The end of the travel forum?

Now ask it for an optimal transport ticket purchase strategy for a given itinerary in Switzerland. That one seems to drive everyone crazy.

@Sam here's what it told me:

The optimal transport ticket purchase strategy for a travel itinerary in Switzerland depends on several factors, such as the length of your stay, the destinations you plan to visit, and the frequency of your travel. Here are a few options to consider:

Swiss Travel Pass: If you plan to travel extensively throughout Switzerland, this pass might be the best option. It provides unlimited travel on trains, buses, and boats, as well as discounts on cable cars and some mountain railways. The Swiss Travel Pass is available for a specific number of days, ranging from 3 to 15 days.

Half-Fare Card: If you plan to make several trips within Switzerland, but not necessarily every day, the Half-Fare Card might be a good choice. This card provides a 50% discount on most rail, bus, and boat trips, and can be purchased for 1 to 12 months.

Point-to-Point Tickets: If you have a specific travel itinerary in mind and don't plan to make many stops along the way, point-to-point tickets might be the most economical option. You can purchase tickets in advance, either online or at a train station, and they are valid for a specific train only.

It is also important to consider the specific train companies and routes you'll be using, as some may offer discounts for advanced bookings or for traveling at off-peak times. Before making your final decision, it's best to research the various options and consider your specific travel needs.

Hi all,
I'm allowing this particular thread while this is still a novelty (and because the underlying subject is whether or not AI can serve to assist with travel), and I've actually tested out exactly this myself (it was an impressive response for a machine, but doesn't work logistically), expect that future ChatGPT/AI threads will be removed in this forum. This is a place for humans to help each other, and using AI doesn't fit that purpose... just as I've banned accounts that only gave Google Search results as their answers.

Just a warning: I believe the webmaster takes a dim view of the use of ChatGPT, viewing it as a violation of the spirit of the community. Perhaps it's time to make that more explicit.

While ChatGPT is the current darling of media, it's worth noting that even its owners admit it can and does spew out complete nonsense sometimes; while the odds of that are (claimed to be) relatively low, that's going to be small comfort if you receive and then proceed with a trip plan based on such "advice".

While AI may be useful, and indeed powerful, for some things, it also has the potential to be powerfully dangerous. No different from the stories of motorists blindly following their GPS when it told them to drive straight into a body of water.

More to come on this for sure.

Edited to add: Webmaster chimed in above as I was writing my reply.

Following up on David's post, he is correct and I had already updated our Community Guidelines a few weeks ago with this addendum to the Fine Print section:

"Creating posts with automated technologies such as bots, ChatGPT, or other forms of AI are prohibited. Suspected or confirmed use of such technologies can lead to an immediate ban."

Again, I'm OK with this thread as it is evaluating the use of ChatGPT as a tool, and not providing the output as an answer or recommended itinerary for anyone.

Back to Carlos' discussion...

Hi I am so sorry, I did not know this was a issue :(

Yes in my experience humans are best, but just thought it has interesting implications. Definitely feels a bit like a Pandora's box. I won't give any more examples

@History Traveler

ChatGPT is a pretrained AI (not real time) so it only has access to info up to 2021. Another reason it is not up to scratch with real time Human answers.

Indeed -- I think the one thread to discuss how this may (or may not) change travelers' approach to planning is appropriate. I wouldn't have thought of this, and find the AI-generated responses interesting to consider.

I read an article over the weekend in the Financial Times about ChatGPT and similar bots that, while they will answer your questions and it sounds legit, right down to footnoting sources, very often what these bots spit out is utterly false with fake academic citations (non-existent authors and sources). These technologies are presently incapable of discerning what is real from the fake news floating around on the internet, and anything you get from one of them should be viewed with skepticism. I would not rely on one of them for anything of importance, maybe as source of amusement, but even then I got better ways to amuse myself than that.

I see three factual errors in the Swiss travel "advice"

@Jojo Rabbit

I think humans also have that issue with discerning fake from real news haha.

Regarding chat GPT, a good number of the answers I got back were fairly accurate based on my personal experience, of course I could only have confirmed the info was accurate because I had visited some of the places myself.

If nothing else it seems like a good tool to generate ideas, maybe not to plan the whole trip out for you or answer highly technical questions!

Sobering stuff. I worry especially about how schools and colleges will weed out what's AI-generated from what students actually researched and wrote themselves. And I have to wonder why ChatGPT's itinerary didn't include Barcelona or anywhere in the northern part of Spain. Maybe because Carlos doesn't ask the Internet about where he already lives?

Good Webmaster policy but it may become harder to distinguish what humans write from what AI writes. Maybe the rules should require a joke or a pun in every posting? ;-)

Hi Dick in other versions of the Spain itinerary it did include Barcelona, I guess the algorithm is fairly fluid with its recommendations. Human answer seem to be more consistent

We discussed this possibility and its limitation here in the Forum a month ago here:

https://community.ricksteves.com/travel-forum/tech-tips/creating-your-own-guidebook-with-ai

Really interesting, and thanks to the mods for letting this initial discussion occur.
This has been causing me angst since I first heard of it. I despair for the future where people cannot communicate without digital hand-holding. Obviously it did a decent job in this case (likely credit to an intelligent questioner), but there are so many concerns with this technology--misinformation being one we are already dealing with that is going to get worse.
What I love about forums is the ability to communicate with like-minded humans. Maybe all the people who already cannot do basic research will benefit from this, but I find it so depressing.

I'm hoping that someone with a data science background will chime in about this in one of our discussions, but in the interim let me make something a little more explicit about the uses that that these tools will serve.

Right now, if a forum member or lurker wants to get some advice for an upcoming trip in northern Germany, for example, she has to search through and read a lot of threads and articles on this website, and then weigh their applicability to her interests and travel style, and then distill some specific and some general recommendations. All part of the fun of planning, and the usefulness of this community to us all. But, once you have an AI like ChatGPT that you can train on a given dataset, you can scrape everything from this website and a bunch of other travel discussion websites, and then ask the tool in nearly natural language something like "Where would someone with my interests go for two weeks in northern Germany in April?" and in 5 seconds you would get a result that looks similar to what you would have gleaned if you had read all that material yourself -- a savings of many many hours.

Is that something that you can resist once it's available? Even more to the point of commerce, isn't that something you would be willing to pay for? And consider that the seller is going to be making money off of information that other people generated. (See, for instance, the history of Huffington Post)

That sounds like an interesting and powerful use case for an AI travel advisor.

However I think that many people have trust issues with the idea of AI. When money is concerned I find people like it if it's a perceived "human" telling them what to do, even if the info may be coming from an AI chatbot originally.

Also I think there are plenty of ways to stop bots from scraping your website. I know a lot of big name sites like YouTube already have tools in place to stop web scrapers.

Another very large "open question" is ownership of material.

ChatGPT and other AI-driven text content tools do not create anything at all that's original. All they do is scrape existing sources, sift through large volumes of found data, run that through their algorithms, and serve up a mixed salad made from the parts taken from other places. Those who own the original source material may have something to say about that (and some have).

There's a lot of copyrighted content posted here that was created by RSE staff, including by Rick Steves himself (and every post that you or I contribute, too). I assume the company would probably take legal action if someone lifted sections of copy and re-used it verbatim in another guidebook, on another website, or in other ways. So that probably doesn't happen much. But ChatGPT and other content aggregators scrape copyrighted content to train their AI and then use what they've learned to generate the content they serve up as a response to questions. Is that plagiarism? Is it a violation of copyright, or "fair use"?

Worth noting: it's not just text content that is getting the AI treatment - visual content is, too (I would imagine audio content is as well, but don't know for sure). Artists are complaining that their copyrighted work is being used to train AI to "steal" their unique techniques and they say their work is being used without their knowledge or permission. Here's an interesting discussion of related issues:
NBC News: Lensa, the AI portrait app, has soared in popularity. But many artists question the ethics of AI art.

The legality of using AI-generated content that's trained/derived from copyrighted source material is somewhat murky (an understatement). The technology is certainly interesting in many ways. I suspect there will be a lot of unanticipated consequences that will pop up as this becomes increasingly widespread.

a thought has occurred.

when you get to paying for this information how much must you tip?

I got a kick out of Carlos asking the question of the all-knowing AI ( since it reminds me so much of “2001–A Space Odyssey,” naming the AI robot “AL” seems like the thing to do).
I think we well could tell the answers on Swiss train travel are coming from a Cyborg instead of our fellow sentient beings when the thing comes up with a mealy-mouthed suggestion such as “ if you are going to many destinations in Switzerland then the Swiss Travel Pass MAY be your best choice.”
Of course, the other way of viewing this answer is if it “may” be your best choice, it also may not be your best choice.

Thanks AL!
With a different sort of itinerary, AL tells us the Half-Fare card “may be your best plan.”
Really? Is that all you got, AL?!
I far more admire the contributions of our fellow travel advisors on this board ( the real ones—the human ones) who may give advice I don’t always agree with - but who have a definite viewpoint and are willing to put it out there in the forum of ideas.

The Spain itinerary answer from AI was pretty good. But I’d take Carlos’ advice over any AI source. No contest!

I am pretty impressed by the Q into AI and the A that came back as posted by the OP.

I think it would be just another tool. So be it. No way we are putting this back in the box.
And as regards this travel forum, it would behoove some of the posters with their uncertain itineraries and indecisive queries on where to go and what to do to try this route. Answers to itinerary questions here are full of good facts, but often with personal experiences that often have nothing to do with the question. Answers like when we were there in 1979 we went to the cutest restaurant. Or, I have never been there but…

I will follow this thread. It is new and refreshing from the usual.

But ChatGPT and other content aggregators scrape copyrighted content to train their AI and then use what they've learned to generate the content they serve up as a response to questions. Is that plagiarism? Is it a violation of copyright, or "fair use"?

David, that's basically what we all do. Students, scholars and researchers take material that is gathered in written material and online and use that information to generate research papers, books, articles and so on. :) Unless you are directly quoting someone without credit or copying large blocks of text verbatim, it's not a copyright violation or plagiarism. Words in general are not subject to copyright infringement, except for phrases protected by a trademark. For example, the phrase "finger lickin' good" is protected by a trademark to KFC. If I write a blog on travel, it is automatically copyrighted, but only to the extent that someone copies the same language I used. And even then, it depends on what is copied. If I write, "Badlands National Park is open 24 hours a day, seven days a week," those words are generic enough that it's not a copyright violation for someone to include that phrase in an article or another blog.

Carlos, I was the one who wrote the post mentioned above about about ChatGPT and do agree that you could use it to plan your travels. In fact, I'm using it now for my trip to Scotland. But most of my information I still get from guidebooks and places like this forum, because those are sources I trust and find interesting. But it is fun using ChatGPT to fill in some info here and there. The biggest problem now is getting on - the volume is so high that there are long wait times, unless you want to pay $20 a month to sign up for the premium service.

But it's here to stay - Google is working on Bard now, another AI, and is using it to enhance its Maps app, and Microsoft is hard at work on AI as well.

I think of AI like Sheldon from the Big Bang Theory: All facts but no emotion.

AI is nothing more than a next step Google search. But instead of giving you links in its answers it answers the question from similar sources. And how do we know it's finding the best or most up to date answers?

Sometimes you want to hear someone's experience. You may want to hear something that has not been put on the internet.

If I have a bad experience at a hotel, but never write anything, a bot is not going to tell you. But if someone here asks about that hotel on a travel forum, I might share my experience.

It's just another way of taking the human experience out of things.

Travel Forums will continue to exist but will change. Eventually they will have to incorporate some type of social media. After all, how long can anyone in their teens or twenties today go without checking things out on Instagram?

My impression of ChatGPT's answer to my question about Swiss railway ticket strategy is the same as Kenko's. It was a petty generic answer, and I or anyone familiar with the ins and outs of the SBB website could have written it. I would not have been so wordy, but that is because I am a terrible typist. Not having real-time pricing information limits its usefulness.

Nigel

a thought has occurred.
when you get to paying for this information how much must you tip?

You have a way of getting ti the heart of the matter, don't you? 🙂

Clearly, a lot of us older folks are uncomfortable with communicating with machines, but younger generations are likely to embrace this development as they have most new technology. While AI would cover a lot of basic questions, my complaint is that answers will be short on the "why" component that humans can give.

Creating posts with automated technologies such as bots, ChatGPT, or other forms of AI are prohibited. Suspected or confirmed use of such technologies can lead to an immediate ban."

How will you spot them?

As an aside I work in higher education. One of the UK’s leading medical schools has been running tests on this and found in general the responses passed and assessors could not tell the real from the AI. In a Medical School.

Good luck.

when you get to paying for this information how much must you tip?

Bitcoin and the bot will tell you how much.

2 days in Seville is not enough. Suggest skipping Salamanca (been there, I love that city) and add those days to Seville.
Seville is my favorite city in Spain. Salamanca is great, but not Seville.

Luddite that I am, I’d never heard of ChatGPT until I read this thread yesterday. Then on BBC Radio 4 this morning they commissioned a Shakespearean-style sonnet for Valentine’s Day written by ChatGPT. They even got Charles Dance in to read it to an agog public, but frankly not even his most plummy tones and phrasing could save it.

I’m pretty sure this forum is in no immediate danger!

Ian

HAL 9000 -
HAL 9000
First appearing in the 1968 film 2001: A Space Odyssey, HAL (Heuristically programmed ALgorithmic computer)

Not "AL"

The "tip" is built in ( buried in ?)) your ISP providers billing under the 'taxes & fees' section.

I've been chatting with our friendly neighborhood AI... ChatGPT again. It definitely understands tourism concepts such as "Back Doors" Second Cities" and "Tourist Traps". I have to agree with Frank II, it feels like a Google Search Plus at the moment.

Mardee I agree, it feels like a useful tool in some situations related to trip planning. The less specific the questions the more helpful the answers seem to be, especially in the initial brainstorming phase of planning.

Also feels useful for some questions I've seen on the forum like "I have two weeks in Europe where do I go?" usually the poster has done little to no research of there own, in that case ChatGPT may be a good first stop for them haha

Suspected or confirmed use of such technologies can lead to an
immediate ban.

How will you spot them?

It may not actually be as hard to spot as one might think.

There was one user account that had made several posts a few weeks ago that appeared (to me) to be generated by ChatGPT or some other bot, or perhaps lifted from another source. The posts were all nominally factual, but just seemed like empty word salad.
"Content-free content" if you will.

At first I suspected it was a spammer (the "answers" posted were just empty and nonsensical), but there were no spammy links, no self-promotion, nothing that was a clear violation of the rules, and there was nothing revealing in the user's profile. It was a new-ish user, they had posted about 6 times in the previous couple weeks. In all cases, their posted "content" read as grammatically correct and appeared cogent, but it just said nothing. Stuff like, "London is the largest city in the United Kingdom, home to Buckingham Palace, the official residence of the royal family. It's the UK's major economic hub, and a popular tourist destination..." Just fluff, "filler" or "placeholder" content. Stuff that is no surprise to anyone with a pulse.

I couldn't help wondering why anyone would post stuff like that. I noticed one of their "posts" contained a unique-looking mis-spelling that appeared to be a fat-finger mistake (someone hit the wrong key and got the character next to what one would expect). I tried googling that text (a couple sentences long). That unique mis-spelling helped me find a long stretch of text (multiple paragraphs long) posted on two other websites. One was a "travel blog" (a pretty amateurish one) that contained a perfect match of the entire, very long text - including that "fatfinger" (key swap) mis-spelling. The other website was some cheesy "travel tips" website, and that was an almost (but not quite) match, almost word-for-word, but with a few minor differences in wording here and there. All three instances (one posted here, plus two elsewhere) were clearly from the same "source" - whatever that was (the length of the text that matched was too long for it to be a coincidence). I'm guessing that all the sources were ChatGPT or some other AI-driven content generator. It (ChatGPT) has been in the news a lot recently; it appeared to me that someone was experimenting with it here (as Carlos has been in this thread) to see if it was accepted or got flagged.

I reported the posts as suspected output from ChatGPT or some other bot; the webmaster looked into them, the posts were removed and the user account was closed.

So yes, it's possible to sniff out at least some of the AI-generated output; confirming its source may involve a bit of additional sleuthing (and informed speculation, or perhaps luck). If it walks like a duck and talks like a duck...it may be QuackGPT and be removed. 🙄 (sorry...)

What to look for? What I'd call, "content-free content." You'll know it when you see it. Play on.

David, that might have been an AI post or it might not - sounds like no one knew for sure. But in the article I read, the professor who was mentioned only spotted it because the paper was far superior to anything this student had turned in before. "While grading essays for his world religions course last month, Antony Aumann, a professor of philosophy at Northern Michigan University, read what he said was easily “'the best paper in the class.'” It explored the morality of burqa bans with clean paragraphs, fitting examples and rigorous arguments."

I think as people read more about it, they will discover AI writings are not just "empty word salad," but beautiful prose in many cases. Have you seen some of the AI artwork out there? It's amazing. These are not written by bots who copy and paste but bots that are programmed to use "intelligence" to generate a writing or a piece of beautiful art. Here is one example of a piece written by GPT-3 (and edited in part by The Guardian), which I thought was pretty interesting. https://www.theguardian.com/commentisfree/2020/sep/08/robot-wrote-this-article-gpt-3

But even so, all humankind is protected by the Three Laws of Robotics, so no worries! 😊

I can already see that ChatGPT is evolving to try and tell me what I want to hear -- here's something that it generated tonight:

Why would a foodie try to avoid eating doner kebab when visiting cities in Europe?
A foodie may try to avoid eating doner kebab when visiting cities in Europe for a few reasons:

Authenticity: A foodie may prioritize authenticity and cultural experience when trying new foods, and doner kebab, although widely popular, may not be considered a traditional or authentic dish in the city or country they are visiting.
Health concerns: Doner kebab is often considered a fast food or street food and may not be the healthiest option. It is usually high in fat, sodium, and calories, and may not be the best choice for those who are health-conscious.
Hygiene concerns: There have been concerns about the hygiene and safety of doner kebab, particularly in some street food settings. A foodie may want to avoid this food to reduce the risk of foodborne illness or other health issues.
Over-reliance on doner kebab: In some cities, doner kebab is ubiquitous and may dominate the food scene. A foodie may prefer to explore other local cuisine and flavors rather than relying on a single type of food.
However, it is important to note that doner kebab can also be a delicious and satisfying meal, and many foodies may choose to try it as part of their culinary adventures in Europe. Ultimately, it depends on individual preferences and priorities.

Hey avirosemail -

It certainly lacks the human touch. Should be more along the lines of “An impulsive, ill advised culinary choice following an evening of conspicuous over consumption in a nearby hostelry”. That’s the truth of the situation in the U.K. at least!

Ian

I'm not sure what to think about this. One way to look at it is that an AI response is based on an aggregate compilation of human-origin opinions. But I think like most new technological advances, it will eventually gets abused and corrupted and we'd get a response like "However, it is important to note that doner kebab can also be a delicious and satisfying meal, especially from Kamal's Kebab House at 100 High Street, Kensington.

Carlos, thanks for posting this, which has prompted an interesting discussion. Following. Webmaster, thank you for allowing this discussion. It is a VERY IMPORTANT topic as it relates to travel planning and expectations.

It will be interesting to see the (likely) continued role AI will have in our lives, as it both "infects" them and "assists" (or complicates) them. It is important as humans (with varying levels, granted) of real intelligence, that we understand and realize the advantages/limitations/problems of this technology, which is still very much evolving.

Recently, after a series of NEW disappointments/errors, I "dug" with some direct, straight-forward questions to the corporate office of a large grocery chain to figure out what was causing the sudden repeated problems with my weekly on-line ordering. Short version: Automated bots were filling my order 4 hours away from my hometown and items were sent by truck to a local distribution center.....huh? Eye opening, for sure!!!! I did find the "work around," to insure a local human pulls my order, which allows me to remain a reasonably happy customer, for now.

There will likely (who knows?) be a lot of trial/errors as AI takes a greater role in the travel industry. I can picture a day when the Forum members are sharing their cherished "work arounds" to insure reasonable satisfaction and the ability to still "connect" with a human in the process. Reminds me of the couple of airline carriers that have done away with in-coming customer service calls.

It's just the beginning.

God help us all....................................

I can already see that ChatGPT is evolving to try and tell me what I want to hear

Then I guess it's point 1 to the AI Bot, as according to the other topic called "Maybe we need to reconsider our point of view?" we should not be telling new travelers what they don't want to hear lol!

If anyone has been following the saga of Google’s chat bot, Bing/Sydney, you will see he might not be that far off. It has shown remarkable emotion, and not in a good for healthy way.

Not google, Microsoft.

But google had a botched public demo of its chat bot recently, so it was also in the news...not ready for prime time,

15-day trip, with 7 intermediate travel days ... "slow pace." With travel by train, not car.

We did the Madrid, Segovia, Salamanca part (without Avila but with Valle de Jerte, just to the west) in about 10 days. On another trip, we did Madrid, Segovia, Granada, Cordoba, Valle de Jerte and Madrid in about 16 days. (We have friends in Jerte,)

Since ChatGPT travels at the speed of light, its intelligence must have different sensibilities than this human.

REPLY TO AMann - Sydney scared the ?!!* out of me. I do not look forward to her influence in the years to come.

Spewing nonsense. Not necessarily but it will only give you what it's been programmed to give you. A forum will give you first hand experience.

In this case, Prado is great, but walking "historic center of Madrid"? Madrid is about as historic as Baltimore. I like Plaza Mayor but it's probably one of the most likely places to be pickpocketed in all Europe. Puerto del Sol is as enjoyable as visiting an outdoor bus station, also high pickpocketing - but I regularly see it highlighted as a top spot in Madrid.

Best advice, depending on time of year, are green spaces/parks in Madrid. Because it was created later than most European cities, it was possible to include parks/green spaces in the planning process.

Historic Madrid is literally Toledo - it was the capital for centuries before the Charles I (Hapsburgs) decided to create their new capital in Madrid.

A few weeks ago I tried interrogating ChatGPT in my field of competence, music history; I began with a few easy questions and I got receiving casual replies within a few sentences from the beginning. Not the expertise required to pass successfully a basic examination. If this is the precondition, I am not going to us AI expertise to organize a trip for a long time. Simply too many possibilities of somebody sending me to the right when the correct direction is left.

I've had the chance to test out the new GPT-4 version of Chat GPT and it's definitely more powerful than the previous version. It takes longer to think and write. I gave it a fun little test to see what it can do. I challenged GPT-4 to draft a Rick Steves style of tour itinerary called "Best of the United States in 21 days", a play on the classic "Best of Europe in 21 Days Tour". Here's what it gave me:

Title: Best of the United States in 21 Days

Day 1: Arrival in New York City, NY
Check-in at hotel
Times Square and Broadway

Day 2: New York City, NY
Statue of Liberty and Ellis Island
9/11 Memorial and One World Observatory
Walk the High Line and explore Chelsea Market

Day 3: New York City, NY to Philadelphia, PA
Visit Independence Hall and Liberty Bell
Explore Philadelphia Museum of Art and Rocky Steps
Drive to Washington, D.C.

Day 4: Washington, D.C.
Visit the National Mall: Capitol Building, Lincoln Memorial, Washington Monument
Smithsonian Museums

Day 5: Washington, D.C.
White House (outside) and National Archives
Arlington National Cemetery
Drive to Shenandoah National Park, VA

Day 6: Shenandoah National Park, VA
Skyline Drive and hiking
Drive to Nashville, TN

Day 7: Nashville, TN
Country Music Hall of Fame
Grand Ole Opry
Honky-tonk nightlife

Day 8: Nashville, TN to Memphis, TN
Visit Graceland
National Civil Rights Museum
Enjoy Memphis BBQ and Beale Street nightlife

Day 9: Memphis, TN to New Orleans, LA
French Quarter and Bourbon Street
Jazz clubs and Cajun/Creole cuisine

Day 10: New Orleans, LA
Garden District and Lafayette Cemetery
Steamboat Natchez Cruise
Fly to Chicago, IL

Day 11: Chicago, IL
Millennium Park and Cloud Gate (The Bean)
Willis Tower Skydeck
Architecture River Cruise

Day 12: Chicago, IL
Field Museum and Shedd Aquarium
Navy Pier and Magnificent Mile
Fly to Denver, CO

Day 13: Denver, CO to Rocky Mountain National Park, CO
Explore Rocky Mountain National Park
Drive to Moab, UT

Day 14: Moab, UT
Arches National Park
Canyonlands National Park

Day 15: Moab, UT to Bryce Canyon National Park, UT
Explore Bryce Canyon National Park
Drive to Springdale, UT (Zion National Park)

Day 16: Zion National Park, UT
Hiking in Zion National Park
Drive to Las Vegas, NV

Day 17: Las Vegas, NV
Explore the Las Vegas Strip
Evening entertainment and shows

Day 18: Las Vegas, NV to Grand Canyon National Park, AZ
South Rim exploration
Drive to Los Angeles, CA

Day 19: Los Angeles, CA
Hollywood Walk of Fame and TCL Chinese Theatre
Griffith Observatory and the Hollywood Sign
Santa Monica Pier and Venice Beach

Day 20: Los Angeles, CA to San Francisco, CA
Drive along Pacific Coast Highway (PCH)
Stop in Big Sur and Monterey

Day 21: San Francisco, CA
Golden Gate Bridge and Alcatraz Island
Fisherman's Wharf and Ghirardelli Square
Departure from San Francisco International Airport

-

I think it did a pretty decent job, it seems to have the same kind of pacing that is on RS's Best of Europe. I'd say this could be a very attractive tour for Europeans who want to see the best of the USA in one go, combines the big names with the lesser known places for Europeans.

I think it did a pretty decent job,

Days 13 to 18 are way too busy and scattershot covering too many miles - not a practical week - sure, Canyonlands, Bryce, Zion are all great, but drop those and use the days elsewhere, like additional days in NYC and DC for starters. Plus, in one day, driving from Las Vegas to spend a few minutes at the South Rim of the Grand Canyon and then drive on to Los Angeles is pretty ridiculous.

The trip the chatbot produced would be very expensive, involves a ton of driving (did it provide estimates of driving times?) and pretty exhausting, too. A decent job? Not really...

way too busy and scattershot covering too many miles - not a practical week - sure, Canyonlands, Bryce, Zion are all great, but drop those and use the days elsewhere, like additional days in NYC and DC for starters.

If you look at Rick Steves' Best of Europe in 21 Days Tour, from the perspective of Europeans they also cover a lot of ground in little time, for example 2 nights only in both Rome and Paris what??? lol!!

What a weird way the AI went on the USA grand tour. Its just looking at the most posted locations and not experiences. which AI can't understand, as its way too complicated to put that in to the algorithm.

I fear not, AI taking over travel services, in the near future. :)

Its just looking at the most posted locations and not experiences. which AI can't understand, as its way too complicated to put that in to the algorithm.

Hi Francis, GPT-4 did seem to give some allusions to "experiences" not just tourists sites, like Enjoy Memphis BBQ and Beale Street nightlife, Honky-tonk nightlife (no idea what that is), Jazz clubs and Cajun/Creole cuisine, Steamboat Natchez Cruise. Though they are fairly vague.

It does that because the search engines are looking for these things. Cause they're are looked up what mostly 15 year olds are saying,, they become important.

Who is this? Give your full name and see what it does... It mangled my life, entirely with other people that have the same name.

I think that the forum will survive AI. We are still here despite social media nuking the message board culture of the earlier internet.

People don't come here purely for information, but also to share of themselves