Paul Stenhouse: ChatGPT-5, Emirates bans use of power banks on board

Chat GPT-5 is here – what does it mean? What does it do?  

They claim it's the best yet at creative writing and coding, and apparently just feels more human-like in its responses. It will also make information up 25% less than GPT-4. Importantly, they've trained the model to fail gracefully when posed with tasks that it cannot solve, which is better than it taking shortcuts or lying about finishing it. It is much better with health related questions too, according to its benchmarks.  

Users can pick from four preset personalities—Cynic, Robot, Listener, and Nerd.  

Something you'll notice if you've never paid for ChatGPT, or used another provider, is a new thing called "reasoning". This is effectively where the AI talks to itself a little, and spends time "thinking" through the question before answering.  

 

ChatGPT is huge  

They say they now have 700 million weekly active users of ChatGPT, 5 million paying business users, and 4 million developers utilizing the API. The company is not yet profitable. It plans to raise $40 billion this year and is on a pace to pull in revenue of $20 billion by year’s end.  

 

If you've got an Emirates flight later this year, there's a new rule about power banks 

Emirates will ban the use of power banks aboard its flights and only allow passengers to bring one such device onto the plane. The airline says the "significant growth" in customers using power banks has led to a rising number of lithium battery-related incidents during flight across the industry. The airline also wants the power banks stored somewhere they can be quickly removed in case they set fire. 

 

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.

Speaker 2 (00:11):
A'd be Jack, Did you mum write that review? Very droll? No?
That was Chat GPT five. It has just been released
by open Ai, their latest GPT model. Paul Stenhouse is
our text Burton. He's been looking into it, Hey, Paul.

Speaker 3 (00:28):
Yeah, it's good at creative writing, they say, Jack. I
tend to agree, wouldn't you.

Speaker 2 (00:31):
I just reckon it's maybe a bit over the top,
like it just needs to tone things down about in
that case, about seventy percent.

Speaker 3 (00:38):
They also say that they've managed to get it to
you know, you know, hallucinate less or make things up less.

Speaker 2 (00:44):
Okay, but I don't know.

Speaker 3 (00:46):
Based on what I just heard you read out, Job, I.

Speaker 2 (00:48):
Believe Yeah, I walked into that very good. You know. Well,
they say that it is a little less sycophantic than
some of the other GPT models, and I'm not convinced
that it is in my case necessarily either. But never mind.
So what does it mean that Chat GPT five is here?

Speaker 3 (01:04):
Yeah? I mean Sir Sam Autmin, who's the c of
open Ai, called it like a step change. He referred
to it as you know when your iPhone got the
retina display or your Samsung got one of their high
definition screens, and you looked at it and you went, WHOA,
this looks incredible. And then you looked back at your
old iPhone and thought, man, did I how did I

(01:24):
look at this? How did I see all these little
dots and pixels on the screen. He thinks that once
you've used chet GBT five you would never want to
go back to four or any other iteration. He says,
it's just better. They say that it feels a little
more human.

Speaker 1 (01:42):
They say it is.

Speaker 3 (01:43):
Apparently great at creative writing and also coding as well.
That two of the things that it's been kind of
trained to do and do well. But I think one
of the things that most importantly it's done, and I
don't know if what you read out says this or not,
but they've trained the model to fail, and fail graciously

(02:04):
because some of the issues with chet GPT four was
when you actually got it to start doing tasks and
it couldn't solve those tasks right, especially in the coding world.
It just took shortcuts. So if you were trying to
get it to return something from a database and it
couldn't figure out how to do it. It would just return
you the same value every time and hard code it,
which obviously isn't what you wanted it to do. It

(02:26):
also sometimes would lie about finishing things.

Speaker 2 (02:29):
Yeah right, okay, and so that's not good.

Speaker 3 (02:32):
So they've been really working to make sure that it
actually fails and when it can't do something, it tries
to say that it can't do something, yeah, which is
which is what you can probably want it to do.

Speaker 2 (02:42):
Right, Okay, that's good. I mean that sounds good. I
saw an example they were doing where they were they
were basically putting in text prompts into the AI bot
and asking it to make games, and it seemed to
be pretty good at that, which is, you know, which
is impressive. But whether or not, I suppose it's, you know,
that much better than the last gptwo model. Will wait
and see. So so check GPT though and open AI

(03:04):
probably still the leaders when it comes to check bots
by some significant margin, right Chap. GPT's absolutely huge at
the moment.

Speaker 3 (03:10):
Oh yeah, oh yeah, I mean they have. Then they
released the numbers this week seven hundred million weekly users,
which is just an astronomical number. Hespirts that's getting interesting
when you're if you're interested in the business aspect of this.
They have five million paying business users four million developers
utilizing the API wow, so that's a little bit of

(03:33):
an interesting ratio with five million paying customers seven hundred
million weekly users, so no surprise the company is not
yet profitable. But some other numbers they mention it plans
to raise from venture capitalists forty billion dollars this year,
so they obviously need the cash and they but they

(03:54):
are on pace to pull in revenues of twenty billion
dollars by the end of the year. So you have
to kind of do some maths. If they need forty
million forty billion, they're going to have twenty billion in revenue.
They must be spending a ton of money. And as
you go through and use these new models, especially this
new CHET GPT five, if you've never paid for Chat GPT,

(04:17):
you've probably never come across a thing called reasoning, which
basically is where the AI starts to talk to itself
a little bit, and so it's spends some time thinking.
But of course that actually makes it kind of more
expensive because it needs to go back and use more
tokens and use more computing time and all of these things.
So I can't imagine their costs are going down anytime soon.

(04:37):
But I think there's a lot of people who use
it daily weekly who can't imagine now a life without it.

Speaker 2 (04:45):
Yeah, yeah, too right. I mean I use it in
all sorts of you know, different parts of my life now.
It is amazing. It often just it just needs a
weird tweak from time to time, right, Like it's kind
of always eighty percent there in my experience and just
needs a wee kind of human you know component, just
but see I like that, Yeah, so do I.

Speaker 3 (05:04):
I like that it's not too perfect. No, like that
it kind of helps refine and then I massage.

Speaker 2 (05:09):
Yeah. Yeah, it's a good way to think about it. Hey,
thank you so much.

Speaker 3 (05:12):
Paul.

Speaker 2 (05:12):
Paul Steinhouse is our textbook.

Speaker 1 (05:14):
For more from Saturday Morning with Jack Tame, Listen live
to news Talks ed B from nine am Saturday, or
follow the podcast on iHeartRadio.