close before I spend 20-30 minutes at each point getting very best readings, instead I did about 5 minutes per point (Setup, level Surveyors pole, let GPS settle for 1-2 minutes, then take a 60-80pt average reading)
Prompt to GROK – can you repeat those calculations out to the full reported decimal places in the Latitude and Longitude as I took the readings with a SparkFun RTK Torch with RTCM/NTRIP corrections, and report the results in feet out to two decimal places please?
Here are the straight-line (great-circle) distances calculated using the full reported decimal places from your SparkFun RTK Torch readings (with RTCM/NTRIP corrections for high accuracy). Results are shown to a much higher precision internally, but reported here with meters to 4 decimal places and feet to 2 decimal places as requested.
These are precise “as the crow flies” distances using the haversine formula and the WGS84 Earth model (mean radius ≈ 6371 km). With RTK-level coordinates like yours, the positional accuracy is typically in the centimeter range, so these distances should be very reliable for most practical purposes. If you’re following an actual ground path (trail, road, boundary, etc.), the real-world travel distance will be somewhat longer.
Let me know if you need any other units, bearings/azimuths between points, or further analysis!
Sorry @gdt I thought it was interesting to show what the AI’s produced, as many people out side the SparkFun Community use them for a lot of things, and it proves what you had mentioned in an earlier thread, they can be wildly incorrect. Since I don’t have a background in surveying, and harvest lumber for a living, I have to rely on the imperfections of AI to some degree to learn. Please take no offence from my posts, I am sorry it was impolite of me to @ you . I won’t do it again (@ you) Please have a great day. Chris
For me it’s useful as a brainstorming tool, but not as a source of truth. In many situations AI helps with a Novel approach to a problem since it doesn’t mind making wild suggestions. But you need to be knowledgeable “enough” in the subject to weed out the stupid ones. So for me AI is helpful……sometimes
But it’s usually extremely frustrating during the process.
yes, I have at least learned enough to see when it’s calculations are crazy, but it has helped me learn so for me as a NON-Surveyor, technical farmer, it is helping me understand the flaws and the adjustments. Overall it’s helpful, but you need to learn enough to spot when it is off the rails CRAZY!!!
My other advice is to learn to use QGIS. Being able to bring all your observations, plans and other data into a common view with background imagery is really useful.
