Using AI to help you plan

[Former member 108819]

I’ll preface this with a disclaimer that I share a lot of common concerns about generative artificial intelligence, including its inability to attribute sources or discern correctness, and its potential to disrupt human workforce. Use it at your own discretion.

If you don’t know what generative AI is, in a nutshell, it takes prompts or questions from you and then estimates answers to your query based on language probability accrued from internet sources. It doesn’t have the ability to determine if what it is saying is correct, so often you need to double check it, but it can be a decent weight on decision making.

Using ChatGPT as a tool for camino planning

ChatGPT is free to use and pretty intuitive. Like googling, researching forums, or talking to friends, it can be a useful tool to help make decisions about your trip planning. Here are some ways I used ChatGPT to help me plan my camino.

1. Weight estimate: I present it a list of items, often including the brand name or details, and it will tell me the estimated weight of each item, as well as the total. It can suggest what you may want to think about leaving behind.

Example: “I own a 34L Osprey Stratos backpack. Will all this fit and what will be the approximate weight including the backpack?”

2. Stages: I present it the length of my trip and tell me will I have enough time to make it.

Example: “I am hiking the Camino del Norte and want to complete it in 30 days. Estimate the km I should aim to complete per day.”

3. Decision making: I had a recent flight disruption and needed to weigh some decisions.

Example: “I am flying to Spain to hike the Camino del Norte for 30 days as part of my sabbatical. My flight to London is potentially disrupted due to a pilots strike. Here are my options: Option 1 - blah blah. Option 2 - blah blah. Which option seems best, and what options may be missing?”

4. Weather: I want to know the temperature range I will face in many different locales.

Example: “I’m hiking from Irún to Santiago de Compostela in May-June. What temperatures will I experience high and low in each region?”

Et cetera!

I hope this is helpful to someone.

 

[Tincatinker]

If you want to really exercise ChatGPT try asking it "What is the melting point of a chocolate fireguard?" Followed by "How would you determine your usefulness compared to a chocolate fireguard?"

Remember, ChatGPT learns more from you than you will ever learn from it.

 

[Turga]

Seems to me to be a fast track to becoming helpless and un-resourceful. Next thing you know, you can have some robot walking the Camino for you following it in real-time via satellite-link from your armchair. Perhaps it could pass as ‘vicarie pro’?

 

[Bradypus]

Sledgehammer. Nut.

 

[David Tallan]

FWIW, there was someone asking questions in the CP question that seemed to be prompted by their doing route planning using ChatGPT.

The questions made no sense, seeming to imply that the Spiritual Variant occurs before the Coastal joins up with the Central route, for example. It was only later that it emerged what research had been done to inform the questions.

The dangers of ChatGPT as a planning tool.

 

[Former member 108819]

I imagine the responses may indicate an unfamiliarity with AI, or perhaps a generational gap. I work in tech so I’m happy to answer any questions!

For example, generative AI is already heavily embedded in your life: cellphones, email, Google Maps, Siri, Google search, Gmail, google docs, spreadsheets, ordering food, your local bank, advertisements, newspapers, magazines, language translation, accessibility for blind, seeing and hearing impaired — these all use generative and general AI to varying degrees. And it’s extremely likely that the words you write here are scraped by AI as well.

The idea that you can you can put your head in the sand about it, or ignore it and it will go away, or that people who use it are backwards — just isn’t realistic. It is a part of our world and will become a bigger part over time, with speed.

I’m happy to answer any questions or help with queries. However if this becomes a debate I recommend the mods shut the thread down as that wasn’t the point.

 

[Annette london]

I’ll preface this with a disclaimer that I share a lot of common concerns about generative artificial intelligence, including its inability to attribute sources or discern correctness, and its potential to disrupt human workforce. Use it at your own discretion.

If you don’t know what generative AI is, in a nutshell, it takes prompts or questions from you and then estimates answers to your query based on language probability accrued from internet sources. It doesn’t have the ability to determine if what it is saying is correct, so often you need to double check it, but it can be a decent weight on decision making.

Using ChatGPT as a tool for camino planning

ChatGPT is free to use and pretty intuitive. Like googling, researching forums, or talking to friends, it can be a useful tool to help make decisions about your trip planning. Here are some ways I used ChatGPT to help me plan my camino.

1. Weight estimate: I present it a list of items, often including the brand name or details, and it will tell me the estimated weight of each item, as well as the total. It can suggest what you may want to think about leaving behind.

Example: “I own a 34L Osprey Stratos backpack. Will all this fit and what will be the approximate weight including the backpack?”

2. Stages: I present it the length of my trip and tell me will I have enough time to make it.

Example: “I am hiking the Camino del Norte and want to complete it in 30 days. Estimate the km I should aim to complete per day.”

3. Decision making: I had a recent flight disruption and needed to weigh some decisions.

Example: “I am flying to Spain to hike the Camino del Norte for 30 days as part of my sabbatical. My flight to London is potentially disrupted due to a pilots strike. Here are my options: Option 1 - blah blah. Option 2 - blah blah. Which option seems best, and what options may be missing?”

4. Weather: I want to know the temperature range I will face in many different locales.

Example: “I’m hiking from Irún to Santiago de Compostela in May-June. What temperatures will I experience high and low in each region?”

Et cetera!

I hope this is helpful to someone.

Heavens above!
How did we manage to walk so many caminos with so little planning and without AI for the past eighteen years.

 

[Stephan the Painter]

I imagine the responses may indicate an unfamiliarity with AI, or perhaps a generational gap. I work in tech so I’m happy to answer any questions!

For example, generative AI is already heavily embedded in your life: cellphones, email, Google Maps, Siri, Google search, Gmail, google docs, spreadsheets, ordering food, your local bank, advertisements, newspapers, magazines, language translation, accessibility for blind, seeing and hearing impaired — these all use generative and general AI to varying degrees. And it’s extremely likely that the words you write here are scraped by AI as well.

The idea that you can you can put your head in the sand about it, or ignore it and it will go away, or that people who use it are backwards — just isn’t realistic. It is a part of our world and will become a bigger part over time, with speed.

I’m happy to answer any questions or help with queries. However if this becomes a debate I recommend the mods shut the thread down as that wasn’t the point.

Its Interesting, I just used Google Bard this morning to research an article I was writing. It was about an art show that I had actually seen in Rome recently.

I just asked a general question which was , “tell me about the Sorolla show at the Spanish Academy in Rome in 2023.”

It got it about half right, but mixed in information from a show about the same artist that was at a different venue a few years before. If I hadn’t actually seen the show, I wouldn’t have known that.

I think that’s about where AI research is now. Not trustworthy enough. And if it’s not trustworthy, you have to do your own research anyway.

It did supply the links where it had gotten the information, and I was able to use those to finish my research. And I think that might be its best value now, as a better search engine when doing something like planning a Camino. We’ll see where it goes…..

 

[Former member 108819]

Heavens above!
How did we manage to walk so many caminos with so little planning and without AI for the past eighteen years.

You probably used a combination of different technologies before AI existed in its current form, such as airplanes, taxis, guidebooks, printed maps, Internet forums, cellphones, GPS. (I am assuming you didn’t swim across the Atlantic.)

By the way, air traffic control uses AI today.

I realize this isn’t for everyone. The point of the post is to educate people who want to try it!

 

[Former member 108819]

Its Interesting, I just used Google Bard this morning to research an article I was writing. It was about an art show that I had actually seen in Rome recently.

I just asked a general question which was , “tell me about the Sorolla show at the Spanish Academy in Rome in 2023.”

It got it about half right, but mixed in information from a show about the same artist that was at a different venue a few years before. If I hadn’t actually seen the show, I wouldn’t have known that.

I think that’s about where AI research is now. Not trustworthy enough. And if it’s not trustworthy, you have to do your own research anyway.

It did supply the links where it had gotten the information, and I was able to use those to finish my research. And I think that might be its best value now, as a better search engine when doing something like planning a Camino. We’ll see where it goes…..

You’re very right. In my opinion “artificial intelligence” is itself a bit of a misnomer. It is not actually intelligent in the sense that you and I are intelligent, because the technology cannot discern right from wrong, good from bad, bad from worse. It lacks ability to judge correctness, so it requires a lot of guardrails. It will always emphasize sounding correct over being correct. This can be tricky for people to learn to deal with.

It will likely never have that capacity to know when it is correct or not. That’s why I think it’s important to keep the queries broad, double-check the response for correctness, and use it as one tool in your tool bag. I would not put anything important on the result from a query.

 

[Tincatinker]

I’m happy to answer any questions or help with queries.

OK, here goes. Even the current advocates of generative AI will concede that it that it has no-where near the reasoning capacity of the human mind. Indeed they've even offered that in defence against the "the Bots are taking over" scare stories in the media.

So, my question. What is the point in using generative AI to aggregate and reflect publicly available information to provide me with an opinion on which I cannot rely?

You probably used a combination of different technologies before AI existed in its current form, such as airplanes, taxis, guidebooks, printed maps, Internet forums, cellphones, GPS. (I am assuming you didn’t swim across the Atlantic.)

By the way, air traffic control uses AI today.

I realize this isn’t for everyone. The point of the post is to educate people who want to try it!

Are you suggesting that generative AI will transport a pilgrim across the Atlantic?

 

[Walkerooni]

I prefer to engage in a non-artificial life as much as humanly possible.

 

[Former member 108819]

OK, here goes. Even the current advocates of generative AI will concede that it that it has no-where near the reasoning capacity of the human mind. Indeed they've even offered that in defence against the "the Bots are taking over" scare stories in the media.

So, my question. What is the point in using generative AI to aggregate and reflect publicly available information to provide me with an opinion on which I cannot rely?

Are you suggesting that generative AI will transport a pilgrim across the Atlantic?

Yes, I think most everyone agrees that it has zero reasoning. It is an aggregator of human language, not a large mind capable of reason. You should not view AI as being capable of reason.

Just like we would not say google search has reason.

It’s basically a more refined google search. When you search for something on google, you don’t rely on every result being god’s word. You have to use your own judgement on what you click on, use context clues, use your cultural knowledge, etc. you might look at the source for the result — NY times might be more acceptable to you than a Tumblr post, as a random example. You might think a page with ads all over is less truthworthy than a text only page. Etc. google searching requires your judgment even if google tries to make the results most useful.

To your question, people use it differently, but I use it as a starting point. If you think of it like a google search, in this way it collates a google search result into a conversational response. From your initial search you may look deeper into other sources, or determine one way isn’t the likely best way.

At heart, AI is just reflecting back what humanity has already said about a topic. It doesn’t know whether humanity is correct or reasonable on any topic. A response can be helpful for broad queries where there is a more common truth. “What is the weather like in Irún?” is likely to return a response that is more helpful than something more nuanced like “What was the weather like in Irún on December 23, 1992?” You couldn’t trust the result on the latter as much as the former.

Over time guard railing will become better — it may even include attributed sources, a rating scale of verification, or other resources to help people determine whether the query and result was time worthy.

 

[Stephan the Painter]

It will likely never have that capacity to know when it is correct or not. That’s why I think it’s important to keep the queries broad, double-check the response for correctness, and use it as one tool in your tool bag. I would not put anything important on the result from a query.

This disclaimer was perhaps what was missing from your original post. It might be the most important thing to know about using AI. And why you got a little flack from people.

It probably would be more helpful to think about it like an occasionally poorly researched encyclopedia you can talk to, with a great bibliography.

And of the questions you used as example, Did you get useful information? Did it give you any useful suggestions for your flight. Was it accurate with the weights?

When I used the open AI interface (which I think is chat GT), it told me it couldn’t answer any questions about things that happened after September 2021 because that’s when it’s training stopped,ha!

 

[C clearly]

ChatGPT may well be a useful planning tool, in the absence of a forum of international camino experts such as we have here. The forum is a primary source of information. As you have pointed out, ChatGPT would likely take all of our posts and regurgitate them, perhaps with somewhat less human intelligence.

Yes AI is wonderful, for what it does, as long as its limitations are understood. But once someone finds the forum, I think ChatGPT can be put aside, for Camino info and advice.

 

[Former member 108819]

This disclaimer was perhaps what was missing from your original post. It might be the most important thing to know about using AI. And why you got a little flack from people.

It probably would be more helpful to think about it like an occasionally poorly researched encyclopedia you can talk to, with a great bibliography.

And of the questions you used as example, Did you get useful information? Did it give you any useful suggestions for your flight. Was it accurate with the weights?

When I used the open AI interface (which I think is chat GT), it told me it couldn’t answer any questions about things that happened after September 2021 be that’s when it’s training stopped.

I agree with you. However, here are the first 2 paragraphs of my post:

I’ll preface this with a disclaimer that I share a lot of common concerns about generative artificial intelligence, including its inability to attribute sources or discern correctness, and its potential to disrupt human workforce. Use it at your own discretion.

If you don’t know what generative AI is, in a nutshell, it takes prompts or questions from you and then estimates answers to your query based on language probability accrued from internet sources. It doesn’t have the ability to determine if what it is saying is correct, so often you need to double check it, but it can be a decent weight on decision making.

I think that emotional responses to AI discussions are common. I don’t hold it against anyone. I have had my own serious concerns, including where I work. I have had deep qualms about how it could affect human workers in particular.

For my queries, I found the information useful. It helped me determine the option I should take for my flight disruption. It helped me determine the weather I might see during my trip. It judged the estimated weight really well. I used this as one weight, in addition to this very forum, my own friends and family, and general google searching

However it was not good at stage planning. It couldn’t understand that I wanted to know the km per day to reach the end in 30 days. It kept giving me confused results suggesting I only walk to Bilbao lol. So you really have to use your own good judgement with this tool.

 

[Annette london]

You probably used a combination of different technologies before AI existed in its current form, such as airplanes, taxis, guidebooks, printed maps, Internet forums, cellphones, GPS. (I am assuming you didn’t swim across the Atlantic.)

By the way, air traffic control uses AI today.

I realize this isn’t for everyone. The point of the post is to educate people who want to try it!

No need to assume anything Intricate
i might have paddled for all anyone knows!

By the way you said that you are starting the Norte soon ……would that be May 2022 or May 2033 as noted on your avatar?
Just assuming that AI got the date wrong here!

anyway enjoy your Camino Intricate
please please let us know how the AI worked for you regarding the planning etc etc
With best wishes
Annette

 

[Former member 108819]

No need to assume anything Intricate
i might have paddled for all anyone knows!

By the way you said that you are starting the Norte soon ……would that be May 2022 or May 2033 as noted on your avatar?
Just assuming that AI got the date wrong here!

anyway enjoy your Camino Intricate
please please let us know how the AI worked for you regarding the planning etc etc
With best wishes
Annette

Thank you for the good wishes!

I am leaving Sunday, so I imagine a human moderator put that 2033 in error!

 

[Annette london]

Thank you for the good wishes!

I am leaving Sunday, so I imagine a human moderator put that 2033 in error!

Ha ha, as far as I know we humans input our own dates, past and future Caminos
Well at least I did anyway

Looking forward to your and AI live posts from the Norte, and for heavens sake, if you do happen to get a blister, please don’t ask AI.
You will get a hundred different answers …..just like the humans on this forum give, bless them
All the best to you
And AI

 

[Goyo]

I imagine the responses may indicate an unfamiliarity with AI, or perhaps a generational gap. I work in tech so I’m happy to answer any questions!

For example, generative AI is already heavily embedded in your life: cellphones, email, Google Maps, Siri, Google search, Gmail, google docs, spreadsheets, ordering food, your local bank, advertisements, newspapers, magazines, language translation, accessibility for blind, seeing and hearing impaired — these all use generative and general AI to varying degrees. And it’s extremely likely that the words you write here are scraped by AI as well.

The idea that you can you can put your head in the sand about it, or ignore it and it will go away, or that people who use it are backwards — just isn’t realistic. It is a part of our world and will become a bigger part over time, with speed.

I’m happy to answer any questions or help with queries. However if this becomes a debate I recommend the mods shut the thread down as that wasn’t the point.

I just prefer using books and maps that I can flip thru and hold, and sure, the internet to an extant. Oh yeah, also people like those on this forum add greatly to my knowledge about my Camino, thank you very much, intricate. Generational gap? Meh…Unfamiliarity? My head isn’t buried in the sand… Disinterest as it applies to mi Camino?

¡Que tenga muy buen Camino!

 

[Molly Cassidy]

I imagine the responses may indicate an unfamiliarity with AI, or perhaps a generational gap. I work in tech so I’m happy to answer any questions!

For example, generative AI is already heavily embedded in your life: cellphones, email, Google Maps, Siri, Google search, Gmail, google docs, spreadsheets, ordering food, your local bank, advertisements, newspapers, magazines, language translation, accessibility for blind, seeing and hearing impaired — these all use generative and general AI to varying degrees. And it’s extremely likely that the words you write here are scraped by AI as well.

The idea that you can you can put your head in the sand about it, or ignore it and it will go away, or that people who use it are backwards — just isn’t realistic. It is a part of our world and will become a bigger part over time, with speed.

I’m happy to answer any questions or help with queries. However if this becomes a debate I recommend the mods shut the thread down as that wasn’t the point.

I don't understand what you gain by asking AI to guess something you could just measure yourself. If you want to know how much stuff weighs, you can weigh it. If you want to calculate how far you'll walk, it's easy to look up the distance and divide by the number of days you intend to walk. Actual planning, of course, involves taking note of where the available accommodation is.

Not sure what you mean by the words here being scraped by AI. Do you mean the spell check changing words to make them wrong?

Unless you believe any old rubbish you read on the internet, it's simpler to do your own research.

 

[VNwalking]

Aggregators can't reason.
So good luck getting a coherent list of stages. Anyway, planning is part of the joy of a camino.
Assuming you have a perfectly good mind, that is. And it does seem so.
Buen camino @intricate! Here and now in real time.

 

[Former member 108819]

ChatGPT may well be a useful planning tool, in the absence of a forum of international camino experts such as we have here. The forum is a primary source of information. As you have pointed out, ChatGPT would likely take all of our posts and regurgitate them, perhaps with somewhat less human intelligence.

Yes AI is wonderful, for what it does, as long as its limitations are understood. But once someone finds the forum, I think ChatGPT can be put aside, for Camino info and advice.

Hi mods! Could you considering closing and locking this thread? Or if you could tell me how to?

At this point the discussion isn’t helpful, and it seems to have touched a nerve to cause tangents which wasn’t intended.

Thank you!

 

[Former member 92300]

Well - I personally have no issue with AI - but just don't see the need on the Camino in the AI form.

Weight - why not just get a scale and weigh your gear and use pencil/paper or an app to add it up? I use the website "lighterpack" to figure out my weights and determine what I want to carry and leave behind.

Backpack - You could just throw your gear in a trashbag and take it with you to buy a bag and see if it fits. Or - you can eyeball your gear and look for a pack you think it will fit. Or do what I do - buy a bag and pack it to ensure the gear fits - a smaller bag then I pack less gear.

Distance and days - a simple calculator will tell me that.

Alternative travel arrangements - there are apps for that.

Temperature - there is an app for that.

And except for the scale and the backpack - you can access all of these websites/apps on your phone and have accurate information in real time.

So - do I feel the need to have some AI program to do this for me? Not really. And personally - I would rather see the answers to my questions in writing instead of AI speaking it to me. Of course, if someone is vision impaired or simply does better with verbal information instead of written information - well - I suppose it might be more useful to them.

 

[trevorcc]

My AI tells me walk out he Albergue door, look for a coffee AH that was nice then smile at those around you and put one foot infront of the other and ENJOY the experience, so may over plan over worry and forget why they are walking to Santiago....

 

[dougfitz]

At this point the discussion isn’t helpful, and it seems to have touched a nerve to cause tangents which wasn’t intended.

This is the nature of human discussion, and certainly not, in my view, to close the thread.

If you think that the issues are tangential, you are discounting the concerns that others have. In your OP, two of the three examples were simple arithmetic - trite in my view - suggesting either that you could or would not research to find basic facts. The third question is more interesting, and presuming that you have given us the facts completely and correctly, would have been difficult for even a human actor to determine. You asked a 'what is best question' without providing your own evaluation criteria. Is the cheapest best? or the shortest flight time? or shortest overall time? You don't say. How do you expect an answer when you don't even know how you would evaluate what 'best' might be.

The last question would appear to be one of aggregation, and would appear to reveal the very worst of using an aggregation engine to search the internet for responses. The first problem is to define something that you might feel is obvious, and that is the range of the temperatures you want to consider. Should the extreme results for a particular region be considered, or only results that exclude anything that will occur less frequently than, say, 5% of the time, or 1%. The second is that you have asked for a forecast inasmuch as you wanted to know what you will experience, rather than asking for what you might expect to experience, something that can be relatively easily determined. The third issue is that in the particular case of weather information, there is generally only one authoritative source for the information you are seeking. That is the nature of the international agreements in place. In this case, scanning the internet with an AI engine will add no useful value to knowing the source to use and being able to consult it directly.

Personally, you appear to have provided no substantial case for the use of an AI engine like ChatGPT. If anything, you have illustrated the frailty of the case for humans who do wish to use it. I wouldn't advocate that everyone develop level research skills, but clearly something more than you have demonstrated would seem to be required if one is to make good use of these emerging AI tools.

 

[Former member 108819]

This is the nature of human discussion, and certainly not, in my view, to close the thread.

If you think that the issues are tangential, you are discounting the concerns that others have. In your OP, two of the three examples were simple arithmetic - trite in my view - suggesting either that you could or would not research to find basic facts. The third question is more interesting, and presuming that you have given us the facts completely and correctly, would have been difficult for even a human actor to determine. You asked a 'what is best question' without providing your own evaluation criteria. Is the cheapest best? or the shortest flight time? or shortest overall time? You don't say. How do you expect an answer when you don't even know how you would evaluate what 'best' might be.

The last question would appear to be one of aggregation, and would appear to reveal the very worst of using an aggregation engine to search the internet for responses. The first problem is to define something that you might feel is obvious, and that is the range of the temperatures you want to consider. Should the extreme results for a particular region be considered, or only results that exclude anything that will occur less frequently than, say, 5% of the time, or 1%. The second is that you have asked for a forecast inasmuch as you wanted to know what you will experience, rather than asking for what you might expect to experience, something that can be relatively easily determined. The third issue is that in the particular case of weather information, there is generally only one authoritative source for the information you are seeking. That is the nature of the international agreements in place. In this case, scanning the internet with an AI engine will add no useful value to knowing the source to use and being able to consult it directly.

Personally, you appear to have provided no substantial case for the use of an AI engine like ChatGPT. If anything, you have illustrated the frailty of the case for humans who do wish to use it. I wouldn't advocate that everyone develop level research skills, but clearly something more than you have demonstrated would seem to be required if one is to make good use of these emerging AI tools.

I asked the thread be closed after people started posting insults. The mod messaged me that they deleted the response and would be monitoring this thread for further disruption. Like yours. You’re calling someone trite and frail for using AI. Are you okay?

I do understand emotional responses to AI discussion. But I don’t appreciate the random insults. At this point I am done with this forum. The replies here are reminding me why I was warned on Reddit that this place is a full of cranky old people. I regret coming here, and I truly hope I don’t run into any of you on the Camino.

Now back to my regularly scheduled life. Adios

 

[trecile]

Thread closed at OP's request.