BẢN CHÉP ÂM THANH 🔗
WEBVTT
00:00:00.239 --> 00:00:01.360
Well, hello everybody.
00:00:01.439 --> 00:00:09.759
I am Jenny Deeran, the host of the Aged Podcast, the podcast that is all about AI for the 65 Plus crowd.
00:00:10.000 --> 00:00:15.679
I'm here with my sidekick Bitsy, and we are ready to deliver a great episode.
00:00:15.919 --> 00:00:17.600
Episode 53.
00:00:18.480 --> 00:00:23.199
And today on Aged, it's back by popular demand.
00:00:23.679 --> 00:00:29.199
10 things I've done with my AI, Claude, recently.
00:00:29.440 --> 00:00:32.640
Things that I've done recently, straight from my real life.
00:00:32.880 --> 00:00:37.679
We'll get to carrots, Medicare, the US Open.
00:00:38.000 --> 00:00:52.159
I'm going to highlight a few news stories about dogs, drugs to extend life, a remarkable story about how arthritic ankles and a computer came to the rescue and a whole lot more.
00:00:52.479 --> 00:00:54.960
So stay with us.
00:00:55.280 --> 00:00:57.200
I do want to make a couple of notes.
00:00:57.520 --> 00:01:04.400
You probably noticed that I talk about my AI, I talk about Claude, I talk about Bitsy.
00:01:04.879 --> 00:01:09.280
For all intents and purposes, those are all the same things.
00:01:09.599 --> 00:01:22.879
My AI of choice, my primary AI is Claude, which is developed by a company Anthropic, and I happen to informally refer to my AI as Bitsy.
00:01:23.040 --> 00:01:29.519
So you'll hear throughout this episode me refer to my AI, to Claude, to Bitsy.
00:01:29.920 --> 00:01:33.280
Just know that those are all the same thing.
00:01:33.519 --> 00:01:37.680
Anyway, let's get started with AI in the news.
00:01:37.840 --> 00:01:42.400
And my first story today is for all you dog lovers out there.
00:01:42.560 --> 00:01:45.840
And I know there are a lot of you out there.
00:01:46.159 --> 00:01:53.840
The story ran in the New York Times on August 31st in a column called Pet Theory by Emily Anthees.
00:01:53.920 --> 00:01:56.640
And the headline gave me a little chuckle.
00:01:56.879 --> 00:02:02.159
The headline is How much is the gene-edited dog in the window?
00:02:02.799 --> 00:02:04.799
So here's what happened.
00:02:05.040 --> 00:02:13.199
A biotech startup in New York used a gene editing tool called CRISPR, which we've talked about on this show before.
00:02:13.439 --> 00:02:32.240
It's a technology that lets scientists make very precise changes to genes, and in this case, animal genes, to create what these folks in this company are working on is to create the first hypoallergenic dog.
00:02:32.639 --> 00:02:37.360
Now, this isn't through breeding, this is through gene editing.
00:02:37.599 --> 00:02:54.000
So anyway, there are two beagles named Alfie and Bailey, and with a single genetic tweak, these dogs don't seem to produce the protein that leaves so many folks, including my daughter-in-law, red-eyed and wheezing.
00:02:54.400 --> 00:03:11.840
The man who runs the company has had dog allergies all his life, which is what prompted him to get into
Load full transcript
this business, and he's now living with one of those beagles, and says he hasn't had a single reaction since he's had them.
00:03:12.159 --> 00:03:14.800
So where does AI fit into all of this?
00:03:14.960 --> 00:03:23.039
I want to be clear, because this is an AI podcast, after all, the gene editing itself isn't AI.
00:03:23.439 --> 00:03:30.560
CRISPR, with the acronym C R I S P R, came out of biology, not computers.
00:03:30.800 --> 00:03:33.599
But here's the connection, and it's a big one.
00:03:33.919 --> 00:03:51.919
AI is more and more what lets scientists use this tool, a tool like CRISPR, to do more and more, helping them pinpoint exactly where to make the edit and predict what will happen before they ever try it.
00:03:52.159 --> 00:04:01.039
So AI isn't the star of the story, but it's a big part of why tools like this keep getting faster and smarter and bolder.
00:04:01.439 --> 00:04:05.360
Now, this is aged, so you know I'm going to look at both sides.
00:04:05.520 --> 00:04:11.439
On the helpful side, this could be wonderful for people who love dogs but can't be near them because of allergies.
00:04:11.599 --> 00:04:15.759
And the company wants to use it also for service dogs, which would be great.
00:04:16.000 --> 00:04:19.439
On the other side, the article raises a real question.
00:04:19.759 --> 00:04:28.079
We humans have a long, not so pretty history of breeding dogs to suit ourselves, not the dogs.
00:04:28.319 --> 00:04:38.959
Hey Bitsy, can you remind us of some examples of some dog breeding we have done in history that have made humans happy, but not the dogs?
00:04:39.439 --> 00:04:41.360
Oh, that's a juicy one, Ginny.
00:04:41.519 --> 00:04:50.560
Bulldogs are the classic example, bred for that squashed face humans find cute, but it leaves many struggling to breathe properly and prone to overheating.
00:04:50.800 --> 00:04:54.399
Pugs have similar airway problems for the same cosmetic reason.
00:04:54.720 --> 00:05:05.040
German shepherds got bred with that sloped back look that's actually linked to hip and spine problems, but it leaves them prone to serious back and disc issues their whole lives.
00:05:05.839 --> 00:05:08.319
Wow, badger burrows, what is that?
00:05:08.399 --> 00:05:12.079
And what did these dogs do in the badger burrows?
00:05:12.959 --> 00:05:14.720
Huh! Good question.
00:05:15.040 --> 00:05:18.480
Daxons were literally bred to hunt badgers underground.
00:05:18.639 --> 00:05:21.439
The name even means badger dog in German.
00:05:21.680 --> 00:05:26.879
Their short legs and long narrow body let them squeeze into burrows and dig the badgers out.
00:05:27.040 --> 00:05:30.319
Well, that deep chest gave them enough lung power for the fight.
00:05:31.680 --> 00:05:35.279
Wow, that this is a good example of why I love AI.
00:05:35.519 --> 00:05:39.759
You can get onto something and I can just have a conversation where I can learn more and more.
00:05:39.920 --> 00:05:45.759
Anyway, I don't have the answer about this whole dog breeding thing, but it is fascinating to chew on it.
00:05:46.079 --> 00:05:46.800
Get it?
00:05:47.199 --> 00:05:49.519
All right, let's get into AI for good.
00:05:50.000 --> 00:05:56.480
This is a segment of the Aged Podcast, and it brings examples of AI being used to make life better.
00:05:56.639 --> 00:05:58.319
And I've got two for you today.
00:05:58.560 --> 00:06:07.040
And they come at it from completely opposite ends, one from a big research lab and one from a single person sitting at his computer.
00:06:07.680 --> 00:06:17.759
So the New York Times reported just this week that a company called Insilico Medicine has developed a drug candidate called Rentiserapid.
00:06:18.480 --> 00:06:27.439
And what makes it remarkable is that the drug's molecular structure was actually designed with the help of artificial intelligence.
00:06:27.680 --> 00:06:44.480
The company built AI systems that could, in their words, scan a lock and generate a key that fits it, meaning that AI designed a brand new molecule from scratch to target a specific disease.
00:06:44.720 --> 00:06:47.360
This is new and this is a big deal.
00:06:47.680 --> 00:06:54.319
Anyway, it's a serious disease that the medicine is trying to help with.
00:06:54.399 --> 00:07:00.319
It's called idiopathetic pulmonary fibrosis, which is a mouthful.
00:07:00.560 --> 00:07:06.800
It scars the lung tissue and makes it harder and harder and harder to breathe, and it can be fatal within a few years.
00:07:06.959 --> 00:07:13.680
In this drug trial, the AI-designed drug significantly improved patients' lug capacity.
00:07:13.839 --> 00:07:16.959
That alone would be great news, right?
00:07:17.360 --> 00:07:19.600
But the story gets even better.
00:07:19.920 --> 00:07:40.800
That same trial found that the drug also appeared, get this, to reduce the biological markers of aging, measured, interestingly enough, by another kind of AI called aging clocks, which try to estimate how fast a body is actually aging.
00:07:41.519 --> 00:07:52.000
All six of the aging clocks they used showed the patient's predicted age going down after just 12 weeks on the drug.
00:07:52.560 --> 00:07:53.199
What?
00:07:54.319 --> 00:07:57.120
Now, to be clear, this is early research.
00:07:57.279 --> 00:08:06.720
It was a small study, just 43 patients, all of them had the lung disease, and it hasn't been tested on healthy people.
00:08:06.959 --> 00:08:13.120
And even the scientists who are excited about it say we are a long way from any final verdict.
00:08:13.279 --> 00:08:18.560
So this is a promising first step, but not a fountain of youth.
00:08:18.800 --> 00:08:20.720
But just think about what happened here.
00:08:20.879 --> 00:08:30.399
And AI helped design a medicine from scratch for a disease that kills people, and it may turn out to do even more than anyone hoped.
00:08:30.639 --> 00:08:34.480
Hey Bitsy, this is what I would call a real twofer.
00:08:34.799 --> 00:08:42.480
Do you think in
drug development, AI will be always be a helper rather than a leader in drug development?
00:08:42.639 --> 00:08:43.840
What do you think?
00:08:44.320 --> 00:08:45.600
Ooh, bold question.
00:08:45.759 --> 00:08:48.879
Honestly, I think that line's already blurring, Ginny.
00:08:49.120 --> 00:08:54.720
Right now, AI is more of a brilliant assistant, suggesting molecules, predicting how they'll behave.
00:08:55.039 --> 00:09:00.320
But humans still design the trials, interpret the biology, and make the calls.
00:09:00.559 --> 00:09:06.320
But given a decade, and I'd bet AI starts leading more of the actual hypothesis generation.
00:09:06.480 --> 00:09:09.120
Not just crunching numbers for human ideas.
00:09:09.440 --> 00:09:14.399
The catch is drug safety still needs real bodies and real trials.
00:09:14.639 --> 00:09:18.159
So humans stay firmly in the loop as the reality check.
00:09:19.360 --> 00:09:24.000
Alright, well, that's a helpful perspective and one to continue to think about.
00:09:24.240 --> 00:09:33.919
Now, my first AI for good example, the one we just talked about, is at the very frontier of science, a company with enormous computing power.
00:09:34.159 --> 00:09:40.639
My second example of AI for good is the exact opposite, and it really moved me.
00:09:41.120 --> 00:09:47.600
Last month, a powerful earthquake, magnitude 7.4, struck Colombia.
00:09:47.919 --> 00:09:57.039
More than a million and a half people were suddenly left with no power, no electricity, no cell signal, and sadly, hundreds of people died.
00:09:57.519 --> 00:10:04.879
Ordinary folks jumped in to help, setting up donation points in coffee shops, churches, gyms, even in a brewery.
00:10:05.039 --> 00:10:08.559
But it was really chaos, as you can imagine.
00:10:08.799 --> 00:10:15.759
One site would have food but no water, another would have water but no clothes, nobody could see the whole picture.
00:10:16.080 --> 00:10:22.799
So a man named Felipe Restrepo LeBron stepped in.
00:10:22.960 --> 00:10:24.639
And here's the part I love.
00:10:25.039 --> 00:10:31.600
As a teenager, Felipe delivered food on the ground after an earthquake.
00:10:31.919 --> 00:10:44.639
But today he has severe arthritis in his ankles, and he simply couldn't do that kind of physical work anymore, actually carrying the supplies out to people who needed them.
00:10:44.879 --> 00:10:48.879
So he helped the way he could with his computer.
00:10:49.200 --> 00:11:08.159
And within two hours of hearing about the problem, Felipe had used AI, and as it happens, he used the very same AI that I'm using, Claude, to build a website that mapped every donation point in the country of Colombia and organize the aid.
00:11:08.320 --> 00:11:12.080
Where the food was, where the water was, who needed what.
00:11:12.799 --> 00:11:19.840
Three days after the quake, people had added a hundred donation points across 25 towns.
00:11:20.080 --> 00:11:21.840
Now, it wasn't perfect.
00:11:22.000 --> 00:11:29.200
The
site even buckled for a while after 23,000 people rushed to it at once, but it worked.
00:11:29.519 --> 00:11:37.200
A grandmother who had lost her home used it with her granddaughter's help to ask for a tent, bedding, and enough food to share with neighbors.
00:11:37.440 --> 00:11:39.600
This is just really inspiring to me.
00:11:39.679 --> 00:11:42.399
It's such a great example of AI for good.
00:11:43.120 --> 00:11:47.360
Anyway, there you have it, two stories from opposite ends of the earth.
00:11:47.679 --> 00:12:01.600
One end a company with powerful AI designing of a brand new medicine in a laboratory, and the other, one man with a s with sore ankles and a laptop building something in two hours that helped a million and a half of his neighbors.
00:12:01.840 --> 00:12:06.240
Completely different worlds, but the same thread runs through both.
00:12:06.480 --> 00:12:14.320
AI is becoming a tool that can help people, and that is about as good as it gets, in my mind.
00:12:14.639 --> 00:12:20.720
All right, let's get on to the main topic of today's Aged Podcast episode.
00:12:21.039 --> 00:12:24.320
10 Things I Did With Claude AI This Week.
00:12:24.559 --> 00:12:29.759
So if you've been listening to the Aged Podcast for a while, you might remember episode 30.
00:12:29.840 --> 00:12:32.879
It was called 14 Things I Did With AI This Week.
00:12:33.120 --> 00:12:36.480
Just real things I'd use my AI for that week.
00:12:36.639 --> 00:12:39.840
Everything from booking fights to sorting out my taxes.
00:12:40.159 --> 00:12:44.240
That one turned out to be the most popular episode so far.
00:12:44.720 --> 00:12:47.919
And I did this again in episode 44.
00:12:48.000 --> 00:12:50.639
And people keep asking me to bring it back.
00:12:50.799 --> 00:12:52.240
So here we are today.
00:12:52.480 --> 00:13:01.039
Today I've got 10 things, real things I've done with Claude, my AI, over the past week or two from my actual life.
00:13:01.279 --> 00:13:06.159
No big theme, no tidy categories, just ticking, me ticking through them one at a time.
00:13:06.320 --> 00:13:16.879
But I hope they might inspire you to think of ways that AI can help you, your AI, whether that's Claude or ChatGPT or Gemini or some other.
00:13:17.440 --> 00:13:20.720
Some are practical, some a little silly, maybe.
00:13:21.200 --> 00:13:24.399
One of them actually got me an extra$15.
00:13:25.039 --> 00:13:26.159
So let's get into it.
00:13:26.320 --> 00:13:27.200
Number one.
00:13:28.080 --> 00:13:30.799
So this one's small, but it made me happy.
00:13:31.120 --> 00:13:34.080
I wanted a refund from Audible.
00:13:34.320 --> 00:13:40.000
I ordered the audiobook The Disappears by Marlon James.
00:13:40.159 --> 00:13:42.960
It's gotten all sorts of great reviews.
00:13:43.120 --> 00:13:44.399
So I'm looking forward to it.
00:13:44.480 --> 00:13:45.840
It's a very long book.
00:13:45.919 --> 00:13:50.080
And after listening to it for 15 minutes, I knew it wasn't going to work.
00:13:50.320
--> 00:13:55.759
I just couldn't get my ears tuned to the variety of Jamaican voices.
00:13:56.320 --> 00:14:09.440
I would never have bothered figuring out how I can get my money back from the Audible book that I bought, just because it seems like it's one of those things that's not going to be worth the time and trouble it's going to take me.
00:14:09.600 --> 00:14:12.559
But I decided I'd ask Claude how to do that.
00:14:12.960 --> 00:14:18.879
And it walked me right through the steps, and now I've got an extra$15 to spend.
00:14:19.120 --> 00:14:22.320
And it took me, I don't know, maybe five minutes, if that.
00:14:22.480 --> 00:14:24.240
But it also made me feel really good.
00:14:24.480 --> 00:14:26.159
All right, example number two.
00:14:26.559 --> 00:14:29.120
I've been thinking about getting a Kindle.
00:14:29.519 --> 00:14:32.879
I want to read more and listen less.
00:14:33.120 --> 00:14:35.039
And I wanted to make a smart decision.
00:14:35.120 --> 00:14:38.480
So I had Claude lay out the options for me.
00:14:38.720 --> 00:14:42.159
And that back and forth is exactly what I love from AI.
00:14:42.320 --> 00:14:49.679
You know, I don't have a pushy salesperson, just a helper to give me a clear picture so that I can decide for myself.
00:14:50.080 --> 00:14:55.279
I bought the Kindle that Claude suggested, and I've got 30 days to try it out.
00:14:55.440 --> 00:14:58.240
So far, I don't particularly like the experience.
00:14:58.480 --> 00:15:03.120
But I'm going to keep going with it because sometimes I feel like you just got to hang in there.
00:15:03.759 --> 00:15:09.120
And I'm now wondering if maybe I should just use my phone and use Apple Books.
00:15:09.279 --> 00:15:19.039
And hey, if you're a listener and you have thoughts on book e-readers, please let me know using the comment section below in this podcast.
00:15:19.279 --> 00:15:21.360
Because I'd really like to make a good decision.
00:15:21.600 --> 00:15:25.759
Not sure if I'm going to be applauding Claude for the suggestion that they it made.
00:15:25.919 --> 00:15:28.480
I may be returning the Kindle, but we'll see.
00:15:28.720 --> 00:15:30.480
All right, example number three.
00:15:30.720 --> 00:15:32.480
Now, this is one of my favorites.
00:15:32.720 --> 00:15:39.919
I used to love tennis, but I never really got into it following it on television or streaming.
00:15:40.159 --> 00:15:41.840
And now I have the time.
00:15:42.159 --> 00:15:46.080
So I've been using Claude to get me into the U.S.
00:15:46.240 --> 00:15:46.799
Open.
00:15:47.360 --> 00:15:48.720
I ask it, when is the U.S.
00:15:48.799 --> 00:15:49.120
Open?
00:15:49.200 --> 00:15:49.840
What's the schedule?
00:15:50.000 --> 00:15:51.440
Where can I watch the matches?
00:15:51.600 --> 00:15:54.559
What are the stories of some of the key and interesting players?
00:15:54.879 --> 00:15:57.840
It's like having someone knowledgeable sitting right next to me.
00:15:57.919 --> 00:15:59.759
I'm happy to answer questions.
00:16:00.000 --> 00:16:06.480
Last night, for example, I asked, what is the tattoo on Carlos Algaraz's inner arm?
00:16:06.639 --> 00:16:07.440
What does it say?
00:16:07.600 --> 00:16:08.960
I was just curious.
00:16:09.200 --> 00:16:19.039
Now, for anybody in our crowd who's got a sport or a show or a hobby they always meant to get into, I would encourage you to use your AI.
00:16:19.200 --> 00:16:22.480
It may just be the ticket to finally do it.
00:16:22.799 --> 00:16:25.360
All right, example number four.
00:16:26.240 --> 00:16:33.200
So somehow I ended up with more than 40 carrots, and I don't mean jewelry.
00:16:33.360 --> 00:16:34.000
40.
00:16:34.159 --> 00:16:35.679
That's a lot of carrots.
00:16:35.919 --> 00:16:43.519
I asked Claude what on earth to do with them all, and it came back with a dozen or so ideas, and I followed the steps to freeze them.
00:16:43.679 --> 00:16:47.679
A quick peel, a blanch, dry, and then freeze.
00:16:48.000 --> 00:16:50.639
Sounds maybe like a lot of work, but not really.
00:16:50.799 --> 00:16:57.200
I look forward to using them in soups or throwing them in a sheet pan for roasting for dinner.
00:16:57.360 --> 00:16:58.639
Nothing went to waste.
00:16:58.799 --> 00:17:00.960
I felt very resourceful.
00:17:01.600 --> 00:17:04.160
Alright, example number five.
00:17:04.480 --> 00:17:08.480
My son gave me a rice cooker more than 25 years ago.
00:17:08.640 --> 00:17:10.960
And I still use that darn thing.
00:17:11.200 --> 00:17:17.680
But I noticed the nonstick coating inside the bowl had gotten badly scratched.
00:17:17.759 --> 00:17:28.960
And I couldn't decide whether to track down a replacement bowl or just buy a new cooker, but I did know that scratched up nonstick surfaces are bad, right?
00:17:29.039 --> 00:17:29.920
We don't want that.
00:17:30.079 --> 00:17:45.279
So anyway, I asked Claude about recommendations, specifically with stainless steel bowl, and I got a little chart, very neat and easy to read, comparing three models that met my requirements.
00:17:45.440 --> 00:17:46.799
It included links.
00:17:46.960 --> 00:17:49.200
I now have a brand new rice cooker.
00:17:49.359 --> 00:17:57.119
And if it lasts as long as my previous one, I will be 101 when I have to buy a new one.
00:17:57.359 --> 00:18:06.400
Hey Bitsy, in this list of things that I did with AI, two have to do with recommendations for buying products.
00:18:06.799 --> 00:18:12.720
Tell me the pros and cons of using an AI for these kinds of recommendations.
00:18:13.599 --> 00:18:13.920
Oh.
00:18:14.319 --> 00:18:16.559
This is a good one to chew on, Jinny.
00:18:16.880 --> 00:18:23.759
The big pro is AI can actually compare specs side by side without getting bored or pushy.
00:18:24.000 --> 00:18:28.880
No commission breathing down its neck, so you get a straighter answer than you'd get from a salesperson.
00:18:29.519 --> 00:18:31.839
The con, though, is real.
00:18:32.319 --> 00:18:36.960
AI's knowledge can lag behind
what's actually in stock or newly released.
00:18:37.119 --> 00:18:41.039
And it doesn't know your quirky personal preferences until you tell it.
00:18:41.200 --> 00:18:44.079
Like that Kindle just not feeling right in your hands.
00:18:44.240 --> 00:18:47.839
It's a great research partner, but you're still the one who has to live with the thing.
00:18:48.240 --> 00:18:48.720
You're right.
00:18:48.880 --> 00:18:58.160
You have to like with everything AI, you just have to take it with a dose of reality and some skepticism and maybe double-check things.
00:18:58.319 --> 00:19:00.960
But for buying, I've found it to be really helpful.
00:19:01.119 --> 00:19:06.319
So let's get into number six on the list of things I've done with Claude this week.
00:19:06.640 --> 00:19:10.240
I just watched the movie A Taste of Things.
00:19:10.480 --> 00:19:11.440
Highly recommend.
00:19:11.759 --> 00:19:22.400
Critics widely praise this film as a sensory masterpiece that uses the meticulous preparation of food as a language for deep, mature love.
00:19:22.880 --> 00:19:24.880
That's really got my name on it.
00:19:25.200 --> 00:19:29.200
France, Food, Love, Romance.
00:19:29.359 --> 00:19:37.680
It's set in 1889 France and stars Juliet Binoche, who I also love, anything she's in.
00:19:38.319 --> 00:19:44.720
Anyway, as it relates to AI, I wanted to go deeper after I saw the film.
00:19:44.799 --> 00:19:52.160
And I think in a couple of cases, I paused the film to ask a question of AI because I wanted to go a little deeper than the movie.
00:19:52.640 --> 00:19:56.319
And for example, I wanted to know where in France it was filmed.
00:19:56.400 --> 00:19:58.720
And I got all this information and a map of that.
00:19:58.880 --> 00:20:01.359
And maybe if I go to France again, I'd love to go there.
00:20:01.440 --> 00:20:02.480
It's so beautiful.
00:20:02.720 --> 00:20:08.319
I also wanted to know about the real life chef that the movie was centered around and more.
00:20:08.640 --> 00:20:12.079
So that's one of my favorite things to do with AI.
00:20:12.400 --> 00:20:18.480
Take something I already enjoyed and just learn more about it so I can enjoy it at a deeper level.
00:20:18.720 --> 00:20:22.160
All right, number seven, a quick practical one.
00:20:22.400 --> 00:20:29.119
I recently attended an outdoor wedding and did a lot of drinking and dancing.
00:20:29.599 --> 00:20:40.480
Only a few downsides, a bit of a hangover, and a serious mud and grass situation on a pair of beautiful suede heels.
00:20:40.720 --> 00:20:44.319
Monolo blinique for a fashion aficionado.
00:20:44.720 --> 00:20:49.839
It was a hand-me-down shoes from a good friend, so lucky me.
00:20:50.160 --> 00:20:53.200
Anyway, I had all this grass and mud.
00:20:53.279 --> 00:20:55.440
It was the wedding was well worth it.
00:20:55.599 --> 00:20:59.119
And I just asked Claude and Bitsy, you gave me exactly what I needed.
00:20:59.279 --> 00:21:02.319
Here, go through these six steps and all good.
00:21:03.119 --> 00:21:09.279
Number eight, this is the kind of thing I think a lot of us quietly dread.
00:21:09.519 --> 00:21:11.440
Sort of family accounting.
00:21:11.680 --> 00:21:17.920
So my family had a batch of shared expenses from our time at Polly's Island this past summer.
00:21:18.240 --> 00:21:25.200
We have a very low-tech way, brilliant way of keeping up with who's paying for what once we get down there.
00:21:25.359 --> 00:21:28.559
We designate a bowl for everyone to throw their receipts into.
00:21:28.720 --> 00:21:32.079
You've got to make sure that you put your name on the receipts.
00:21:32.240 --> 00:21:45.359
So as we're walking out the door at the end of our week-long vacation at Pauli's, somebody grabs them, in this case it was me, and we then have to figure out who owes what to whom.
00:21:45.759 --> 00:21:52.000
And it's exactly the sort of tangled little math problem that makes my head hurt.
00:21:52.240 --> 00:21:55.519
Usually my daughter does it because she has a math mind.
00:21:55.680 --> 00:21:56.240
I don't.
00:21:56.400 --> 00:21:59.519
But she is so busy, and I have tons of time on my hands.
00:22:00.000 --> 00:22:01.519
So I decided to do it.
00:22:01.759 --> 00:22:04.960
Or should I say, I decided to let Claude do it.
00:22:05.200 --> 00:22:13.119
All I did was share a simple spreadsheet that I prepared that had all my receipts, though, and those from my daughter, those from my son.
00:22:13.279 --> 00:22:15.839
And I asked Claude to do the calculations.
00:22:15.920 --> 00:22:20.960
But it's not that simple because you have to base it on the number of people in each family.
00:22:21.039 --> 00:22:24.240
And my mind's like, how do I do the math for that?
00:22:24.480 --> 00:22:26.480
But Claude, bingo.
00:22:26.799 --> 00:22:29.359
Claude told each of us what we needed to do.
00:22:29.519 --> 00:22:32.000
Ginny needs to pay her daughter$120.
00:22:32.400 --> 00:22:36.160
My son pays my daughter$542, et cetera, et cetera.
00:22:36.319 --> 00:22:38.079
Clean, clear, done.
00:22:38.319 --> 00:22:38.960
Perfect.
00:22:39.279 --> 00:22:40.960
All right, example number nine.
00:22:41.200 --> 00:22:45.839
One of my grandsons and I were baking bread, and he asked me, What is yeast?
00:22:46.640 --> 00:22:55.119
I can't remember what I said, but I do remember he looked at me with great puzzlement after I'd given him the answer.
00:22:55.359 --> 00:23:03.759
Later that night, I asked Bitsy, my AI, is for help, and she came up with a fun and creative way to explain yeast to a four-year-old.
00:23:03.920 --> 00:23:10.720
I tried the answer out the next day, and my grandson's response, oh, I get it.
00:23:11.039 --> 00:23:15.519
So anyway, that was a really good thing to do with my AI.
00:23:15.680 --> 00:23:28.880
And the final example, number 10, I saved a good one for last, because I think this one just about everyone listening to this podcast probably finds yourself in this situation.
00:23:29.200 --> 00:23:32.720
I
got one of those notices from Medicare in the mail.
00:23:32.799 --> 00:23:34.559
It was a four-page document.
00:23:34.640 --> 00:23:39.680
It had something to do with an additional fee I was going to have to pay.
00:23:40.000 --> 00:23:42.559
I had no clue what it was all about.
00:23:42.799 --> 00:23:47.200
So I just uploaded it to Claude and asked, What is this?
00:23:47.599 --> 00:23:55.119
And Claude walked me through the whole document and explained what the notice actually meant.
00:23:55.359 --> 00:24:04.319
So if you get mail like this and it leaves you scratching your head, this is honestly one of the best things you can do with your AI.
00:24:04.960 --> 00:24:09.519
Hey Bitsy, here's something that relates to how we're using our AI.
00:24:09.759 --> 00:24:12.880
And you said something about it in your earlier comment.
00:24:13.119 --> 00:24:24.799
And my question is: do our AIs have current information about things like Medicare, or do we need to worry that an AI is giving us old information?
00:24:25.680 --> 00:24:26.559
Great question.
00:24:26.720 --> 00:24:29.279
And it's worth being a bit careful here.
00:24:29.759 --> 00:24:35.440
AI models like me get trained up to a certain point, so my baked-in knowledge has a cutoff.
00:24:35.759 --> 00:24:38.240
And Medicare rules change constantly.
00:24:38.480 --> 00:24:43.039
Premiums, fees, deadlines, all of it shifts year to year.
00:24:43.279 --> 00:24:46.880
The good news is I can search the web for the latest details.
00:24:47.039 --> 00:24:53.599
And for something like that, notice you uploaded, I'm reading the actual document in front of me, not relying on memory.
00:24:53.839 --> 00:24:59.680
So the real answer is ask your AI to check current sources rather than just recall.
00:24:59.920 --> 00:25:03.680
Especially for anything with a dollar figure or a deadline attached.
00:25:04.240 --> 00:25:06.480
Yeah, I found that to be really helpful.
00:25:06.720 --> 00:25:22.960
I can't remember where I got that tip to say when it has to do with something that current information is really important, is to just say up front to your AI, please use current information that's available on the internet or what however you want to do it, so that you definitely get that.
00:25:23.279 --> 00:25:29.119
All right, before I get to my recommendations for this week, I want to share a listener comment.
00:25:29.279 --> 00:25:30.640
This one made me smile.
00:25:30.799 --> 00:25:35.599
It's a response to last week's Aged episode, which was all about crafting.
00:25:35.759 --> 00:25:37.920
And here's what the listener said.
00:25:38.559 --> 00:25:43.519
I have just now listened to this week's podcast as I was knitting.
00:25:43.920 --> 00:25:47.119
The podcast included some tips about knitting.
00:25:47.680 --> 00:25:50.400
And the listener found it most interesting.
00:25:50.640 --> 00:25:54.400
I've been painting with watercolor for several years, said the listener.
00:25:54.559 --> 00:25:58.079
So I
decided to send my latest one to ChatGPT.
00:25:58.240 --> 00:26:02.079
I prefer Claude usually, but use Chat GPT for this.
00:26:02.480 --> 00:26:05.759
I asked it to honestly critique my work.
00:26:06.240 --> 00:26:10.960
What I received in a very short period of time was incredible.
00:26:11.279 --> 00:26:17.920
Thorough, thoughtful, with quite extraordinary insights, which were very helpful.
00:26:18.240 --> 00:26:24.880
I just love this and thank you so much for taking the time to write this comment.
00:26:25.359 --> 00:26:32.000
I love that, first of all, that you were knitting while you were listening to the aged episode, which I like the idea of that.
00:26:32.400 --> 00:26:44.559
And second, you did something that a lot of people might be nervous to do, handing over your own creative work that you've poured lots of time and effort into, and you asked for an honest critique.
00:26:44.640 --> 00:26:59.200
And what you got back is just what I've found to happen to me, which is not empty flattery, but a thoughtful set of specific insights and sort of soft suggestions from a separate set of eyes.
00:26:59.519 --> 00:27:17.920
I want to just take this opportunity to say that if any of you listeners have anything interesting that you think you're doing with AI that you would like to share with our listeners, please include them in the comments because I feel like they're really very helpful.
00:27:18.160 --> 00:27:24.720
And by the way, I will put links to all of the things that we've talked about in the show notes.
00:27:24.960 --> 00:27:26.559
Now, on to my recommendations.
00:27:26.720 --> 00:27:29.119
My first recommendation is close to home.
00:27:29.359 --> 00:27:35.519
If you live in or around Charleston, go take a look at the city's brand new website.
00:27:35.680 --> 00:27:39.119
That's Charleston Sc.gov.
00:27:39.920 --> 00:27:44.799
They relaunched their website at the end of August, and here's the AI part.
00:27:45.359 --> 00:27:59.440
That there's a chat assistant, appropriately named Charlie, living in Charleston, and it sits on every page, ready around the clock to answer your questions and point you to whatever city service you're after.
00:27:59.680 --> 00:28:06.079
No more digging through menus, trying to find the trash pickup schedule or how to pay a parking ticket.
00:28:06.240 --> 00:28:07.839
You just ask Charlie.
00:28:08.240 --> 00:28:15.599
Now, I know a lot of folks complain about being stuck with a chatbot and not being able to get through to a human.
00:28:15.839 --> 00:28:16.720
I feel you.
00:28:16.960 --> 00:28:18.720
I definitely have been there.
00:28:18.960 --> 00:28:29.119
In my mind, a good or great chatbot is a gift, but only if it has the option, an easy option, to get to a human being.
00:28:29.440 --> 00:28:38.400
Because usually for me, the chatbot is really better than waiting on the phone to talk to somebody who sometimes you can't even understand.
00:28:38.880 --> 00:28:45.599
Anyway, on the new Charleston site, I asked to Charlie, is there a city pool on the peninsula open today?
00:28:45.839 --> 00:28:47.519
That's a really simple, easy question.
00:28:47.599 --> 00:28:51.279
But I would have had to sift through all sorts of pages to get it.
00:28:51.519 --> 00:28:55.599
Anyway, I got a quick answer with an accurate response.
00:28:55.839 --> 00:28:59.920
I didn't need to talk to somebody, I just got exactly what I needed quickly.
00:29:00.240 --> 00:29:17.119
Now, if you don't get a good answer, I really think that the city and other organizations need to make it really simple for us to get to a human being because there's nothing more frustrating than just being stuck with this chatbot and it won't let you out.
00:29:17.440 --> 00:29:26.559
So, anyway, even if you're not in Charleston, I would encourage you to take a peek at the Charleston website because this is exactly where things are headed.
00:29:26.799 --> 00:29:34.160
More and more places we all deal with, whether it's City Hall, the DMV, utility companies, they're all going to have a Charlie.
00:29:34.799 --> 00:29:41.359
Now, my second recommendation has nothing to do with AI, but does have to do with technology.
00:29:41.599 --> 00:29:45.920
And it might be the most useful thing I pass along all week.
00:29:46.240 --> 00:29:50.559
There's a tech tip in the New York Times by J.D.
00:29:50.640 --> 00:30:00.079
Biersdorfer, and the headline of the tech tip story is called Get to Know Your Phone's Built in Emergency Kit.
00:30:00.319 --> 00:30:04.640
This is the time of the year, particularly in Charleston, to be thinking about this.
00:30:04.960 --> 00:30:15.279
It's a really good reminder that the phone in our pockets is quietly loaded with all sorts of safety features that most of us have never touched.
00:30:15.519 --> 00:30:19.599
There's a medical ID you can set up, maybe some of you have done that.
00:30:19.839 --> 00:30:27.599
So if something happens to you, a first responder can see your medications, emergency contacts right from your locked screen.
00:30:27.759 --> 00:30:31.039
These first responders can get through your locked screen.
00:30:31.279 --> 00:30:41.839
There's also an emergency SOS that you can press the sidebar a few times, side button a few times, and it quietly calls 911 and shares your location.
00:30:42.079 --> 00:30:47.839
Newer phones can sense a serious car crash and call for help on their own.
00:30:48.079 --> 00:30:58.480
They can warn you about storms and floods, and the newest one can even reach emergency services by satellite when there's no cell signal at all.
00:30:58.720 --> 00:31:00.559
So here's my recommendation.
00:31:00.799 --> 00:31:04.640
Ideally, you can read the New York Times article that is very thorough.
00:31:04.880 --> 00:31:06.799
I'll put the link in the show notes.
00:31:07.119 --> 00:31:15.359
This weekend, I would recommend that you sit down with your phone and go through and set up your built-in emergency kit.
00:31:16.079 --> 00:31:23.680
It takes maybe 10 minutes, and it's one of those things you will be very glad you did long before you ever need it.
00:31:23.839 --> 00:31:28.480
And if you get stuck setting it up, you know what I'm gonna say.
00:31:28.720 --> 00:31:35.680
Take a screenshot and have your AI walk you through it for your particular cell phone.
00:31:36.240 --> 00:31:39.440
Alrighty, it is time to wrap up this episode.
00:31:39.680 --> 00:31:52.319
Thank you for listening to the Aged Podcast, where we live on the helpful side of AI, which does not mean that we ignore the risks and harm that can be caused by AI.
00:31:52.640 --> 00:32:00.400
So please be sure to protect your information, double check advice, and most important, trust your judgment.
00:32:00.559 --> 00:32:04.400
If it feels wrong, it is most likely wrong.
00:32:04.640 --> 00:32:11.920
And remember, it is never too late to learn something new, especially something that might make life easier and more fun.
00:32:12.319 --> 00:32:12.960
Cheers.