This is just the first page:

You talked me into uploading my photo. I asked for my celebrity look alike. I got Zoe Saldana. Here are two photos of her. I don't think we look anything alike but I wish we did. I am guessing that I am old enough to be her mother. She also appears to be about the size of a twig. Let's just say I am not...

ETA I just googled her age. I am only 8 years older than her. I was guessing more like 25 years older. Remind me to never ask ChatGPT a question like that again. Depressing!

ONly that you are both very pretty!

Thank you. It was a fun little task. I also asked how old I look. It gave an unrealistic answer. Then I uploaded the same photo again and it said it was too difficult to tell.🙄

Based on Likefriday's experience with lookalikes, I won't be asking ChatGPT.

Interesting experiment, Mth.

Very interesting. I was really interested to read the page you shared, because I did visit Iguassu ( though many years ago). Edited to say now I googled how to enlarge the page and am studying it :) So you are doing both sides. We only did Argentinian side. I will say that optional boat ride was absolutely outstanding so if you can, make it happen. and wet is an understatement.

I'm tempted to ask for my lookalike.... now how do we do this? I know generally when I get any option for AI stuff I say no so maybe I've banished it too far from my puter?

I haven't read all the articles yet but the NYTimes just had several on how AI is changing travel. The topics were: How AI can help plan a trip to NY; help you make a connecting flight; guide you through a vacation in Norway and improve weather forecasts and delays.

Tempting to ask for a lookalike though.

Try this: take your own picture, or someone else's if you don't want to use your own. Ask Chat GPT to reproduce it without changing anything. Take that one and repeat the instructions. Keep taking the result and ask Chat GPT to reproduce it without changes. See what happens.

I belong to a vintage car group on FB and there are posts every day of cars that are misidentified. Maybe they're just off by a year or two. But frequently there are posts of cars that never existed --real cars, but with specific details that are incorrect.

The problem is that if you know a lot about the subject, you recognize the information as partly incorrect, or possibly false in the guise of accurate.

But many people say "Cool, I have never seen that particular model before. "

But there were many South American aand Australian variants of American cars that I am not conversant with, so I have little clue as to their actual existence.

So as a frequent world traveler, you can probably verify that GPT gave you a "really good itinerary " but this is based a lot on your preexisting knowledge. You recognized a "pat answer" for what it is.

So is it a good tool of convenience for someone who already has information? Sure. The problem is that people will use it For complete information and take it all at its face value. We can ofte n tell in teaching who uses Chat GPT to help them compose a better answer to a homework question vs. someone who uses it To Answer the question.

The problem with all of this is that more people will Misuse it than utilize it. And I think it's going to make people more Stupid.

I know people who don't know how to get someplace after driving it ten times because they never pay attention to the actual trip. They just do what the App tells them at the moment. They don't Learn anything from the experience.

DD wanted to take the urchins to NYC for a long weekend. I gave ChatGPT the children’s ages, interests, and things I especially wanted them to see, like Temple of Dendur. I was very pleased with the itinerary it returned. I’m going to try it for Reykjavik, too!

off to get a picture and see who I look like….

ETA: Do y’all use an app? I always access it by googling for it. There is probably a better way.

More ETA:

It would not give me who I look like, per se. How did y’all get it to give you the name of a person(s)? This is what it returned. If only …

Which Movie Character You’d Be

You’d totally be a blend of:

• Professor McGonagall (from Harry Potter) – not for the sternness, but for the wisdom, quiet strength, and clever wit.
• Diane Keaton in Something’s Gotta Give – stylish in an understated way, independent, and deeply thoughtful with a warm, dry sense of humor.

💃 Style Inspiration

You give off a timeless, elegant aesthetic that aligns with:

• Diane Keaton (again!) – especially her love of clean lines, glasses, and neutral tones with character.
• Meryl Streep off-screen – relaxed sophistication with personality and depth.

🖼️ Vibe Summary:
"Intellectual elegance with a creative soul. Someone who would host a book club, give heartfelt advice, and recommend the best indie films."

I’m going to try it for Reykjavik, too!

Yes, I thought of you and that trip specifically. I really appreciated the legwork on transfer times, opening and closings. No fun in that.

Pal, keep in mind I am using the free version of what is still pretty early AI in the scheme of things. I notice it has a disclaimer that says it can be wrong. It definitely can be wrong! It is a tool that has to be used judiciously.

One interesting aspect was that it asked me if I wanted a pdf. I said no, I wanted a Google doc. It said it could not do that but offered me a docx. The docx had odd formatting ...not wrong so much as odd. I told ChatGPT and it fixed the formatting in a new version. I find it interesting that it understood what I meant by"odd" and was able to fix it, but thought it was ok in the first draft.

@salonva, would love to hear your thoughts on IF; where else did you visit in SA?

Just go to ChatGPT.com. It is free. It might prompt you to do this (that was the way I found it). If not jsut say, can you find my celebrity look alike? It will prompt you to d/l a photo

HAHAHAHAHA

I wish!

Tried again with a better picture.

So it must be true. Julia Roberts is my twin.

I think that most people who use it as a shortcut rather than an assist will use whatever the free version is.

I tried ChatGPT very briefly, was underwhelmed by the experience and deleted the app. If in the end if you still have to verify all of the information, what’s the point? At some point people will have to disclose that AI assistance created some or all of the information presented in their lecture, or their brochure, or in their book, or post operative instructions or whatever…..and the sooner that is required the better, IMO.

I don't use it as an app, i just go to the website. Depending on the use and your own knowledge, I don't feel I need to verify all the info. given. No source of info. is 100% infallible. I think the error rate is pretty low, but people sometimes have a sense of false precision with AI.

I tried it out for an itenerary for a trip to Italy and it at least helped me out with a starting point. I suck at planning trips so I'll take any help I can get even from Chatgpt

I like planning trips, as I am sure you all can tell, but there are still parts of it I'd just as soon delegate.

Ok I tried asking for my celebrity lookalike and it said it couldn't do that but I had a good smile.....I guess I don't know how to do this and will try and spend some time figuring out what everyone else seems to get easily.......

Mtn, I checked and our trip was 2007 (!). It was DD2 semester abroad in Buenos Aires so of course we visited her. We spent probably 5 to 7 days in BA (and there was so much to see and do, and eat ) which we loved. I cannot find it but they had trips to the pampas and barbecue- party which we did do and it was not cheesy but so much fun. I just remember that there were not many English speakers but we had a great time. Really, there was so much in Buenos Aires we could have spent more time. Iguazu Falls was absolutely worth all the build up.

We did not get to Mendoza but DD2 did and raved about it.

We did a lot of walking in BA and did a lot of walking tours as well. At that time, 2007, I used a lot of forums on Tripadvisor but most things needed email to communicate ( no online sign ups).

We had planned a ferry trip to Uruguay- Colonia de S but the only day that we had left, was heavy rain so we did not do that.

oh well it finally said-

Your look brings to mind the confident, expressive style often seen in iconic performers or actors—especially those with natural, silver hair and a warm, engaging presence. Think along the lines of someone like Bette Midler or Diane Keaton in their more relaxed, candid moments—not an exact match, but similar in vibe and charisma.

Pretty much women of a certain age. I do not believe I look anything like either one...well I do have wild hair so there's Bette but facially, not at all.

Thank you, Salon. We are only spending 3 days in BA, mainly as a stopover. since we have been. But last time we were basically trying to get in all the famous sites, so this time we can do more offbeat things. It is a great city.

We actually will fly into Montevideo and drive along the coast to Colonia, then get the ferry back to BA. We don't have Mendoza in our plans, but will go to Patagonia/Bariloche. Glad to hear IF was worth it; it is logistically a pain for sure.

I first used it to plan an itinerary for my MI trip last summer and was blown away. One of DD's tech friends said recently, "Nobody should be using Google anymore, just go straight to Chat GPT -and I find myself doing just that more and more. Just yesterday I used it to help me write a post for a rescue dog. And the day before that, I used it for help with my living room, which I'm updating a bit.

Nobody should be using Google anymore, just go straight to Chat GPT

I have avoided it, not entirely sure why, but I agree it is an amazing tool used carefully. It can't do my work writing (yet) but it is a much better proofreader than Word or Gdocs.

I am just starting a test of AI for our work product. Some parts are really great. Meeting notes! Yeah - it will listen to the meeting and produce notes. It will include every bad dad joke or sneeze if you just tell it to take notes. If I ask it to summarize the key points and list takeaways I get a much better result, but it sometimes misses some of the things I think are important. It saves me having to take notes during the meeting, but I have to review before sending out the notes and make sure they are complete and accurate.

Knowing how to refine your ask is half the battle, and looking at the design dilemmas I wonder how many people are actually capable of asking the right questions.

Will AI know to answer the question "What color should I paint my kitchen cabinets?" with "What color are your floors?"

I worry about potential misuse: When will the algorithms start introducing bias? Will they promote certain pieces of information and limit other pieces of information based on what the app owner believes?

I worry about accuracy - I work with data - garbage in / garbage out. How do we keep the machine from learning from idiots?

I worry about the lack of scrutiny of the data it may return. Ask about cost of weddings and it will bring back the most recent survey data from The Knot. The Knot surveys members who have gotten married in the past year. But the members - people who signed up for the services offered by The Knot isn't going to provide the results as a sample of all couples in the US who got married in the past year, including those who may have eloped or had a small wedding at their local church or got married in front of the Justice of the Peace and held a reception at a local VFW.

I wonder if it will limit creativity. Sometimes I see a design dilemma and start pondering possibilities. 10 people will come up with similar ideas and then one of our truly creative Pros comes through with the perfect answer that no one else thought of. Will we end up getting the cookie cutter responses of what is familiar and common and lose the truly innovative and creative answers?

Itinerary question: DH loves to plan trips, too, and when anyone is traveling he likes to contribute his expertise. So when his brother was returning from a trip abroad, flying in to Dorval in Montreal, then taking the train 4 hours to the city closest to his home, DH let him know how to get to the train from the terminal, and where to find food (Tim Horton’s, Subway, or a pub) because there is no food at the station. And how much time to allow. Will AI tell him that? I’ll bet it will!

I'm glad you started this topic. I was struggling tonight to create a calendarized checklist for medication management using iOS tools on an iPad. I remembered this discussion and tried ChatGPT out on the task. The result wasn't perfect, but it was good enough and took me minutes instead of the agonizing slow progress I was making before. I think I will play around with it a little more on some more complex spreadsheets type tasks.

Fun, my friend used it to create an Excel spreadsheet for care homes, medication lists, and more, for her MIL. Still had to tweak them but as you said, it saved her so much time and headache.

DH takes a lot of meds.... I believe it was from Web MD that he has been using a listing. We both carry small printouts in our wallets of his meds. Knock on wood we haven't needed it lately but several years ago we had our runs to the ER etc and it's just handy anyway now that each dr visit reviews what meds you are taking. That way you know you're not omitting something.

I still don't seem to have the swing of it to use Chat GPT. I'd like to try it for a few things that some ofyou have mentioned but I just get very generic responses. I am using the webiste, like Mtn as I am not wanting to do it on my phone.

@salonva, others with more ChatGPT experience will chime in, I hope, but I wanted to share what I’ve found about using it.

If I think the response is too generic, I ask it to amend the response and give it more parameters. For example, yesterday I asked it what to do in [name of town] Iceland that would be interesting for children aged [gave ages] during [gave the dates]. After it answered, I asked it to amend and gave it the number of hours we would be there. At this point ChatGPT figured out it was a cruise stop. Next I asked it to amend putting one of the activities last. One activity required a taxi to go to a nearby town. So I asked for a list of taxis that would go from A to B.

I have found that a fair amount of specificity is needed. When I do a Google search, I start very broad. With ChatGPT, I’m fairly specific.

I forgot to add that the first thing I ever did with ChatGPT was have it concoct a cocktail for me. I told it what was in the liquor cabinet and it came back with a recipe that was actually very good!

btw this description of me Someone who would host a book club, give heartfelt advice, and recommend the best indie films is so totally inaccurate. I would neither be a member of nor host a book club unless my cookbook club counts. Book club to me is too much like required reading. I can’t stand indie films. However, I will tell you what I think you ought to do, but I don’t know that it would be characterized as ”heartfelt.” I’m not very warm and fuzzy (fluffy, yes; fuzzy, no).

ETA: Just used ChatGPT to translate text from Icelandic to English. Couldn’t find my Translate app.

Quick question for those of you using the app - did you sign up/log in or are you using it more privately (yeah, right!) as a 'guest'? Upsides/downsides?

I just asked Chat GPT: If a renter builds onto your home without your permission?

It's reply, short version: Stating lease violation. Document the problem & the conversation you have with the tenant. Ask the tenant to take down the structure. I LIKE that!

My property manager said: Well....they pay rent on time don't they!

ChatGPT is reinforcing what is legal. What I already know to be my legal rights. It's logical. I think I'll have more "discussions" with a logical device in the future!

Thanks Mtn! It's not just good for travel itens.

This article pops up in my feed lately.

An accidental, inaccurate term in AI

Andrew Clements who wrote Frindle must be laughing. Maybe we could get something goofy going! But also, this ”feature” could be abused.

DLM, I signed in with Apple, admittedly with some misgivings. When I deleted the app it purportedly deleted all the information. I only test drove it— with random speculative questions and scenarios. The one benefit I could see is that it could allow access to information from sites that may otherwise require you to sign in or sign up, which would limit your online footprint to some extent.

I signed in with Google. I use it in temporary mode - check dotted circle in upper right corner. Allegedly, nothing I ask is saved.

@DLM2000-GW .I just go to chatgpt.com. I don't pay or sign up.

And yes, the best use of ChatGPT is when you keep honing and refining. Just like anything you do yourself, right?

Use it just the way you'd instruct a human. If you don't like the first results, tell them why. EG "that was too long" or "i'd like a less formal style" or "please add addresses to my itinerary" etc etc

Thanks for the info. I tested a couple of queries yesterday but it kept prompting me to sign up for better feedback - done through the .com, not an app.

I downloaded the app only because I wanted to play along with ”who do I look like.” I thought I might get better results. I didn’t. I usually go through the .com for queries. I’m signed in with Google there, too.

Bbstx, who did Mr Chat say you look like? (I haven’t asked because I prob would not like the answer.)

Interesting, it never pesters me to login or sign up and pay. It will, however, at a certain point, cut you off for 24 hours. I rarely ask it enough for it to do that to me.

I am on a quest for a vintage toy pond boat. My area is crawling with antique stores and I want to check them out. I asked Chatgpt to plan an efficient route for me to go to all antique stores within a 30 minute radius. It did so, warned me that one was "reported closed (permanently)" and that a good one was 35 min away, if i wanted to add it. I then told it to delete 3 or 4 I had been to. It did so and then sent me a Google maps link so I can use it on Car Play.

coda - the closed place? it has a pic of exactly what i want! : (

KSWL, I am sure you'd like the answer! Like I said, they never choose character actors or seniors they always err on the side of the very most attractive celebrities! Also they really go on coloring more than features. I don't think it's their highest and best use, TBH... but it's their fault, they suggested it!

At the library we can use the computers anonymously. i wonder i could access ChatGPT there?

of course, it is just a website

Kswl, it said:
You give off a timeless, elegant aesthetic that aligns with:
Diane Keaton (again!) – especially her love of clean lines, glasses, and neutral tones with character.
Meryl Streep off-screen – relaxed sophistication with personality and depth.

I think they meant Meryl Streep when she wasn’t wearing make-up, but had her glasses on, and probably a sweatsuit.

Y’all who have met me know I look nothing like either of those women! I only wish….

If I thought it would compare me to either of those women I would ask 🙄 I’d also take Natasha Richardson in the blink of an eye. Or Cate Blanchette.

Apparently, ChatGPT limits the number of queries or the number of characters or the number of something (I’m trying to find out) that you can consume in 5 hours, then it cuts you off. I got cut off. I waited a couple of hours. When I attempted a new query, I was allowed to ask.

So far today, I’ve had it plan a kids’ trip to Chicago, answer a grammar question, and give me the name of taxi companies in a small Icelandic town.

Interesting article about Gemini AI chatbot becoming available to children. Google warns families of the risks. This is a ”gift article” so even if you don’t subscribe to NYT, you ought to be able to read it.

https://www.nytimes.com/2025/05/02/technology/google-gemini-ai-chatbot-kids.html?unlocked_article_code=1.Ek8.FJQo.gbyrdmhlrq55&smid=url-share

From the site on my Mac, I dont get the paper clip option to upload a photo. On my phone, when I click on the paperclip it wants me to log in/sign up. I don't think I'm up for that just yet.