Welcome, everybody, to Midday Magazine for this Tuesday, November 12th, 2024.
Have your host, James J. Mailoff here, and it's time for Reality Radio.
We are going to speak with David Farmboro right now, local broker here in our area, and
David and I are going to talk about real estate in general, but we're taking a different
approach to it today, David.
We're going to be talking a little bit about AI in real estate, and if you don't mind,
I think one of the good ways to do this, just to make sure to cover our bases, if you
will.
Let's get into what AI is a little bit.
I feel like you and I were talking about this in our pre-game a little bit.
So many people hear the term AI, and they may think they know what it is, or they may
not really have a comprehension of what it is.
Dive into a little bit of what AI refers to and what it means.
Hello, James.
It's great to be back with you, and yeah, there are a few things that I was planning on
saying, and then I realized that probably people don't want to hear a lecture about the
history of learning machines going way back to the Babbage's analytical engine.
Nice.
Wow.
In the Victorian era.
As a side note, I want to hear that episode, so we'll do that on a side, or just as a side
podcast.
We'll talk about that some other time.
But one of the things that we've striven for as the human race has been to make machines
which learn in the way that we do.
So anyone can make a machine that just does what you tell it, and you can program your
rumble to go around the floor and clean the floor.
But what you want it to do is to learn where the objects are and then to avoid them.
And artificial intelligence does those things, but it also takes it a step further.
And the reason we can do that is that year on year, the capability of microprocessors
gets more and more powerful.
So there's a thing called Moore's Law, which I'm sure you've heard of, which is a saying
which was established about 50 years ago by one of the founders of Intel, the microprocessor
company.
That is that every two years the number of transistors built into an integrated circuit doubles.
And you can extend that to say that basically the capability of what we do with our computers
doubles every two years.
Now that's obviously an approximation.
But if you imagine something doubling and doubling and doubling, well, first of all it's
not very much.
Instead of two, you've got four, then you've got eight, then you've got sixteen.
But it's not long before you're going up to one thousand and twenty four, two thousand
and forty eight.
It gets very big, very quickly.
And this is what's happened with our technology.
The rate of increase is astonishing.
We have gotten to a point and we got to this a while ago, actually, where the technology
advanced quicker than we could advance with it.
I think that's a very good way of talking about AI.
And in fact that's one of the reasons that people are saying, a lot of people are saying
I'm a bit scared of AI.
And there's good reasons for that because we invented the atom bomb and then we didn't
really know how to control it.
So the AI is something, sorry, AI is something which we use all the time.
It's all around us.
And we really don't notice it unless we know it's there.
So for example, I watched a video on YouTube a couple of days ago about President Grover
Cleveland.
And I was watching and I was thinking there's something weird about this.
The voice is mispronouncing a few words.
It's putting the slight wrong emphasis on some of the words.
And then I realized the whole thing was computer generated.
So the video was made by somebody in India, but the voice was an American voice, but
it was AI generated.
And also all the pictures in the video were generated by AI.
And I would be very surprised if the script was not also generated by AI.
So we do have AI in so many places.
If you have one of those home speaker devices, I was going to say the name of it,
L-E-X-A.
But if I said the name, everybody's going off.
So if you have one of those, if you have an echo in your house, the voice that talks
back to you is not a recording, it's not a synthesis, it's an AI voice.
So just to briefly explain one of the ways that it works, it uses a thing called oppositional
or adversarial networks.
So what you have there is rather than just having a computer generating something, you
have two computers talking to each other.
So one generates, and then the second gives it feedback and says, this doesn't look
right.
And then the first, the generating AI says, oh, okay, I'll change it a little bit.
And the second one said, that's better, or well, no, you still need to tweak this.
And it's not just the four stages that I've outlined, it just goes on and on very, very,
very quickly.
And until you get these amazing things, like if you've ever seen somebody can generate
an AI photograph of themselves that looks like it's really them.
I did that once for my Facebook profile.
And then somebody saw it.
In fact, it's on our community theatre website because Linda, the wonderful lady who does
our web page, just took my photo from Facebook and she had not realized it was AI generated.
So my picture on the community theatre website is actually an AI Artifice.
It's that good.
Wow.
And it explains to so much of this explains, I think, where we've gotten to this point,
where we have AI, whether it is, you know, where we're texting somebody and autocorrect
correct something, or some of the bigger things that, like you'd the Grover Cleveland video
that you refer to.
I think that a lot of this too, when we're talking about these things, whether we're repeating
anything anybody knows or not, it is normalizing and getting people more and more familiar
with these terms, with what this is.
I don't believe that AI is something too fear.
I think what human beings do with it certainly can be a concern.
But especially when we're talking about the entertainment industry and some other aspects
of it.
But the more we understand AI, the better we, I think the more comfortable we get with
it.
We keep an eye on it certainly, but also at the same time, oh, wait a minute, I don't
have to be afraid of that.
If you've ever, if you've been afraid of the dark at any point in your life, at some point
or another, hopefully you get to a place where you realize, well, when I, when I'm looking
at a dark alley, it looks really dark, but once I get inside it, well, it's actually
kind of lighten here.
It's not that dark at all.
I can actually see my hand and things.
I feel like a lot of this is that.
Once we understand something, we don't fear it as much.
We may not like it.
We may not be a fan of it, but that's very different than having a fear of something.
The more we talk about these things, and part of the reason I'm happy to have this conversation
is, hey, we're talking about AI.
We're familiarizing people with it, getting you more and more familiar with it.
That's absolutely right.
And I think we also have to look at it in historical context.
So if we look at the industrial revolution, there were a lot of people who were very much
against some of the machines.
So one of the typical examples is the threshing machine.
People would have the threshing machines really simplify and speed up the work on a farm
at Harvest.
And all the people whose job it had been to manually take those bales of crops and thresh
them, now said, well, hang on, I'm being put out of a job.
And they sabotaged the machines and they got in terrible trouble.
Nowadays, 200 years on from that, we look at that and we think, well, hang on a minute,
those people could have been threshing machine operators.
So when we look at something like AI, it's slightly different because there will be some
jobs which won't be so important.
So I'll give you a case in point.
I'm wearing a shirt now with my Rapids Homes logo on it.
And I designed that using AI.
Now I don't want to say AI did it because if AI had done it, it wouldn't have been good.
What I did was I had an idea.
So you know, the initial imagination is mine.
Right.
I fed that into an AI and the AI generated several examples.
I wasn't happy with those.
So I refined my instructions and made it generate some more.
When it came up with one, they thought that looks really good.
I took that and then I had to adjust it.
Why did I have to adjust it two reasons?
Firstly, the words Rapids Homes had garbled and it came out with something like Roman
Homes or something.
So that's one of the limitations of image generation that people are having six fingers
which you sometimes see.
The other thing was that there were just a few tweaks I had to make, but after that,
I took it to a professional graphic designer who actually made some really good tweaks
which depended on his skill knowing things like which colors work best and how to balance
the design.
So yes, I used AI, but I didn't just generate something and go straight ahead and print it.
Right.
When it comes to, I'm a horrible speller.
I am one of the world's worst spellers.
So when it comes to AI, often in text or in email or something like that, it saves me
David.
It makes me sound a little bit smarter than I am because it corrects words for me.
So that is helpful and it works a lot of way.
But it has its limitations.
Certainly, anybody who has ever tried to text a swear word understands this.
And to that point.
I don't give a shift.
Nicely done.
So to what David and I are talking about here as far as what AI is capable of and its
limitations, I think for one, to go back to what we were saying before, it gives a little
less fear I would think in people like, hey, this isn't something that can necessarily
do much without your assistance and without your help.
That's right.
And I think that's a key part of this conversation.
No computer has ever turned itself on.
No AI system is, you know, thinking for itself or anything, David kind of broke this down
for you of how that works and how your participation in this goes.
And with that kind of shows not only the limitations, but I also feel like it shows the control
that we have over these things.
Yeah, that's absolutely right.
We have to realize that it is within our control.
So for example, I do notice that some realtors use AI to create their listing descriptions.
And you've got to be careful because often what happens is it will fill it full of cliches.
I've even seen it happen where somebody fed in to chat GPT that they want to write a
listing description for a house with five bedrooms and three bathrooms.
And the listing description that was spat out was quite good.
But it also said that there was an in-ground swimming pool.
And the reason for that is that the AI just took typical other examples of listings with
five bedrooms and three bathrooms and obviously thought that swimming pool would be a good
thing to add.
Now, if that agent had then just gone ahead and published that, that would have led to
problems.
So you've got to read it and you've then got to correct it and improve it.
We're speaking with David Farnborough and Reality Radio right now.
And we've talked a little bit about AI, what AI is and how it is.
It's limitations, the work that it can do.
Let's talk about AI in real estate, David, and get into how it applies with the real
estate industry.
OK.
Well, there's a couple of examples I just gave, like the logo generation and the listing description.
One of the ways that I use AI a lot is in smoothing and upscaling images.
So a lot of the time, some of the images I take, they may have problems because of low
light, which are just functions of the building that I'm photographing in.
And an AI can take that image and it can brighten, wear appropriate, just so that people
can see what they're meant to be seeing.
Sometimes there are things like you're not getting quite the definition you need on an
image and an AI can upscale that.
So for video generation, they can do a lot of, they can actually generate a video from
a 3D image that I take with my 3D camera, which can be immensely helpful.
I also use them for virtual tours because I do virtual tours for all of my listings.
And although I'll only take a 3D photograph of each room, the AI will be able to put together
the transition from room to room so you actually feel like you're moving through the property.
The other thing that it will, it will often do, in fact I also do this for every listing
I have, is that I use it to create plans.
Now floor plans are really important for houses because sometimes you're looking at the photographs
and you're thinking, well, I'm in it, which bedroom is which and you need to know things
like whether you have to walk through a bedroom to get to another bedroom and a floor plan
can really help to make that clear before you've even set foot in the property.
But again, you have to be careful because it does make mistakes and I've come up with,
you know, often I've been given the results back that are not quite accurate and so I
think, well, I've got to tell it where the staircase is and so I have to go back and
correct.
But that's the important thing, you need to have oversight and you need to not be lazy.
That's the thing.
This doesn't mean you can be lazy, it means you can work harder on other things.
The intent of AI was always to do what, really what you're referring to here, David, I'm
doing my job, hey, can you fill in the blanks AI, hey, can you help polish this a little
bit maybe or do something that I just physically cannot do or my camera can do or something
like that?
It's an assistant.
It's a fork.
It's a spoon.
It's something that assists us in our job and assists us in what we do.
That's right.
And it doesn't mean the graphic design is being put out of work because I still had to go
to the graphic design.
Yeah.
And I think that part of this is interesting too because especially, well, we're talking
about real estate and all these things that you mentioned there are great.
I love the idea of anything that can help out our agents out there and certainly helping
out home buyers and sellers.
Are there things that people should be aware of like, hey, you want to keep an eye on
this part of it?
Because we're, you know, some of the, I don't want to say fear or anything like that.
But, hey, this is something that could be an issue, could be an off-app price to look
out for.
Yeah.
I'll give you a very good example and that is online valuations.
I went around to do a listing presentation at a person's house the other day and I pulled
up online valuations from four different websites and the lowest was 115,000.
The highest was 230,000.
Now if those are done by AI, why aren't they accurate?
Well, the AI has not been inside the house.
The AI doesn't know about the additional bedroom they put in.
The person hadn't sold this particular home in 20 years so there's no listing data that
the AI can take from the MLS.
So the AI's can be working with very limited data and they do their best job but you just
have to be cautious.
So if you're looking at the zestimate or whatever realtor.com calls their estimate, you got
to be careful because that does not mean that's the value of the property.
Well, and, you know, from what the little bit of research I was able to do on this, a
lot of your bigger conglomerate kind of like agencies might use these things.
It's part of why we encourage you to always, you know, buy local support local and especially
in this industry.
When you're talking to a local, you're talking to a community member.
When you're talking to maybe a future neighbor or just at least a community member, somebody
who is certainly, I want to say has stake in this because all realtor's I believe do have
stake in what they're doing.
It's so much of a reputation business.
But when it's a local person you're talking to, you're really getting somebody who has
been in that building, has been in that house and really understands what they're talking
about and is talking to somebody that they want to put in that home.
They want to help.
They want to put in the right home, not just any home.
That's right.
And also a local person will know the difference between a home that's, you know, just to one
side of the airport, like the one I was looking at the other day and one which is maybe right
next to eighth street.
And you know, I was looking at the valuations and I was, and actually I use AI tools to
help me with valuations as well, but I use my brain too.
My AI suggested I used a house that was over in Grand Rapids, you know, way out in the
country because it said, okay, it's got a Wisconsin Rapids mailing address.
It's got five, four, four, nine, four and it's three bedroom, two bathroom, but it was
completely different.
It was completely different in terms of the area and in terms of the style of house.
So looking at, looking ahead here, you've mentioned it, you use AI a little bit and everything.
Do you see this being a real asset to the real estate industry as we kind of fine tune
this morning, we get AI to be more harness to what we need it to be and everything?
It already is and I've been using it probably for at least two years.
I think that the things to be careful of is not to be overconfident to supervise it,
you know, just like you would any assistant and to make sure that you use the time you've
saved with the AI to work harder on other things.
Hmm, very well said, a great topic today, David.
This was a lot of fun.
This was great getting into this and hearing how this impacts an industry that I don't
think we get to talk about nearly enough with AI in that.
So appreciate the topic today and as always appreciate the time with you.
We're looking forward to next month again already in another episode of Real It To Radio.
David, if people have followed questions one and the more about whether what we talked
about today or just you as a realtor, how can they get in touch with you?
Well, they can look me up on Facebook and they can tell me any corrections they need
to about my technical description of AI on my Facebook realtor page, David Farmer
Realtor, or they can text me on 715-323-8594.
I encourage you to do that everybody by local support, local support those that support
this community and support us.
We appreciate you, David.
Thank you again so much for the time looking forward to next month and I'm sure we'll have
some fun things to talk about and I know that you and I got a pregame a little bit here.
So we'll have some good stuff.
I'm sure.
We'll have another Midday magazine lined up for you tomorrow, everybody.
Looking ahead at the rest of the week here at Midday Magazine from 4-5, be sure to join
us tomorrow.
We'll have family, natural foods, and with us Steven and Katrina Hentner, and in part
2, encourage.
On Thursday, we'll be talking with our friends from UW Extension and Amitia will be with
us on Friday, Senator Patrick Tesson, and then we'll wrap the week up with breakfast
with Santa.
Pete's going to be in for that.
Be sure to be listening for community stories with Melissa K, 5-6 Tuesdays and Thursdays,
plenty more right here at 97-5 FM 13-20 AM WFHR.
We are locally grown radio.
