Paul Stenhouse: ChatGPT adds app integration and Amazon's prescription vending machine

Apps have arrived inside ChatGPT  

Another big moment in the company's offerings – this is where ChatGPT has gone from a tool to a platform. Now inside ChatGPT you can ask it to directly ask an app to do something. So, you can ask Spotify to create a playlist for you that is 150bpm for a 45 minute run. Or you can ask it to research hotels in Paris on Booking.com, and it'll return real results you can then book with a click.  

You don't need to search through those sites or apps, ChatGPT becomes the interface for that. The apps at launch are selected by OpenAI and more will be added over time  including retailer Target, restaurant booking app OpenTable, and Uber. They’ll also be adding them for the EU users "soon".  

Travel and personal assistant type use cases seem to be the top priority for this launch.  

  

Amazon is launching a prescription vending machine  

It owns a doctors clinic network in the USA called One Medical, and inside those locations will be a vending machine for commonly prescribed medicines, dispensed by Amazon Pharmacy. It'll debut in Los Angeles starting in December.  

How it works: after the patient visits the doctor, they can choose to send their prescription to the vending machine. You then login to your app, scan it, and away you go. It'll have commonly prescribed medications including antibiotics, inhalers, and blood pressure medications, with each location's inventory selected based on needs (I smell some AI here!). Controlled substances and medications requiring refrigeration are not available via the vending machine. 

 

LISTEN ABOVE 

See omnystudio.com/listener for privacy information.

Episode Transcript

Speaker 1 (00:07):
You're listening to the Saturday Morning with Jack team podcast
from News Talks at B.

Speaker 2 (00:12):
Apps are coming to chat GPT. I know that seems
confusing because if you're like me, you might have chat
GPT as an app. But basically they're turning from chat
gpt just being a tool into a platform. I know
this is all kind of jargony stuff. That's why our
Textbert paulstine House is here to explain exactly what this means.
In Paul, this is a big change.

Speaker 3 (00:35):
Yeah, this is a big change because it is showing
that chat GPT is trying to be so much more right.
It is trying to be a platform. It is trying
to be an experienced I don't know how else to
describe it, Jack, but it wants to be the thing
that you use to do everything.

Speaker 2 (00:51):
Yes, basically, I think that's the way to talk about it.
It's going to be it's like the portal through which
you access the rest of the Internet. Think about it
like that.

Speaker 3 (00:59):
Yeah, and exactly. So we started out with directories like
Yahoo those yeah, where you like literally was like the
Yellow Pages, and then we went into search with Google.
That was pretty cool, and now this is sort of Yeah,
the next big leap where it all gets merged together.
So you can imagine a time when put on your
put on your imagination for a second. You're looking for

(01:21):
a playlist, right, so instead of just chat GPT giving
you a random list of songs and then you have
to go into your app and find them all. If
it's got these apps inside of chat GPT, you can
ask Spotify, say, to create you a playlist for you
that's good for a run. Or you can say, I'm
going to Paris booking dot com app and I want

(01:42):
you to give me some results, and I only want
to spend this much, and I want the hotel to
be here in the city and blah blah blah blah blah. Basically,
you can use it to browse the internet and do
things for you. Yeah, now, of course browse the internet
not wide out. I mean it can browsey open, browse
the internet like that. But these specialist apps will mean

(02:03):
you can do specialist things. Worth pointing out that there
are only limited ones at the moment that are being
in this sort of launch phase. They say more of
them are coming later this year, including in the US.
The big retailer Target is also a restaurant booking app
over here called open Table, So you can be finding
a restaurant for Jack and if you go for dinner,

(02:24):
be great, wouldn't it. And Uber so you could even
say like hey, you know, if it's Uber Eats or something,
you're like, hey, I want Indian tonight, like find me
the best Indian around, and it would do it.

Speaker 2 (02:34):
I mean, so it seems to me, be your personal assistant,
that this is going to be Yes, it seems to
me like that there's a couple of really important dimensions here.
First of all, from a user experience, I can see
how this would be extremely convenient, right, Like if you
said that to use your holiday example, I'm want to
go on holiday here, I only want to spend this
March and like make me an itinery and you say yes,

(02:57):
I like that Booker and it does it. I can
see how that's extremely convenient. Presumably, the trade off is
that you have to that your data is being shared
with these big other companies as well. So if it's
Booking dot Com that is integrated into chat GPT, then
all of Paul Stenhouse's preferences are being feed into Booking

(03:17):
dot Com, so that Booking dot Com then knows it
about you as well, and then conspiracy to you. How
do we know that If you say to chat GPT,
what is the best way to book a trip? Then
instead of just giving you the best way to book
a trip based on your preferences, it doesn't immediately say why, Paul,

(03:38):
you should use booking dot com.

Speaker 3 (03:41):
Exactly and you know what ads are coming to yeh
GPT right right by figuring out how that's going to be.
And that's I mean, that's the Google Search model we
started out by. Just you know, you think, go back
to those examples they gave. You have a directory, you
want to be the promoted one in the directory, you
pay a bit more on your top Google search exactly
the same. You want to get the keyword, you get
to the top. I can imagine it's going to be

(04:03):
exactly the same. Now you do have to invoke this thing,
so you have to see, hey Booking dot com do
this thing, or hey Expedia do this thing. And if
you look at their launch partners, they have got a
couple of them. It will be interesting to your original
point of how much data does get sent, whether it's
just the data in that chat that you were having
at that time and when you invoke Booking dot Com, say,

(04:26):
but they also have just integrated the ability to pay
as well with them chat GPT, So you could go
through the whole experience, say using Booking dot Com, and
you've never gone through Booking dot com website, You've never
done anything on your phone. You have searched for it all,
you've browsed it, or you've selected your hotel, you pay
for it everything. The only thing you're getting probably is

(04:48):
your ow itinerary email to you, and that's the only
interaction you really have with them outside of chat gpt.
Kutty crazy, but it's interesting, it's interesting looking to it.
But the types of companies that they've done for these
launch partners, a lot of travel in there, and a
lot of personal assistant type things like a lot of food,
a lot of reservations, a lot of you that type
of stuff, that kind of personal assistant type.

Speaker 1 (05:09):
Yeah.

Speaker 2 (05:09):
Right, I read one other crazy little fact with chat
gpt this week. So at the moment, the company says
that it's being held back from expanding as quickly as
it wants because it simply can't get enough computing power. Right,
So it's got lots of ideas for different ways to
use their technology and to integrate it into different things
and to launch different products. But they just can't get

(05:29):
enough computing technology and computing grunt. So they are doing
this massive expansion. They're spending a trillion dollars, that's right,
a trillion dollars, one thousand billion US dollars and building
all of these data centers and buying these incredible computing chips.
But someone did the maths right on how much electricity

(05:50):
you would need in order to power all of the
different data centers that they're going to purchase with one
trillion dollars. Apparently, just to keep the lights on and
to power the computers, they're going to need at least
eight nuclear actors. Like it is just crazy how hungry

(06:11):
these things are. So yeah, it is going to be Hey,
do you want to know something else interesting? Yeah?

Speaker 3 (06:14):
Please? They can they can move markets. So they announced
this week that they were going to be looking at
potentially buying AMD or. There was some speculation I.

Speaker 2 (06:21):
Think they're going to be buying AIM which is which
is a chip manufacturer.

Speaker 3 (06:24):
Yeah, yeah, yeah, And stock went up instantly thirty eight
percent that day, Yeah, thirty eight percent that's amazing because
of the potential of a deal.

Speaker 2 (06:33):
Yeah, incredible.

Speaker 1 (06:34):
Wow.

Speaker 2 (06:34):
Hey, thank you so much, Paul. You take care. Watch
the space keenly as we always do. That's paused in
house our text, but spelling it out in simple terms
for us this morning, which we always appreciate. On the show.

Speaker 1 (06:46):
For more from Saturday Morning with Jack Tame, listen live
to news talks he'd be from nine am Saturday, or
follow the podcast on iHeartRadio.