Is anyone using AI to find targets

A little more specific.

ME

"Can you provide a list of nebulae located more than 35 degrees above the horizon and visible in June at the 47th parallel and suitable for focal lengths ranging from 135mm to 400mm"

CHATGPTSure, here is a list of nebulae that should be visible from the 47th parallel in June, located more than 35 degrees above the horizon and suitable for focal lengths ranging from 135mm to 400mm:

 

1. North America Nebula (NGC 7000)
• Constellation: Cygnus
• Approximate Altitude: High in the sky, typically above 60 degrees.
• Focal Length: 135mm to 200mm


2. Pelican Nebula (IC 5070)
• Constellation: Cygnus
• Approximate Altitude: High in the sky, typically above 60 degrees.
• Focal Length: 200mm to 400mm


3. Veil Nebula (NGC 6960, NGC 6992)
• Constellation: Cygnus
• Approximate Altitude: High in the sky, typically above 60 degrees.
• Focal Length: 200mm to 400mm


4. Crescent Nebula (NGC 6888)
• Constellation: Cygnus
• Approximate Altitude: High in the sky, typically above 60 degrees.
• Focal Length: 300mm to 400mm


5. Lagoon Nebula (M8)
• Constellation: Sagittarius
• Approximate Altitude: Around 35-45 degrees above the horizon.
• Focal Length: 135mm to 300mm


6. Trifid Nebula (M20)
• Constellation: Sagittarius
• Approximate Altitude: Around 35-45 degrees above the horizon.
• Focal Length: 200mm to 400mm


7. Omega Nebula (M17)
• Constellation: Sagittarius
• Approximate Altitude: Around 35-45 degrees above the horizon.
• Focal Length: 200mm to 400mm


8. Eagle Nebula (M16)
• Constellation: Serpens
• Approximate Altitude: Around 35-45 degrees above the horizon.
• Focal Length: 200mm to 400mm


9. Ring Nebula (M57)
• Constellation: Lyra
• Approximate Altitude: High in the sky, typically above 50 degrees.
• Focal Length: 200mm to 400mm


10. Dumbbell Nebula (M27)
• Constellation: Vulpecula
• Approximate Altitude: High in the sky, typically above 50 degrees.
• Focal Length: 200mm to 400mm

These nebulae should provide excellent targets for your focal lengths, offering a range of sizes and complexities to capture. Be sure to use a star chart or an astronomy app to confirm their exact positions and best viewing times.

I agree, I don't really see the value when comparing it to the likes of Telescopius or Stellarium for that matter (I prefer the latter one). One of the more enjoyable things for me during off season (may-august) is to scout new projects off the beaten path in Stellarium and/or Telescopius and making notes. I get that this part isn't for everyone, but it adds another dimension to the hobby. I for one am not in a hurry to have lots of objects handed to me by AI, I would find it rather stressfull. The examples also seemed to be pretty basic well known objects. After a while you know which constellations are present or when they will become present. Just like any other map really, it gets familiar. 

But to each their own!

Eric Gagne:
I had no idea Telescopius could do that.  To be honest I am not a big fan of it, I know it is very powerful but I find the interface quite cumbersome.  Maybe I should set time aside to take another good look at it.

IA = Intelligence Artificielle..................I forgot to translate it,  I fixed it

They are extremely useful tools. I agree with you on Telescopius though, I much prefer Stellarium myself - but the one thing Telescopius has which is a huge advantage is the possibility to frame objects but also create mosaics and actually export plans. Those plans can be imported to the Asiair, which I use, and executed from there. Stellarium also offers framing, and if you use NINA you can transfer those framings into NINA, but as far as I know you can't export anything to a file or make mosaic plans there. 

What I've started doing is using Stellarium like I always have, but planning the projects in Telescopius for exporting it

Yes I've been using it for a couple years now. The only issue with the AA is that it's sky atlas is utterly rubbish - at best. Might as well be star hopping as far as I'm concerned

You won't be able to see most of the stuff out there except for the really obvious objects which is a shame. I'd love for them to implement surveys in there. Which in return would make framing and creating mosaics in there viable. 

​​​​​​

Thats why we use stellarium with HIPS…

In order to find targets, I'm using an iOS app called Observer Pro. What I specifically like about it is that 1) it takes into account my custom horizon (no need to import it in a certain format, the app lets you capture it with your phone camera - and export it to SkySafari and other formats!) 2) it takes into account the DSO magnitude, altittude, location vs the moon, moon phase to calculate its visibility, 3) it calculates the total available imaging time based on the custom horizon, 3) then computes quality imaging time for each DS per night. I also like the fact that it gives you a view of the same info by day and by month (so that you know if this object will become more or less visible in a few days or month).

It also has a FoV simulator based on DSS images (hence not great, but OK)

After many research online this is the one that fits all my needs.

Gael Gibert:
In order to find targets, I'm using an iOS app called Observer Pro. What I specifically like about it is that 1) it takes into account my custom horizon (no need to import it in a certain format, the app lets you capture it with your phone camera - and export it to SkySafari and other formats!) 2) it takes into account the DSO magnitude, altittude, location vs the moon, moon phase to calculate its visibility, 3) it calculates the total available imaging time based on the custom horizon, 3) then computes quality imaging time for each DS per night. I also like the fact that it gives you a view of the same info by day and by month (so that you know if this object will become more or less visible in a few days or month).

It also has a FoV simulator based on DSS images (hence not great, but OK)

After many research online this is the one that fits all my needs.

I can fully agree with that. I've been using it for several years and it's one of the best pieces of software for astrophotographers.

Eric Gagne:

• Give me a list of targets suitable for a range of focal length like I asked ChatGPT.   Why would I spend hours looking at hundreds of DSO objects only to find out that over 95% of them I cannot shoot with my 135mm and 400mm scopes ?
• Filter on Az/Alt.  Yes it has RA and Dec but those aren't very friendly.  From my backyard I can only from WNW to ENE.  I could ask AI to give me a list of targets that are visible tonight in that part of the sky.

This takes seconds if not a minute on Telescopius. I think one just needs to get familiar with the UI once to make it practical. Every parameter you can use is already there. I define my scope and it remembers my scope. Then for that day, I define my parameters: Show me objects 30 degrees above the sky at least X hours, but also these should be brighter than this amount and the size should be between this and this. Doing this is only a few seconds for me.

ChatGPT is a LLM, it is not capable of such calculations. The only reason it is giving correct answers at a time is because that specific information was 'luckily' involved in the training set somehow. It might give you a completely wrong answer as well. Because it does not calculate when you ask that question, it predicts depending on the common words in the training set. 

We tried even training it for simple physics calculations, let alone doing spherical geometry calculations that require a complete dataset of objects. It fails the calculation, this is obvious from the experiments and no surprise because of what it is. And this area lacks the required training set.

It will fail you if you rely on that. Use it if you find it practical, but always check if it gives you the correct answer.

*See what I refer to as 'luckily': https://x.com/petergyang/status/1793480607198323196

Kay Ogetay:
=14pxChatGPT is a LLM, it is not capable of such calculations. The only reason it is giving correct answers at a time is because that specific information was 'luckily' involved in the training set somehow. It might give you a completely wrong answer as well. Because it does not calculate when you ask that question, it predicts depending on the common words in the training set.

The responses in the original posts aren't even correct. M16, M8 and M20 are all well below 35 degrees from 47 degree North. (M8 and M20 don't go higher than ~20 degrees above the horizon.)

Xavier Ducrohet:

Kay Ogetay:
=14pxChatGPT is a LLM, it is not capable of such calculations. The only reason it is giving correct answers at a time is because that specific information was 'luckily' involved in the training set somehow. It might give you a completely wrong answer as well. Because it does not calculate when you ask that question, it predicts depending on the common words in the training set.

The responses in the original posts aren't even correct. M16, M8 and M20 are all well below 35 degrees from 47 degree North. (M8 and M20 don't go higher than ~20 degrees above the horizon.)

No surprise. I'd be very surprised if it was all correct. I didn't make that claim because I did not check. Thanks for looking into that. 

There is a way to use it though. If you know what you are doing, you can ask for a Python script that makes the calculation. But I have to say, it'll require a lot of testing and debugging. You can make your own Telescopius-like script. And for many, of course, this isn't a more convenient way than just using the Telescopius.

Scott Badger:
Another resource is Gary Imm's Compendium. Old school spreadsheet, but as or more powerful than anything else, I think, and a truly massive endeavor.
https://www.garyimm.com/compendium

Cheers,
Scott

Yes, this is  what I do.  Ai has produce pedestrian results.

There is quite literally zero benefit to using LLMs and generative AI to select targets for you, it doesn't have any concept of the sky and it cannot recommend anything to you that it hasn't been trained on (Or that it finds on the internet). This feels like another unnecessary "I used AI because I can" type of deal with no real justification or rationale. One Google search of "Beginner targets northern hemisphere" would yield the same results without any potential for hallucinations.