September 5, 2024 • 51 mins
This introduction is written by an AI bot provided by the podcast broadcaster. The description is edited slightly but it saves an hour of work every time a new episode is published. Bob Morrell and Jeremy Blake unravel the intriguing evolution of artificial intelligence, from its mythological origins to today’s cutting-edge innovations. You'll learn how ancient legends, like medieval golems and modern cinematic portrayals like Ex Machina, have shaped humanity's understanding of AI. With humorous anecdotes and personal stories, your hosts simplify complex concepts such as reasoning, algorithms, and neural networks.
The journey through history highlights the contributions of early thinkers including Aristotle and Alan Turing, whose pioneering work laid the groundwork for AI as an academic discipline. Let's explore the advantages AI brings to business, from boosting productivity and improving user experiences to offering strategic competitive advantages. However, let's not shy away from discussing AI’s limitations, especially in capturing human nuance, emotion, and creativity. By referencing popular culture and sharing personal experiences, this podcast makes these sophisticated topics easily accessible and enjoyable.
Finally, let's find out how AI can transform business strategy, using real-world examples to illustrate its practical applications. From automating processes and decisions to enhancing marketing strategies and website improvements, AI's potential seems limitless. But there are potential pitfalls, including inherent biases and the challenge of mimicking human conditions like ADHD. Whether you’re a business leader or just AI-curious, this episode promises to equip you with invaluable insights to navigate the ever-changing landscape of AI from a layman's perspective, all while understanding its fascinating quirks and limitations.
Bob and Jeremy's new book, Whose Side Are You On? Disloyal Bonding and Strategic Lies, is released on 1st October 2024. Pre-order your copy now on Amazon.co.uk!
For more info, free resources, useful content & our blog posts, please visit realitytraining.com.
Reality Training - Selling Certainty
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:00):
Welcome to the next episode of the Reality of
Business.
I'm Bobby Murrell, one of yourguides through the ever-evolving
landscape of business, andjoining me today is the man who
once tried to program his coffeemaker to work out his tax bill.
Speaker 2 (00:14):
Jeremy Blake.
Hey, bobby, it almost workedtoo, until I ended up with a
refund in coffee beans.
Speaker 1 (00:22):
Today we're diving into the fascinating world of
artificial intelligence.
That's right, dear listeners.
Ai, the thing that powerseverything, from your phone's
autocorrect to that uncanny adthat knows you need new socks
before you do.
Speaker 2 (00:38):
AI is like the office assistant who never sleeps,
never takes a coffee break andwon't eat your lunch from the
fridge.
Well, usually.
Speaker 1 (00:49):
But like any good assistant, AI has its quirks.
Imagine an assistant whosometimes sends your emails in
Klingon or decides your meetingroom should be a broom closet.
Ai can be a bit unpredictable.
Speaker 2 (01:04):
Sounds about right.
So today we're going to takeyou on a journey through the
history of AI, its advantagesand disadvantages, and share
some examples, including ourfirst experience with ChatGPT.
Speaker 1 (01:18):
Exactly.
We'll start by tracing AI'sorigins, from its humble
beginnings to the powerful toolswe have today.
Then we'll explore the pros,like boosting sales and
streamlining operations, and thecons, like the occasional
tech-induced chaos.
Speaker 2 (01:34):
And trust me, folks, it's not all chat and robotic
overlords, Although I did onceask ChatGTP for a joke and it
told me to brace myself foralgorithmic humor.
Speaker 1 (01:46):
Which I've always thought your humor was, Jeremy.
If I'm honest, We'll also sharesome stories from our first
uses of ChatGPT, includingwriting this very introduction
Wow, we're doing the robot'sbidding by using it.
Speaker 2 (01:59):
So whether you're a tech geek ready to embrace our
robot future, or someone whostill uses a fax machine, yes,
stay tuned.
We promise to make AI as funand simple as explaining it to
your grandma or your grandpa.
Speaker 1 (02:16):
Grab your cup of coffee, sit back and let's
unravel the magic and madness ofAI in business.
By the end of this episode, youmight even be inspired to get
your toaster to start takingyour breakfast orders.
Speaker 2 (02:28):
Now, that's a future worth waiting for.
Let's get started.
Cue the music.
Speaker 1 (02:52):
Well, welcome to the reality of business.
And the more observant amongstyou will probably have worked
out that that introduction allabout this episode and AI was
written by ChatGPT.
So, for those of you who don'tknow, I'm Bobby Murrell and I'm
Jeremy Blake, and today we'regoing to focus in on AI and I
think in some ways it's a bit ofa layman's introduction to AI,
(03:14):
because we know lots oforganizations who are embracing
this new technology and runningwith it and there are some who
aren't touching it at all.
There are some who aren'ttouching it at all.
So we thought if we could do anepisode where we focused in on
what AI actually is and what aresome of its advantages and
disadvantages and what you canreally use it for, then that's
(03:34):
going to help everyoneunderstand it more and also
point out some of the thingsthey could be using it for.
And, of course, a lot of thesethings which you can find online
right now, like ChatGPT, areentirely free, and I think
that's the really fascinatingthing about it.
There are free bits of softwarewhich you can use right now.
(03:56):
That will really make adifference.
Free for now, yes, indeed, andI think as it becomes more
useful.
That's an interesting questionwhether people will start
charging for things like that.
There is, of course, a mountainof AI software that you can
purchase, use, install and runwith, and that's only going to
grow as time goes on.
Speaker 2 (04:16):
I mean, I should say, before we get into what your
discoveries and I'm excitedabout what you're going to share
For me, ai hit its peak in 2001.
And why do I say that?
Because, of course, film fansamongst you will know, spielberg
produced in 2001 that rathermarvellous film called
(04:37):
Artificial Intelligence, whichfocused on a child and a boy
being the embodiment of anartificial intelligent being
that could replicate the son ofthe professor played by William
Hurt which I love I find thefilm still very moving, but
maybe that will be where AI getsto.
It will become part of beingsthat we'll interact with.
Speaker 1 (05:00):
But well, that's, that's fascinating.
I'm really interested that youfound that out, because that
absolutely is the subject of thefirst part of of this podcast,
because I've done a bit ofresearch into the history of ai,
right.
So before I launch into that,how far back in history do you
(05:22):
think we need to go before wecome across a story that relates
to AI?
How many you know?
Where do you think we're goingto in history where we can start
really talking about AI as asubject?
Speaker 2 (05:34):
great question.
I'm going to plump, for Ireckon they're going to be
playing with it around thesecond world war period.
Speaker 1 (05:43):
Okay, that's a really good guess, and of course, the
20th century is where the modernversion of ai and the outputs
of that research begins.
But we're going back muchfurther, are we really?
Speaker 2 (05:59):
oh yeah, okay, give me a moment.
Okay, I'm tempted, because I'minterested in film, to go back
to it, having a play with filmand misinformation.
So I'm going to go for 1890.
Speaker 1 (06:14):
What's lovely about you, jeremy, is that you're very
much a sort of Victorian eraonward person.
I think so you're telling mewe're going back further.
Speaker 2 (06:22):
Oh yeah, oh way, way, way further.
But hang on a minute, itdepends on the definition.
Person, I think.
So you're telling me we'regoing back further.
Yeah, oh way, way, way further.
But hang on a minute, itdepends on the definition.
If you're talking about thecreation of artificial
intelligence or the, the, theperception, are you?
Speaker 1 (06:35):
talking about the perception of artificial
intelligence, okay, okay, so theterm beings beings from outer
space?
Not quite, but I suppose itcould have that connection.
I've looked at the history ofAI and, of course, ai is
artificial intelligence, or Ithink another way to think about
it is a robot, a mechanicalmachine that has the
(07:03):
intelligence of a human.
That's what we're looking athere, okay, and one of the
earliest examples of this isactually found in Greek
mythology.
We're going back 1500, 2000 BC,so we're going back about 4,000
years.
Okay, bearing in mind thatthere's only been writing for
(07:26):
about six and a half thousandyears, it's quite extraordinary
that you've suddenly got thisidea that we are thinking about
a metal humanoid character.
In this instance, it was calledTalos.
Talos was a giant constructedof bronze and he was a guardian
(07:47):
of the island of Crete, and thestory was that he would throw
boulders at invading ships andhe would walk around the entire
island three times a day toprotect it from invaders.
Okay, that's a greek myth of abronze statue that could think
(08:09):
and feel and do things underinstruction and did.
Speaker 2 (08:12):
Did in the story.
In the story because I knowquite a bit of greek mythology
but I don't know, that was hebuilt by man.
Speaker 1 (08:20):
Well, that's the interesting thing.
So Jason, jason and theArgonauts they were attacked by
Talos and to defeat him theyremoved a plug which was near
his foot and out of that hole inhis foot poured this stuff
called ichor, i-c-h-o-r ichor,which was also the blood of the
(08:43):
Greek gods, and that renderedthe bronze machine inanimate.
So the idea was that this ichorgave the metal structure
characteristics of a human being, because it was given by the
gods.
Speaker 2 (08:58):
But it had the power of the gods, so it was almost
like a half being, becausethere's also that in mythology
which they're playing with nowin lovely films for kids, about
children of half Greek God andhalf woman and all that stuff.
Ah, yes, okay.
Speaker 1 (09:13):
I'd never have guessed that.
Well, there we go into ancienttimes.
We've always had this idea thatwouldn't it be great if we
could create something that hada brain that could actually do
stuff that would be useful andhandy.
I mean, we don't need to do it.
Okay, now that is going to be atheme of this podcast.
(09:36):
What is it that ai does?
That takes away the task fromus.
It means we don't have to do it.
That is the fundamental thingwhich backs it all up.
I love the idea of jason andthe argonauts coming ashore
fighting this huge, great bronzething and then realizing, if
they just pull a plug from hisfoot, uh he, he doesn't move
(09:58):
anymore, but it, you know, it'sa bizarre thing to to consider
that was his achilles heel.
Uh, well, yes, he.
Speaker 2 (10:05):
That's a very great illusion, and only medieval
scholars or ancient scholars, oh, I think reference most of us
know about being dipped in thesticks marvelous so then we come
forward.
Speaker 1 (10:15):
So how much further do you think we're coming
forward from ancient greece?
Where do you think we're comingto now, the next stage in the
history of ai?
Speaker 2 (10:22):
I have to check your sources.
I mean, is this one website orone person?
Who's got off on this?
Speaker 1 (10:27):
I don't think you need to go down that road at all
.
I've done a very easy to findresearch.
Speaker 2 (10:31):
I like the idea of the Greeks.
But then you're saying there'sa sort of downtime where people
are having a rollies and notthinking about it, and then
we're coming back to a periodwhere it starts again.
Are you you want me to pick?
A time period where asignificant thing on a timeline
that is the embodiment of AI.
Speaker 1 (10:47):
Yes, I think we've got to stick with this idea of a
human functioning.
Okay, I'll tell you where I'mgoing Off the top of my head.
Speaker 2 (10:55):
I'm going to go to warfare, misdirection, that sort
of stuff.
But I'm trying to think of whatwar.
But I may be completely off.
Speaker 1 (11:06):
It war, but I may be completely off.
It's not that it's okay, it'smuch simpler.
So what this does, if you go tomedieval times, especially in
europe, we had the idea of agolem okay, not golem as in the
lord of the rings, but a golemg-o-l-e-m.
It's a germanic word what thatis is a being that's created
mechanically, and anotherversion of this is a word we
haven't used in this podcast fora long time a homunculus
(11:28):
beautiful, which is a beingcreated in a jar without
frankenstein.
Yeah absolutely frankenstein'smonster that was the next stage
is frankenstein frankenstein'sidea yes, the idea of creating
something which has the abilityto think, but not the emotional
connection.
You then bring it tofrankenstein.
(11:49):
So the whole point offrankenstein's monster is that
it is a man brought back to life, but because it's been
artificially created, it doesn'tunderstand love.
Speaker 2 (12:03):
However, as shelly wrote it does.
That's the twist.
It as Shelley wrote it does.
That's the twist.
It wants it, it craves it itcraves it.
Speaker 1 (12:08):
Yeah, absolutely right, and in fact, you
mentioned the film earlier.
There's another film which ismore recent than that, which you
must have seen.
Poor Things, no, no, I wasthinking of Ex Machina.
Speaker 2 (12:20):
Oh gosh, yes, Horrendous, horrendous.
Speaker 1 (12:23):
Absolutely.
Now, there you've got this ideaof this chap creating these
beautiful robots with artificialintelligence, and they're, you
know, the sort of perfectconstructions of a movable
machine, and yet there is notthat emotional connection.
Speaker 2 (12:42):
Yeah, I think it's a stereotype of lack of empathy,
isn't it?
It's a trope used in fiction aswell, and most recently I
didn't watch it, but, uh, someof my family watched it on our
family holiday on the plane.
Uh, poor things, which emmastone won an oscar for.
She is a reanimated being who'ssaved through inputting
(13:03):
artificial intelligence, lacksempathy and goes on these weird
quests.
I haven't seen it, so can'tfully comment okay.
Speaker 1 (13:08):
So then we sort of bring this idea in terms of what
ai really is.
It's actually a kind offormalization of reasoning.
So you have a thing which is amachine that you've created in
some form, and what you do isyou give it a sequence of
elements which allows it toreason on various subjects.
(13:31):
Now I've got some more on this.
So AI is based on the assumptionthat the process of human
thought can be mechanized.
Okay, now, interestingly, goingback a bit again, aristotle
actually considered the idea insome of his writings.
He thought that there was a waythat you could mechanize
thought.
Okay, because of the way thatit worked and there was a chap
(13:55):
called Ramon Lul, from Spain,1232 to 1315.
15 he actually started creatinglogical machines that combine
basic and undeniable truths witha simple logic.
Okay, and he had to be verycareful because in doing this it
(14:16):
was going against a lot ofreligious perspectives at the
time you can imagine thecatholic church wouldn't have
been very happy about that.
So but if you combine that idea,if here are some undeniable
truths, here is simple logic.
If we combine those, can thissequence of cogs or whatever it
may be come up with an output,an output exactly yeah.
So then you come slightlyfurther forward, 17th century.
(14:38):
You've got thomas hobbes andrenee descartes oh descartes
yeah I knew you, I knew you'dlove a french uh, yeah, you love
a bit of that.
And they came up with this ideathat all rational thought could
be made systematic with algebraand geometry.
So if you use algebra andgeometry as your basis, then
rational thought can be made tobe that too what's interesting
(15:00):
about those chaps who I know alittle about from my interest in
Florence's A-levels?
Speaker 2 (15:07):
quite weird.
The early philosophers dabbledmassively in maths and physics
because they believed that youcould only get reason in a
mathematical precision.
So they struggled to just befully philosophical without some
grounding in science.
Speaker 1 (15:23):
Well, that's Sir Isaac Newton, of course.
Yeah.
You know, that's that same kindof principle.
So we now arrive in the 20thcentury, and what you're getting
in the 20th century isformalizing maths so it can be
used to arrive at differentsolutions.
And, of course, one of thegreat friends of our podcast,
someone we've spoken about a fewtimes, alan Turing, was one of
(15:43):
the great friends of our podcast, someone we've spoken about a
few times.
Alan turing was one of themajor actors in all of this, and
in 1941 he was actuallythinking about this during the
war, working at bletchley park,which is a place that I know
we'd urge listeners to go andvisit.
He was the first person toseriously investigate the
(16:04):
theoretical possibility ofmachine intelligence.
Yeah, and that was in 41.
The actual academic disciplineof ai emerged in 1956, and it's
been an academic discipline eversince, but it wouldn't have
been anything without alanturing.
So did he coin machine learning, as they say that?
I haven't got that as a note.
(16:26):
He may well have done, but hewas definitely one of the major
proponents of this.
That's great.
So then we look at what are thedifferent elements that make up
AI.
Well, there are things likereasoning, which we mentioned.
How does a machine reasonsomething?
Well, of course, it usessomething that we all understand
from social media algorithms.
So algorithms work out whereyou're going to be, how you're
(16:47):
going to be there, why have youlooked at that product, why have
you looked at this?
And the internet is very goodat studying algorithms and
working out what we may or notbe interested in.
So then you have neuralnetworks.
So neural networks are thebrain and how one thought might
pass to another element toarrive at a point.
(17:09):
And what they did was createmachines with weights, knowing
that when a neural thought wentin a different direction, it
would use different levels andengage different amounts of
weight to then allow it to gofurther on.
And I like the idea of weightsbecause you and I know that when
(17:31):
decisions need to be made, weweight different elements of
that process in our brain.
Speaker 2 (17:36):
Yeah, it's good Waiting waiting.
Speaker 1 (17:38):
Absolutely.
Then of course, there's all thestuff around language.
So where does AI and languageconnect?
And of course, if you go toJapan, there's plenty of robots
which respond to language andunderstand language and get that
element of it.
Then of course, there is logicitself, and logic is very
(17:58):
difficult to program, and onegreat thinker made the point
that human beings understandlogic.
We know what it is butbizarrely, we rarely use it to
solve problems, and I thinkthat's really interesting.
If you look at major problemsin the world many of us you know
, you and I've had discussionsin the past about things that
(18:19):
happen and you go well, surelycan't they see that logically,
this must happen because blah,blah, blah.
Speaker 2 (18:26):
Well, that comes back to other podcasts and things
we've done where we use aheuristic, we use a shortcut to
determine that lots are based onprevious experience.
We won't necessarily use logiceach time We'll go.
I think it's one of these.
Speaker 1 (18:40):
Whereas what you have with AI is this idea that, look
, this is the logic, here areall the various neural networks
and the weights, here is thelanguage, here is the reasoning.
By using that, we can use it tosolve problems.
Now, the other thing which youcan do is program biases,
because this is something youand I do a lot on, which is
(19:02):
surely an AI solution won't bebiased.
But there's this kind of acaveat there, because if your AI
system is created by a human,which of course it is, then
quite often there is an inherentbias within that creation.
Speaker 2 (19:20):
Well, here's a question.
It might be coming a bit earlyand I know we've got sections.
Well, here's a question.
It might be coming a bit earlyand I know we've got sections,
so you don't just program it andit works.
Speaker 1 (19:49):
You create this network which allows the machine
to go?
Yes, let me take that questionand do something with it.
Speaker 2 (19:54):
It can make different connections that we may not
make Absolutely right.
Okay, so the information hasbeen there, but it might make
connections that a human mighttake longer to reach or may
never reach.
Speaker 1 (20:05):
Well, I think the speed is the thing which we'll
come on to.
But just to summarize thissection, in the last 20 years,
of course, we've now created AIapplications that are being used
.
For many years now.
They've been used behind thescenes in things like banking
and doing predictive stuff andthat sort of thing, and from
2011 onwards, ai has been usedto manage big data, learning,
(20:30):
trends and things like that, and, of course, we know all about
algorithms and how that allworks on the internet.
And then what you've actuallygot from 2022 onwards only a
couple of years is the emergenceof these large language models
like chat, gpt, which thenbecomes a user-friendly
(20:50):
interface for people to engagewith the technology.
So that is the history of AI.
It starts in ancient Greece, itcomes through the medieval era,
we have these really fantasticpeople researching this and
understanding the kind ofconcept of it, and then geniuses
like Alan Turing showing us thepathway to actually bringing
(21:13):
these different elementstogether and making it something
usable, absolutely fascinating,great, lovely start.
Speaker 2 (21:20):
Thank you very much.
So what we're?
Speaker 1 (21:21):
going to do is we're going to take a short break and
when we come back we're going tolook at what are the advantages
of AI, what are thedisadvantages of AI, and we're
going to share with you some ofour experiences of using ChatGPT
for the first time.
Speaker 3 (21:41):
You're listening to the reality of business brought
to you by Reality Training.
Selling Certainty.
We're a leading sales trainingand coaching company based in
the UK.
For information on how we helpour clients improve their
businesses, check out ourwebsite, realitytrainingcom.
Speaker 2 (22:02):
Welcome back.
Let's get into some more AI.
So, bobby, you want to putforward to me, for my brain as
the listener, the advantages anddisadvantages of AI.
Is that where we're going?
Speaker 1 (22:14):
now, absolutely Now.
Ai has a number of differentways that it gives you an
advantage.
So the broad headings for thesethere's broadly two things
operational effectiveness andoffering strategic competitive
advantage.
Okay, offering strategiccompetitive advantage okay, but
(22:34):
I would say that the fundamentalone thing that underpins
everything here is time, ismoney.
If you think about the time wespend as humans on tasks and
activities and all that sort ofthing, this is going to change
that, no question.
So one advantage it is aproductivity enab.
You can automate a businessprocess and improve the time it
(22:55):
takes.
Number one, no question.
Number two enhance a user andcustomer experience by speeding
up that process and making itquicker and more enjoyable.
What it can do, which I think isvery interesting, is that
sometimes you'll have businesseswhich have got problems that
they never seem to be able tosolve, and I think there's a
(23:18):
number of ai applications thatwill allow you to solve that
business problem permanently.
We've never been able to dosuch and such.
Click a button, that problem'snow solved, wow, okay, that's
significant.
And then, lastly, you canidentify and improve strategies
for competitive advantage usingAI to do that work for you.
(23:41):
So let's go into that inslightly more detail so AI can
deliver insights that drivevalue.
So it could say something likeyour website isn't very
user-friendly compared to others.
Speaker 2 (23:56):
Okay, well, that's worth knowing and just to check.
Does it do this because itchecks a number of websites and
makes that determination?
Speaker 1 (24:05):
I think it must learn what are the characteristics
that make something effective,and if you haven't got it, it's
going to tell you that.
Speaker 2 (24:12):
Perhaps because it's got other data of Google
rankings and it says you'relower ranked and so it makes
determinations yeah, okay.
Speaker 1 (24:19):
Now it will automate processes.
So that's something which, ifyou've got something which is a
manual thing that you've donefor years, ai can probably
automate it.
You won't have to do it anymore.
The interesting one is it canalso automate decisions.
Now I find that interestingthat you know we have things we
want to decide to do?
Speaker 2 (24:39):
do you mean order them?
What do you mean by automate?
Speaker 1 (24:42):
so what you can do is you can say every week we need
to decide whether we are goingto pay our payroll on a friday
at five o'clock based on thesechoices.
And a senior person might sitdown on a friday and look at all
the money coming in and themoney going out and all that
sort of thing and, and you know,worry about that.
(25:04):
But actually the ai canprobably go boom, boom, boom,
yeah, fine, okay, no, it's allthere, we off, we go.
So you can actually removeyourself from that task automate
decision making yes and the uigoes pay those people.
What you're actually saying isabdicate decision making.
Well, you are.
You're also thinking well, mytime is more valuable.
Yes to for evaluation, exactly.
(25:27):
So what you're doing is justtaking that task away from
yourself.
Now, what that does, of course,is increase, but I'm going to
add into that what you can alsodo is that you can use AI.
It can be trained to useunbiased data, which means that
what it will do when it makesdecisions, it will create more
(25:48):
equitable choices.
It will create more equitablechoices.
So, whereas you or I might lookat a decision and have a bias
in some form using that AIsystem, it will go.
no, this is the right decisionand this is a fair way of doing
it, and I think that's actuallyquite interesting.
Speaker 2 (26:03):
Well, that's very exciting because that means it
might ultimately enable peoplewho cannot make different
decisions and head downdifferent paths to have to make
them.
Or their positions becomeuntenable because they're unable
to listen.
Speaker 1 (26:19):
Absolutely.
That's an amazing difference, Ithink.
Now there's a few other things.
It's faster, so you'll getfaster results, and you and I
did a project a few years agowhere we asked a team of people
from the company to examinecompetitors web offerings.
Okay, we can now do that inseconds using ai.
(26:40):
No, we don't need to do thatanymore.
Okay, that's days and days anddays.
You can, of course, check it.
Speaker 2 (26:45):
Surely you'd probably want to, you know, you'd
probably want to check some ofit and have some visual
representation, because I thinkthe other point to make to
listeners from my very limitedexperience yours is more, of
course is you're sent backlanguage and text.
You're not given visuals.
Speaker 1 (27:00):
Well, that's just one thing we've looked at, jeremy.
There's lots of other softwareyou can get which I'm sure would
give you that Produces otherthings, absolutely.
So a few other little benefitsare AI operates 24-7.
So if it's doing a task for you, then you can get it doing that
all the time.
Now, one example of this is youcan have chatbots on your
(27:21):
website having conversationswith customers 24 hours a day,
seven days a week, so there isno downtime.
Now, that is a reallyinteresting usage, usage which I
think we should think about.
What it does do over time is itdecreases human error, because
it's not a human that's doing it, it is a machine that is doing
this as well.
Speaker 2 (27:41):
So you're not going to have human doesn't need a
break, a coffee, a walk, yogapilates.
Speaker 1 (27:47):
It doesn't get ill and it doesn't die.
Doesn't miss days, work, yeah,yeah yeah yeah, I've employed ai
.
What's it cost you?
Oh, it's cost me ten thousand ayear.
Mind you, the human was a bitmore.
Oh yeah, so that's quiteinteresting, and it diverts you
from repetitive tasks so you canconcentrate on more creative
tasks, and we'll come on to that.
(28:07):
That's a disadvantage in aminute.
The other thing it will do isit will acquire data for you and
analyze it for you, but the keything here is that AI for this
sort of thing is reliant on thequality of data.
You can use AI to have betterquality data, and it will spot
and fill in gaps for you.
It'll look at anomalies andinaccuracies and it'll fix them
(28:29):
too, but data quality suddenlybecomes very, very, very
essential.
So those are the mainadvantages of AI.
It's pretty good, isn't it?
Speaker 2 (28:38):
It is pretty good.
But I'm hoping and I'm suresome listeners who put things on
their Facebook page like I washoping I could be a poet and an
artist rather than AI, I washoping AI would clean the dishes
.
You know, you see quite a lotof that humor humorous but also
humorless comments being madeabout AI.
(28:59):
So I think for people who arestruggling, maybe concrete
disadvantages now rather thanjust emotional feelings.
Speaker 1 (29:08):
Well, okay.
Well, there's some keydisadvantages.
You know, if you're going to beusing this a lot for certain
solutions, you are going to havea costly implementation.
So if you've got a largeorganization and you think we
are going to bring in an AIsystem to do all this stuff for
us that human beings werepreviously doing, that's costly,
(29:30):
that's going to cost a lot ofmoney in lots of ways.
Now, another disadvantage isyou could have job losses here,
because some of these systemswill do the job of a human for
much less and, in some cases,much better.
There was an old woody allenjoke where he says my father
(29:52):
retired after 50 years doing thesame job uh, just, um, sitting
there, uh, making things in afactory uh, for 50 years.
And then they replaced himovernight with a machine that
did everything my father did,only did it much better.
The sad thing was the next daymy mother ran out and bought one
only did it much better.
The sad thing was the next daymy mother ran out and bought one
(30:14):
Nice punchline Because.
I didn't see that coming, that'snice.
Speaker 2 (30:19):
Well, I can just come in there.
There's been quite a lot ofpress in some of the North
American magazines like Inc andFast Company and stuff about
tills.
I don't know what the otherword is.
You know, when you go to asupermarket and you flip
yourself, they often fail.
A number of supermarkets in theUS removed people from running
checkouts and the evidence wasthat people who lived on their
(30:41):
own and were over 70 liked tochat to the people because they
met them, they knew them.
They removed that conversation.
That was very, very damaging,and so they actually went to
supermarkets that had people ontills to chat to.
So they reinstalled them.
So is that ai is?
Speaker 1 (30:59):
a till ai.
I suppose it must be to somedegree but, I think I would say
it's only ai, if it's a bit likeminority report, when, when you
walk out of the thing and itand it chooses that you've it
welcomes you back or yeah, youknow, does that sort of thing.
Now there, what we're gettingtowards here is one of the real
disadvantages.
Here there is a lack of emotionand creativity.
(31:19):
You can ask your ai system towrite something for you, to
create it for you, but itdoesn't have that yet.
It doesn't have that yet.
It doesn't have that humancreative element.
No, exactly, exactly, and Ithink whether this will change
or not again is another matter.
(31:39):
Right now, ai can't learn fromexperience or mistakes.
Oh, really, I think so.
They don't think it.
If a machine did the same jobfor 10 years, it wouldn't go
well.
What I've learned over the last10 years is I now know to do it
like that.
I don't think it would havethat consultative aspect applied
(32:01):
to it, and if it made a mistake, I think, unless you pointed
that out, it would continue tomake the mistake, because in the
end, it is a machine.
Still, it doesn't have a kindof personality that goes oh,
maybe I'll do it differently.
There are other disadvantages,but they're mostly around this
idea that it's expensive andit's also transitionally
(32:26):
emotionally tough because youmight have departments of people
who are paid to do somethingwhich the machine will do in
moments.
Speaker 2 (32:33):
Well, if we just come back into this section, so back
to Alan Turing again.
He invented a thing or it wasnamed in his honor called the
Turing test, which is, if amachine can convince a human
it's not a machine, then we'vereached a point where AI, then
we've reached a point where AIis equal to us.
(32:53):
So far, no one's passed theTuring test and some of the apps
that have tried to do it and Ithink of I won't mention who it
is, but a dear relative of minehas been on various dating apps
and he's discovered AI is in thedating apps.
Speaker 1 (33:07):
Oh, yeah, I found that out too.
Speaker 2 (33:09):
And you found that out too Well, this is my bro,
not that he listens, listens.
He might start listening untilI.
Speaker 1 (33:14):
I was on one last year where there was a woman who
her pictures were ai generatedright.
Speaker 2 (33:20):
Well, worse than that .
Simon was talking to a womanwho said she was in oxford and
he got a response about going.
I like doing lots of things.
It just sounded odd, so I saidask her about her favorite part
of oxford, because we both knowoxford extremely well.
If she could name ageographical spot.
We'd know what she went and shecame back saying I like both the
busy areas and the quiet areaswent right.
(33:42):
What's?
This so that was a company thattook money from you for the
amount of interactions you had,but the ai was crap, so that
failed the train test and hecame off that website brilliant.
Speaker 1 (33:53):
well, there we are, so, but then, if you think about
it, that's something asrelatively trite as a dating app
, which you know is utilizingthis technology, and it's an
interesting one because it'sit's also using our human
behaviors for advantage, and Ithink that's an interesting
point.
Speaker 2 (34:13):
Well, that's getting very interesting.
So I think of films.
So there was the other filmAgain.
Often they're short stories byL Ron Hubbard or they're short
stories by Clark.
A lot of the AI work in fictionis now getting closer, but
there's a Willith film wherethese droids come out.
Do you know the film I mean?
And the droids create anupgrade of themselves.
(34:37):
An upgrade of the droids iscreated like these robots, okay,
and they try to kill all theother robots and then take over
and the head droid convincessomeone that this is all
masterminded by a human.
But it's not.
It's masterminded by the droids.
That's a great film with WillSmith.
I have to search it justbecause listeners might want to
(34:59):
see it.
Speaker 1 (34:59):
Well, of course you know, if you look at the
Terminator, it's all about that.
You know the fact that theworld is taken over by computers
and tries to wipe out humanity.
I mean, that's the worst casescenario.
Speaker 2 (35:11):
It's called iRobot and that's probably a short
story.
Isaac Asimov, that's who I'mthinking of.
It's based on a book by Asimovwhich he wrote in 1950.
You see, you talk about yourtimeline.
Asimov was born in Russia in1920.
By 1950, he's creating thesesort of AI envisaging books.
(35:32):
The other author who theycategorize him as?
Robert A Heinlein, of course.
So Arthur C Clarke, heinlein,asimov, you know very scary
science fiction that's gettingcloser to the truth.
So any more disadvantages fromyou, or does that cover that?
Speaker 1 (35:50):
No only except to say there was a very big report
done on AI in 2021 called theAppen State of AI Report, and
the conclusion there was thatall businesses or commercial
businesses have a critical needto adopt AI or risk being left
behind.
Being left behind, and now,having done some experimentation
(36:11):
with some of the AI stuff, I'mcoming around to that feeling as
well, actually.
Speaker 2 (36:17):
What does GPT stand for?
Do we know that?
Speaker 1 (36:19):
No idea, but I do know that it is an extremely
interesting website.
So give me what you got.
Give me what you got.
The first thing I'll say aboutit is it is free, and that's the
thing I can't quite get my headaround that this service is
entirely free, because if youwere to pay a company to do some
(36:40):
of the questions I'm about totell you about, they'd charge
you thousands for this kind ofstuff.
So that's the first thing thisis a free thing.
So I went on this for the firsttime and I asked it first of all
to assess the appeal andeffectiveness of our reality
training website.
Okay, and it did it pretty well.
(37:03):
Now I would say that some of itwas a little bit by rote, so
color, visual stimulation,pictures, calls to action, all
that sort of stuff.
So it's assessed it on thecriteria that that most of we
would if we looked at a website.
But it did pick out some thingswhich we didn't have.
(37:23):
So I thought, well, that'sinteresting.
So then I thought I know whatI'll do.
I'll ask to compare our websiteto one of our competitors in
terms of commercial appeal andlead generation capabilities.
Okay, Now, that's a pretty tallorder If I asked you to do that
, jeremy.
That would be a couple of hoursof your life.
Speaker 2 (37:44):
Well, it would also, based on what you've already
said it would have a bias.
I'd go.
Well, they think they're likethat, you know.
Yeah, I'd go.
Oh, that guy, he doesn't do anywork, he's a figurehead.
She's semi-retired.
She only works the summerholidays.
Speaker 1 (38:03):
Brilliant.
Speaker 2 (38:03):
They're owned by a massive company.
I would chuck in a load ofbiases, even though I'd tell you
I hadn't.
There we are, yeah.
Speaker 1 (38:10):
Good point, good point.
Speaker 2 (38:11):
That's good.
Speaker 1 (38:12):
So then I went broader and I asked the question
how could AI deliver salestraining to companies?
Now, there's lots of ways thatAI could deliver sales training
to companies, but then I askedwhat are the disadvantages of
this?
And of course, that's when itcomes back to this thing we
talked about earlier the idea ofnuance and emotion and
(38:35):
creativity and thinking upthings on the fly and coming up
with responsive language quicklyand naturally.
There is always going to be agap on that.
When you're using a machine, itjust it doesn't have that
reflexive way that the peopletalk currently sorted well, an
ai machine doesn't have adhd no,it doesn't, but it could
(38:58):
probably be programmed.
Programmed to have it, itprobably could.
So then I asked it how could weadapt our site so we compare
more favorably to our competitor?
Speaker 2 (39:10):
and it came up with some really good ideas that I
hadn't thought and not that youneed to say it aloud, but you
can matter on the screen.
You picked one competitor, didyou?
Speaker 1 (39:17):
yes, I did someone we've lost some work to yes,
okay, and I thought it wasreally interesting that, whilst
we're very different companies,there are things that we could
definitely do.
That would certainly show anadvantage over them if we were
to do them.
Speaker 2 (39:34):
But I'm just wondering if AI does the
comparison and makes us a bitmore like them.
Speaker 1 (39:39):
That's a very interesting question, but my
question was how do we adapt oursites so we compare more
favorably To them?
Speaker 2 (39:47):
in the eyes of a buyer.
So it's given us a couple ofthings that this company doesn't
do, that we should do, and theydon't do either Okay.
Speaker 1 (39:55):
Absolutely.
So that was quite interesting.
Then it said as a singlequestion back to me would you
like a detailed plan of how tomake these changes to your site?
So I said, yes, please, that'sall I said yes please.
So I said yes, please.
That's what I said yes, please.
Massive time-bound series oftasks of what we would create,
(40:16):
do change and how quickly we canchange our website.
Enough marketing for months,months, really, well or less.
I mean, you could just followtheir timeline and in two to
three weeks you've got animproved and brilliant website.
Speaker 2 (40:31):
Got to get people to find it.
Got to get people to click.
It doesn't end.
Does it marketing?
No, it doesn't end.
Speaker 1 (40:36):
It doesn't end.
But you see, one of the thingsit suggested that we do.
Rather than us do it, I couldask the chat GPT to do it for us
.
Do what bit?
Sorry, I could get it to writesomething for us.
Oh, I see.
Now let me tell the listenerssomething.
So at the beginning of thisepisode we read an introduction,
a scripted introduction thatI'd asked ChatGPT to create.
(41:00):
I said we're doing this podcastall about the advantages and
disadvantages of AI.
I gave it a few ideas of whatwe were going to be doing.
I gave it a few hints as towhat type of people we were, and
then I asked it to write theintroduction, which it did.
It took four seconds to writethe introduction, the scripted
elements.
I changed about three or fourbits of it to make it a bit more
(41:24):
like things that we would bemore likely to say, but that
task took three minutes.
Speaker 2 (41:29):
So can I?
I ask did it come up with thosecoffee jokes trying to?
Make it a right?
It actually thought a way toconnect to us would be to make
jokes about coffee beans andmachines and tax returns and
yeah, and all I said was it wasuh, light-hearted.
Speaker 1 (41:43):
I said it was going to be a podcast to introduce ai
in layman's terms, to people,and it took all that on board
and created that.
Now I don't think it'samazingly brilliant, but it's
certainly a different approach.
Speaker 2 (41:58):
Well, I don't think we'd expect AI to be a writer,
quite yet.
You know, novelists, goodwriters, still have a place,
whether they self-publish orthey are represented by book
houses, but I don't think peopleare asking AI to tell them
stories yet well, I think we'renot far off it, because we know
(42:19):
that people use AI in education.
Speaker 1 (42:21):
We know people use it for you can write books because
, especially if a factual book,if you've got a non-fiction book
about a subject, I think if youput in enough information about
that subject and some of thethings you were looking for, I
think ChatGPT will come up withsomething pretty damn close to
what you were looking for.
Speaker 2 (42:41):
So, based on what you've just said about writing a
factual book, martha and I lastnight and this is so middle
class, you'll love this we madetwo sticky toffee puddings.
Yes, as you know, now you'rejoking you'd love her to make
you one because you're coming tostay next week.
You'd love to have one beforeshe goes.
Martha and I researchedwebsites together, some of the
(43:05):
classic food recipe websites.
Her sticky toffee pudding isoff the charts.
Good.
I only helped a bit.
She did nearly all of itbecause I took a call halfway.
Could I get chat halfway?
Could I get ChatGTP to write mea cookery book?
But here's my question Would itonly access recipes from books
(43:25):
and information that's beengiven freely?
It wouldn't be able to go intoa published book that you could
only buy.
Speaker 1 (43:30):
I can't answer that.
I don't know the answer to thatquestion.
One thing I do know is, as youknow, I'm trying to sell some
copies of a book I've recentlywritten and I asked chat GPT
just to make sure we're clear onthat how I could sell more
copies of my book.
It gave me a few answers that Iknew, like advertising and
(43:56):
looking at search terms and thatkind of thing.
But it also suggested somedifferent routes that I could
use, based on the title of thebook, to promote it into
websites.
I'd never heard of thatactually review books when
they're launched and that sortof thing.
I had no idea about them.
So there's no doubt that it isconnected.
(44:18):
So there's no doubt that it isconnected.
Chat GPT is connected to everypossible outlet that you could
want on the internet and usesthose connections and I suppose
the trends and the usage ofthose things over years of data
gathering to make that point.
Speaker 2 (44:37):
So your book is about walking the Camino way.
Did it suggest Camino groups tosend it to, or just book review
sites that are generic?
Or did it say these are travelbooks?
How specific was it?
Speaker 1 (44:49):
It looked specifically at vloggers and
social media influencers, but italso went to general book
review sites to say stick it onhere and you'll get the thing
reviewed.
And it's followed by readers.
I don't even know what sitesthey are.
Speaker 2 (45:06):
You've written them down, I'll send them through.
I've got them Because there'sanother book coming out that
we'll need to do this with aswell.
Speaker 1 (45:11):
Exactly this is the key, because what you have here,
if you think about a Googlesearch, you might need to do 20
Google searches to find theinformation you need to create a
marketing plan for a book.
Okay, I think what ChatGPT doesis bring together that entire
task and, in seconds, will giveyou, at the very least, the
(45:36):
basis of that plan, but probablya number of very good, positive
actions that you can take thatwill make this thing work and
that is impressive.
Speaker 2 (45:45):
Tell you what's just gone off in my head.
Hearing this is De Bono's sixthinking hats might be a very
good method to distill it anddebate it once.
Chatgpt has done a lot of thework for you.
Speaker 1 (45:56):
I don't know.
Is that because you don't trustthe information?
Speaker 2 (45:59):
it's going to give you no.
No, it's the group of people.
So I'm thinking of a company,not our company.
I'm thinking of a company witha board yeah people are
reluctant to allow it.
It's agreed that we're going toreview a project where maybe
we're making a component.
I don't know, chat gpt does thework and they assembled it.
The human biases would come inagain against it.
(46:20):
So I'm thinking, if you saidright, what do we feel
emotionally about this?
What's a good idea that wecould build on?
Okay, if we all black hat someof it straight away, what's the
white hat, what's the missinginformation?
It just might be quite a goodway to disseminate.
That's in a way, you're right.
It's if you're not fullytrusting it.
Speaker 1 (46:36):
Yeah, it is, but I think also what's interesting
about that is that that's a bitlike jeff bezos amazon meetings.
This is your starting pointfrom which you build, but I
think what I then realized isthat I can have that as my
starting point and then I can go.
Do you know what?
If I ask it a slightlydifferent question, it will go
(46:57):
deeper.
Yes, and if I give it anotherquestion, it's going to go even
deeper.
So I don't need to keep onusing the starting point to do
more work.
I can actually keep asking thething questions until it gives
me everything that I need.
Speaker 2 (47:12):
Well, that's my very limited experience.
I asked about my middledaughter's university choices
and I asked them to compare twouniversities and it sort of was
quite I don't think it's rightto say sat on the fence, but it
got me to do more work.
It said well, you need to lookat the lecturers.
What are their papers?
What's the employment prospectslike?
What's their alumni?
Like you'll need.
(47:32):
It actually made me go gosh.
Yes, good point.
How can it be so simple?
If I then said who are thenotable alumni of x university
and who are the notable alumniof other, I could have compared
the alumni, but just as you'resaying that would have been me
asking it to do a secondaryquestion.
Speaker 1 (47:47):
So in a way informs you of the questions you've not
even asked it, which is prettyamazing well, I heard a great
thing the other day which wasexplained to me and it's similar
to what we've done, but they'vetaken it further.
So a gentleman said he askedchat gpt to assess his website
against 20 competitors and tocome back to him with ways that
(48:12):
he would have an advantage overall of those competitors if he
changed his offering.
Speaker 2 (48:17):
Okay, offering or design, or off.
Speaker 1 (48:20):
Actual everything.
Everything you know calls toaction, the whole lot.
And that comes back to you andthen you go, wow, now you can
either look at it and go, well,no one's worried about that.
But then you think, hang on aminute, what if one of those
competitors has done the exactsame thing and is currently
changing their website to havean advantage over me at the
moment?
Then it becomes important.
(48:40):
And, yes, we're at the earlystages of this, but it is
changing really, really quicklyand the interface with it is
easy, simple, so simple to use.
And what I think now we and allof our listeners must challenge
ourselves to do is think right,how can I make this work for my
(49:02):
organization, for my company,for my business?
Speaker 2 (49:05):
Even at an early stage, just experiment.
I tested something personalabout university choices.
I also tested about the companyJust start to question it and
build up ways of using it.
But this has been veryenlightening.
Speaker 1 (49:21):
Well, I think we'll leave this subject at this point
and we will return to thissubject, I'm sure, in future
episodes, because it's going tobe a movable feast in the years
to come.
And I'd also like to name checkone other podcast which talks
about this subject.
It's the rest is politicsleaders, and they recently
(49:43):
interviewed nick clegg, formerlythe head of the lib dems and
deputy prime minister, and hewas now, of course, he's a
global president of facebookmeta, and he talks all about
this, all about ai, what therisks are and what the future is
and how exciting it is, andalso, interestingly, about the
type of people who work inSilicon Valley on this sort of
(50:06):
thing.
And if you're interested inthis subject, you could do a lot
worse than listen to thatexcellent interview on the Rest
Is Politics Leaders.
Speaker 2 (50:14):
Well, thank you for some AI discovery.
Thank you for tuning in.
If this is your first listen,you'll have joined us today and
thank you so much.
We will be producing more.
If you've already been alistener to Bob and Jeremy's
Conf Lab, you'll have noticedwe've become the reality of
business and nothing changes.
We are your hosts.
We just thought it was a morereflective name than what we had
(50:38):
previously.
So thank you ever so much,bobby.
We'll be back with you soonwith a new episode okay, take
care.
Welcome to the next episode of the Reality of
Business.
I'm Bobby Murrell, one of yourguides through the ever-evolving
landscape of business, andjoining me today is the man who
once tried to program his coffeemaker to work out his tax bill.
Speaker 2 (00:14):
Jeremy Blake.
Hey, bobby, it almost workedtoo, until I ended up with a
refund in coffee beans.
Speaker 1 (00:22):
Today we're diving into the fascinating world of
artificial intelligence.
That's right, dear listeners.
Ai, the thing that powerseverything, from your phone's
autocorrect to that uncanny adthat knows you need new socks
before you do.
Speaker 2 (00:38):
AI is like the office assistant who never sleeps,
never takes a coffee break andwon't eat your lunch from the
fridge.
Well, usually.
Speaker 1 (00:49):
But like any good assistant, AI has its quirks.
Imagine an assistant whosometimes sends your emails in
Klingon or decides your meetingroom should be a broom closet.
Ai can be a bit unpredictable.
Speaker 2 (01:04):
Sounds about right.
So today we're going to takeyou on a journey through the
history of AI, its advantagesand disadvantages, and share
some examples, including ourfirst experience with ChatGPT.
Speaker 1 (01:18):
Exactly.
We'll start by tracing AI'sorigins, from its humble
beginnings to the powerful toolswe have today.
Then we'll explore the pros,like boosting sales and
streamlining operations, and thecons, like the occasional
tech-induced chaos.
Speaker 2 (01:34):
And trust me, folks, it's not all chat and robotic
overlords, Although I did onceask ChatGTP for a joke and it
told me to brace myself foralgorithmic humor.
Speaker 1 (01:46):
Which I've always thought your humor was, Jeremy.
If I'm honest, We'll also sharesome stories from our first
uses of ChatGPT, includingwriting this very introduction
Wow, we're doing the robot'sbidding by using it.
Speaker 2 (01:59):
So whether you're a tech geek ready to embrace our
robot future, or someone whostill uses a fax machine, yes,
stay tuned.
We promise to make AI as funand simple as explaining it to
your grandma or your grandpa.
Speaker 1 (02:16):
Grab your cup of coffee, sit back and let's
unravel the magic and madness ofAI in business.
By the end of this episode, youmight even be inspired to get
your toaster to start takingyour breakfast orders.
Speaker 2 (02:28):
Now, that's a future worth waiting for.
Let's get started.
Cue the music.
Speaker 1 (02:52):
Well, welcome to the reality of business.
And the more observant amongstyou will probably have worked
out that that introduction allabout this episode and AI was
written by ChatGPT.
So, for those of you who don'tknow, I'm Bobby Murrell and I'm
Jeremy Blake, and today we'regoing to focus in on AI and I
think in some ways it's a bit ofa layman's introduction to AI,
(03:14):
because we know lots oforganizations who are embracing
this new technology and runningwith it and there are some who
aren't touching it at all.
There are some who aren'ttouching it at all.
So we thought if we could do anepisode where we focused in on
what AI actually is and what aresome of its advantages and
disadvantages and what you canreally use it for, then that's
(03:34):
going to help everyoneunderstand it more and also
point out some of the thingsthey could be using it for.
And, of course, a lot of thesethings which you can find online
right now, like ChatGPT, areentirely free, and I think
that's the really fascinatingthing about it.
There are free bits of softwarewhich you can use right now.
(03:56):
That will really make adifference.
Free for now, yes, indeed, andI think as it becomes more
useful.
That's an interesting questionwhether people will start
charging for things like that.
There is, of course, a mountainof AI software that you can
purchase, use, install and runwith, and that's only going to
grow as time goes on.
Speaker 2 (04:16):
I mean, I should say, before we get into what your
discoveries and I'm excitedabout what you're going to share
For me, ai hit its peak in 2001.
And why do I say that?
Because, of course, film fansamongst you will know, spielberg
produced in 2001 that rathermarvellous film called
(04:37):
Artificial Intelligence, whichfocused on a child and a boy
being the embodiment of anartificial intelligent being
that could replicate the son ofthe professor played by William
Hurt which I love I find thefilm still very moving, but
maybe that will be where AI getsto.
It will become part of beingsthat we'll interact with.
Speaker 1 (05:00):
But well, that's, that's fascinating.
I'm really interested that youfound that out, because that
absolutely is the subject of thefirst part of of this podcast,
because I've done a bit ofresearch into the history of ai,
right.
So before I launch into that,how far back in history do you
(05:22):
think we need to go before wecome across a story that relates
to AI?
How many you know?
Where do you think we're goingto in history where we can start
really talking about AI as asubject?
Speaker 2 (05:34):
great question.
I'm going to plump, for Ireckon they're going to be
playing with it around thesecond world war period.
Speaker 1 (05:43):
Okay, that's a really good guess, and of course, the
20th century is where the modernversion of ai and the outputs
of that research begins.
But we're going back muchfurther, are we really?
Speaker 2 (05:59):
oh yeah, okay, give me a moment.
Okay, I'm tempted, because I'minterested in film, to go back
to it, having a play with filmand misinformation.
So I'm going to go for 1890.
Speaker 1 (06:14):
What's lovely about you, jeremy, is that you're very
much a sort of Victorian eraonward person.
I think so you're telling mewe're going back further.
Speaker 2 (06:22):
Oh yeah, oh way, way, way further.
But hang on a minute, itdepends on the definition.
Person, I think.
So you're telling me we'regoing back further.
Yeah, oh way, way, way further.
But hang on a minute, itdepends on the definition.
If you're talking about thecreation of artificial
intelligence or the, the, theperception, are you?
Speaker 1 (06:35):
talking about the perception of artificial
intelligence, okay, okay, so theterm beings beings from outer
space?
Not quite, but I suppose itcould have that connection.
I've looked at the history ofAI and, of course, ai is
artificial intelligence, or Ithink another way to think about
it is a robot, a mechanicalmachine that has the
(07:03):
intelligence of a human.
That's what we're looking athere, okay, and one of the
earliest examples of this isactually found in Greek
mythology.
We're going back 1500, 2000 BC,so we're going back about 4,000
years.
Okay, bearing in mind thatthere's only been writing for
(07:26):
about six and a half thousandyears, it's quite extraordinary
that you've suddenly got thisidea that we are thinking about
a metal humanoid character.
In this instance, it was calledTalos.
Talos was a giant constructedof bronze and he was a guardian
(07:47):
of the island of Crete, and thestory was that he would throw
boulders at invading ships andhe would walk around the entire
island three times a day toprotect it from invaders.
Okay, that's a greek myth of abronze statue that could think
(08:09):
and feel and do things underinstruction and did.
Speaker 2 (08:12):
Did in the story.
In the story because I knowquite a bit of greek mythology
but I don't know, that was hebuilt by man.
Speaker 1 (08:20):
Well, that's the interesting thing.
So Jason, jason and theArgonauts they were attacked by
Talos and to defeat him theyremoved a plug which was near
his foot and out of that hole inhis foot poured this stuff
called ichor, i-c-h-o-r ichor,which was also the blood of the
(08:43):
Greek gods, and that renderedthe bronze machine inanimate.
So the idea was that this ichorgave the metal structure
characteristics of a human being, because it was given by the
gods.
Speaker 2 (08:58):
But it had the power of the gods, so it was almost
like a half being, becausethere's also that in mythology
which they're playing with nowin lovely films for kids, about
children of half Greek God andhalf woman and all that stuff.
Ah, yes, okay.
Speaker 1 (09:13):
I'd never have guessed that.
Well, there we go into ancienttimes.
We've always had this idea thatwouldn't it be great if we
could create something that hada brain that could actually do
stuff that would be useful andhandy.
I mean, we don't need to do it.
Okay, now that is going to be atheme of this podcast.
(09:36):
What is it that ai does?
That takes away the task fromus.
It means we don't have to do it.
That is the fundamental thingwhich backs it all up.
I love the idea of jason andthe argonauts coming ashore
fighting this huge, great bronzething and then realizing, if
they just pull a plug from hisfoot, uh he, he doesn't move
(09:58):
anymore, but it, you know, it'sa bizarre thing to to consider
that was his achilles heel.
Uh, well, yes, he.
Speaker 2 (10:05):
That's a very great illusion, and only medieval
scholars or ancient scholars, oh, I think reference most of us
know about being dipped in thesticks marvelous so then we come
forward.
Speaker 1 (10:15):
So how much further do you think we're coming
forward from ancient greece?
Where do you think we're comingto now, the next stage in the
history of ai?
Speaker 2 (10:22):
I have to check your sources.
I mean, is this one website orone person?
Who's got off on this?
Speaker 1 (10:27):
I don't think you need to go down that road at all
.
I've done a very easy to findresearch.
Speaker 2 (10:31):
I like the idea of the Greeks.
But then you're saying there'sa sort of downtime where people
are having a rollies and notthinking about it, and then
we're coming back to a periodwhere it starts again.
Are you you want me to pick?
A time period where asignificant thing on a timeline
that is the embodiment of AI.
Speaker 1 (10:47):
Yes, I think we've got to stick with this idea of a
human functioning.
Okay, I'll tell you where I'mgoing Off the top of my head.
Speaker 2 (10:55):
I'm going to go to warfare, misdirection, that sort
of stuff.
But I'm trying to think of whatwar.
But I may be completely off.
Speaker 1 (11:06):
It war, but I may be completely off.
It's not that it's okay, it'smuch simpler.
So what this does, if you go tomedieval times, especially in
europe, we had the idea of agolem okay, not golem as in the
lord of the rings, but a golemg-o-l-e-m.
It's a germanic word what thatis is a being that's created
mechanically, and anotherversion of this is a word we
haven't used in this podcast fora long time a homunculus
(11:28):
beautiful, which is a beingcreated in a jar without
frankenstein.
Yeah absolutely frankenstein'smonster that was the next stage
is frankenstein frankenstein'sidea yes, the idea of creating
something which has the abilityto think, but not the emotional
connection.
You then bring it tofrankenstein.
(11:49):
So the whole point offrankenstein's monster is that
it is a man brought back to life, but because it's been
artificially created, it doesn'tunderstand love.
Speaker 2 (12:03):
However, as shelly wrote it does.
That's the twist.
It as Shelley wrote it does.
That's the twist.
It wants it, it craves it itcraves it.
Speaker 1 (12:08):
Yeah, absolutely right, and in fact, you
mentioned the film earlier.
There's another film which ismore recent than that, which you
must have seen.
Poor Things, no, no, I wasthinking of Ex Machina.
Speaker 2 (12:20):
Oh gosh, yes, Horrendous, horrendous.
Speaker 1 (12:23):
Absolutely.
Now, there you've got this ideaof this chap creating these
beautiful robots with artificialintelligence, and they're, you
know, the sort of perfectconstructions of a movable
machine, and yet there is notthat emotional connection.
Speaker 2 (12:42):
Yeah, I think it's a stereotype of lack of empathy,
isn't it?
It's a trope used in fiction aswell, and most recently I
didn't watch it, but, uh, someof my family watched it on our
family holiday on the plane.
Uh, poor things, which emmastone won an oscar for.
She is a reanimated being who'ssaved through inputting
(13:03):
artificial intelligence, lacksempathy and goes on these weird
quests.
I haven't seen it, so can'tfully comment okay.
Speaker 1 (13:08):
So then we sort of bring this idea in terms of what
ai really is.
It's actually a kind offormalization of reasoning.
So you have a thing which is amachine that you've created in
some form, and what you do isyou give it a sequence of
elements which allows it toreason on various subjects.
(13:31):
Now I've got some more on this.
So AI is based on the assumptionthat the process of human
thought can be mechanized.
Okay, now, interestingly, goingback a bit again, aristotle
actually considered the idea insome of his writings.
He thought that there was a waythat you could mechanize
thought.
Okay, because of the way thatit worked and there was a chap
(13:55):
called Ramon Lul, from Spain,1232 to 1315.
15 he actually started creatinglogical machines that combine
basic and undeniable truths witha simple logic.
Okay, and he had to be verycareful because in doing this it
(14:16):
was going against a lot ofreligious perspectives at the
time you can imagine thecatholic church wouldn't have
been very happy about that.
So but if you combine that idea,if here are some undeniable
truths, here is simple logic.
If we combine those, can thissequence of cogs or whatever it
may be come up with an output,an output exactly yeah.
So then you come slightlyfurther forward, 17th century.
(14:38):
You've got thomas hobbes andrenee descartes oh descartes
yeah I knew you, I knew you'dlove a french uh, yeah, you love
a bit of that.
And they came up with this ideathat all rational thought could
be made systematic with algebraand geometry.
So if you use algebra andgeometry as your basis, then
rational thought can be made tobe that too what's interesting
(15:00):
about those chaps who I know alittle about from my interest in
Florence's A-levels?
Speaker 2 (15:07):
quite weird.
The early philosophers dabbledmassively in maths and physics
because they believed that youcould only get reason in a
mathematical precision.
So they struggled to just befully philosophical without some
grounding in science.
Speaker 1 (15:23):
Well, that's Sir Isaac Newton, of course.
Yeah.
You know, that's that same kindof principle.
So we now arrive in the 20thcentury, and what you're getting
in the 20th century isformalizing maths so it can be
used to arrive at differentsolutions.
And, of course, one of thegreat friends of our podcast,
someone we've spoken about a fewtimes, alan Turing, was one of
(15:43):
the great friends of our podcast, someone we've spoken about a
few times.
Alan turing was one of themajor actors in all of this, and
in 1941 he was actuallythinking about this during the
war, working at bletchley park,which is a place that I know
we'd urge listeners to go andvisit.
He was the first person toseriously investigate the
(16:04):
theoretical possibility ofmachine intelligence.
Yeah, and that was in 41.
The actual academic disciplineof ai emerged in 1956, and it's
been an academic discipline eversince, but it wouldn't have
been anything without alanturing.
So did he coin machine learning, as they say that?
I haven't got that as a note.
(16:26):
He may well have done, but hewas definitely one of the major
proponents of this.
That's great.
So then we look at what are thedifferent elements that make up
AI.
Well, there are things likereasoning, which we mentioned.
How does a machine reasonsomething?
Well, of course, it usessomething that we all understand
from social media algorithms.
So algorithms work out whereyou're going to be, how you're
(16:47):
going to be there, why have youlooked at that product, why have
you looked at this?
And the internet is very goodat studying algorithms and
working out what we may or notbe interested in.
So then you have neuralnetworks.
So neural networks are thebrain and how one thought might
pass to another element toarrive at a point.
(17:09):
And what they did was createmachines with weights, knowing
that when a neural thought wentin a different direction, it
would use different levels andengage different amounts of
weight to then allow it to gofurther on.
And I like the idea of weightsbecause you and I know that when
(17:31):
decisions need to be made, weweight different elements of
that process in our brain.
Speaker 2 (17:36):
Yeah, it's good Waiting waiting.
Speaker 1 (17:38):
Absolutely.
Then of course, there's all thestuff around language.
So where does AI and languageconnect?
And of course, if you go toJapan, there's plenty of robots
which respond to language andunderstand language and get that
element of it.
Then of course, there is logicitself, and logic is very
(17:58):
difficult to program, and onegreat thinker made the point
that human beings understandlogic.
We know what it is butbizarrely, we rarely use it to
solve problems, and I thinkthat's really interesting.
If you look at major problemsin the world many of us you know
, you and I've had discussionsin the past about things that
(18:19):
happen and you go well, surelycan't they see that logically,
this must happen because blah,blah, blah.
Speaker 2 (18:26):
Well, that comes back to other podcasts and things
we've done where we use aheuristic, we use a shortcut to
determine that lots are based onprevious experience.
We won't necessarily use logiceach time We'll go.
I think it's one of these.
Speaker 1 (18:40):
Whereas what you have with AI is this idea that, look
, this is the logic, here areall the various neural networks
and the weights, here is thelanguage, here is the reasoning.
By using that, we can use it tosolve problems.
Now, the other thing which youcan do is program biases,
because this is something youand I do a lot on, which is
(19:02):
surely an AI solution won't bebiased.
But there's this kind of acaveat there, because if your AI
system is created by a human,which of course it is, then
quite often there is an inherentbias within that creation.
Speaker 2 (19:20):
Well, here's a question.
It might be coming a bit earlyand I know we've got sections.
Well, here's a question.
It might be coming a bit earlyand I know we've got sections,
so you don't just program it andit works.
Speaker 1 (19:49):
You create this network which allows the machine
to go?
Yes, let me take that questionand do something with it.
Speaker 2 (19:54):
It can make different connections that we may not
make Absolutely right.
Okay, so the information hasbeen there, but it might make
connections that a human mighttake longer to reach or may
never reach.
Speaker 1 (20:05):
Well, I think the speed is the thing which we'll
come on to.
But just to summarize thissection, in the last 20 years,
of course, we've now created AIapplications that are being used
.
For many years now.
They've been used behind thescenes in things like banking
and doing predictive stuff andthat sort of thing, and from
2011 onwards, ai has been usedto manage big data, learning,
(20:30):
trends and things like that, and, of course, we know all about
algorithms and how that allworks on the internet.
And then what you've actuallygot from 2022 onwards only a
couple of years is the emergenceof these large language models
like chat, gpt, which thenbecomes a user-friendly
(20:50):
interface for people to engagewith the technology.
So that is the history of AI.
It starts in ancient Greece, itcomes through the medieval era,
we have these really fantasticpeople researching this and
understanding the kind ofconcept of it, and then geniuses
like Alan Turing showing us thepathway to actually bringing
(21:13):
these different elementstogether and making it something
usable, absolutely fascinating,great, lovely start.
Speaker 2 (21:20):
Thank you very much.
So what we're?
Speaker 1 (21:21):
going to do is we're going to take a short break and
when we come back we're going tolook at what are the advantages
of AI, what are thedisadvantages of AI, and we're
going to share with you some ofour experiences of using ChatGPT
for the first time.
Speaker 3 (21:41):
You're listening to the reality of business brought
to you by Reality Training.
Selling Certainty.
We're a leading sales trainingand coaching company based in
the UK.
For information on how we helpour clients improve their
businesses, check out ourwebsite, realitytrainingcom.
Speaker 2 (22:02):
Welcome back.
Let's get into some more AI.
So, bobby, you want to putforward to me, for my brain as
the listener, the advantages anddisadvantages of AI.
Is that where we're going?
Speaker 1 (22:14):
now, absolutely Now.
Ai has a number of differentways that it gives you an
advantage.
So the broad headings for thesethere's broadly two things
operational effectiveness andoffering strategic competitive
advantage.
Okay, offering strategiccompetitive advantage okay, but
(22:34):
I would say that the fundamentalone thing that underpins
everything here is time, ismoney.
If you think about the time wespend as humans on tasks and
activities and all that sort ofthing, this is going to change
that, no question.
So one advantage it is aproductivity enab.
You can automate a businessprocess and improve the time it
(22:55):
takes.
Number one, no question.
Number two enhance a user andcustomer experience by speeding
up that process and making itquicker and more enjoyable.
What it can do, which I think isvery interesting, is that
sometimes you'll have businesseswhich have got problems that
they never seem to be able tosolve, and I think there's a
(23:18):
number of ai applications thatwill allow you to solve that
business problem permanently.
We've never been able to dosuch and such.
Click a button, that problem'snow solved, wow, okay, that's
significant.
And then, lastly, you canidentify and improve strategies
for competitive advantage usingAI to do that work for you.
(23:41):
So let's go into that inslightly more detail so AI can
deliver insights that drivevalue.
So it could say something likeyour website isn't very
user-friendly compared to others.
Speaker 2 (23:56):
Okay, well, that's worth knowing and just to check.
Does it do this because itchecks a number of websites and
makes that determination?
Speaker 1 (24:05):
I think it must learn what are the characteristics
that make something effective,and if you haven't got it, it's
going to tell you that.
Speaker 2 (24:12):
Perhaps because it's got other data of Google
rankings and it says you'relower ranked and so it makes
determinations yeah, okay.
Speaker 1 (24:19):
Now it will automate processes.
So that's something which, ifyou've got something which is a
manual thing that you've donefor years, ai can probably
automate it.
You won't have to do it anymore.
The interesting one is it canalso automate decisions.
Now I find that interestingthat you know we have things we
want to decide to do?
Speaker 2 (24:39):
do you mean order them?
What do you mean by automate?
Speaker 1 (24:42):
so what you can do is you can say every week we need
to decide whether we are goingto pay our payroll on a friday
at five o'clock based on thesechoices.
And a senior person might sitdown on a friday and look at all
the money coming in and themoney going out and all that
sort of thing and, and you know,worry about that.
(25:04):
But actually the ai canprobably go boom, boom, boom,
yeah, fine, okay, no, it's allthere, we off, we go.
So you can actually removeyourself from that task automate
decision making yes and the uigoes pay those people.
What you're actually saying isabdicate decision making.
Well, you are.
You're also thinking well, mytime is more valuable.
Yes to for evaluation, exactly.
(25:27):
So what you're doing is justtaking that task away from
yourself.
Now, what that does, of course,is increase, but I'm going to
add into that what you can alsodo is that you can use AI.
It can be trained to useunbiased data, which means that
what it will do when it makesdecisions, it will create more
(25:48):
equitable choices.
It will create more equitablechoices.
So, whereas you or I might lookat a decision and have a bias
in some form using that AIsystem, it will go.
no, this is the right decisionand this is a fair way of doing
it, and I think that's actuallyquite interesting.
Speaker 2 (26:03):
Well, that's very exciting because that means it
might ultimately enable peoplewho cannot make different
decisions and head downdifferent paths to have to make
them.
Or their positions becomeuntenable because they're unable
to listen.
Speaker 1 (26:19):
Absolutely.
That's an amazing difference, Ithink.
Now there's a few other things.
It's faster, so you'll getfaster results, and you and I
did a project a few years agowhere we asked a team of people
from the company to examinecompetitors web offerings.
Okay, we can now do that inseconds using ai.
(26:40):
No, we don't need to do thatanymore.
Okay, that's days and days anddays.
You can, of course, check it.
Speaker 2 (26:45):
Surely you'd probably want to, you know, you'd
probably want to check some ofit and have some visual
representation, because I thinkthe other point to make to
listeners from my very limitedexperience yours is more, of
course is you're sent backlanguage and text.
You're not given visuals.
Speaker 1 (27:00):
Well, that's just one thing we've looked at, jeremy.
There's lots of other softwareyou can get which I'm sure would
give you that Produces otherthings, absolutely.
So a few other little benefitsare AI operates 24-7.
So if it's doing a task for you, then you can get it doing that
all the time.
Now, one example of this is youcan have chatbots on your
(27:21):
website having conversationswith customers 24 hours a day,
seven days a week, so there isno downtime.
Now, that is a reallyinteresting usage, usage which I
think we should think about.
What it does do over time is itdecreases human error, because
it's not a human that's doing it, it is a machine that is doing
this as well.
Speaker 2 (27:41):
So you're not going to have human doesn't need a
break, a coffee, a walk, yogapilates.
Speaker 1 (27:47):
It doesn't get ill and it doesn't die.
Doesn't miss days, work, yeah,yeah yeah yeah, I've employed ai
.
What's it cost you?
Oh, it's cost me ten thousand ayear.
Mind you, the human was a bitmore.
Oh yeah, so that's quiteinteresting, and it diverts you
from repetitive tasks so you canconcentrate on more creative
tasks, and we'll come on to that.
(28:07):
That's a disadvantage in aminute.
The other thing it will do isit will acquire data for you and
analyze it for you, but the keything here is that AI for this
sort of thing is reliant on thequality of data.
You can use AI to have betterquality data, and it will spot
and fill in gaps for you.
It'll look at anomalies andinaccuracies and it'll fix them
(28:29):
too, but data quality suddenlybecomes very, very, very
essential.
So those are the mainadvantages of AI.
It's pretty good, isn't it?
Speaker 2 (28:38):
It is pretty good.
But I'm hoping and I'm suresome listeners who put things on
their Facebook page like I washoping I could be a poet and an
artist rather than AI, I washoping AI would clean the dishes
.
You know, you see quite a lotof that humor humorous but also
humorless comments being madeabout AI.
(28:59):
So I think for people who arestruggling, maybe concrete
disadvantages now rather thanjust emotional feelings.
Speaker 1 (29:08):
Well, okay.
Well, there's some keydisadvantages.
You know, if you're going to beusing this a lot for certain
solutions, you are going to havea costly implementation.
So if you've got a largeorganization and you think we
are going to bring in an AIsystem to do all this stuff for
us that human beings werepreviously doing, that's costly,
(29:30):
that's going to cost a lot ofmoney in lots of ways.
Now, another disadvantage isyou could have job losses here,
because some of these systemswill do the job of a human for
much less and, in some cases,much better.
There was an old woody allenjoke where he says my father
(29:52):
retired after 50 years doing thesame job uh, just, um, sitting
there, uh, making things in afactory uh, for 50 years.
And then they replaced himovernight with a machine that
did everything my father did,only did it much better.
The sad thing was the next daymy mother ran out and bought one
only did it much better.
The sad thing was the next daymy mother ran out and bought one
(30:14):
Nice punchline Because.
I didn't see that coming, that'snice.
Speaker 2 (30:19):
Well, I can just come in there.
There's been quite a lot ofpress in some of the North
American magazines like Inc andFast Company and stuff about
tills.
I don't know what the otherword is.
You know, when you go to asupermarket and you flip
yourself, they often fail.
A number of supermarkets in theUS removed people from running
checkouts and the evidence wasthat people who lived on their
(30:41):
own and were over 70 liked tochat to the people because they
met them, they knew them.
They removed that conversation.
That was very, very damaging,and so they actually went to
supermarkets that had people ontills to chat to.
So they reinstalled them.
So is that ai is?
Speaker 1 (30:59):
a till ai.
I suppose it must be to somedegree but, I think I would say
it's only ai, if it's a bit likeminority report, when, when you
walk out of the thing and itand it chooses that you've it
welcomes you back or yeah, youknow, does that sort of thing.
Now there, what we're gettingtowards here is one of the real
disadvantages.
Here there is a lack of emotionand creativity.
(31:19):
You can ask your ai system towrite something for you, to
create it for you, but itdoesn't have that yet.
It doesn't have that yet.
It doesn't have that humancreative element.
No, exactly, exactly, and Ithink whether this will change
or not again is another matter.
(31:39):
Right now, ai can't learn fromexperience or mistakes.
Oh, really, I think so.
They don't think it.
If a machine did the same jobfor 10 years, it wouldn't go
well.
What I've learned over the last10 years is I now know to do it
like that.
I don't think it would havethat consultative aspect applied
(32:01):
to it, and if it made a mistake, I think, unless you pointed
that out, it would continue tomake the mistake, because in the
end, it is a machine.
Still, it doesn't have a kindof personality that goes oh,
maybe I'll do it differently.
There are other disadvantages,but they're mostly around this
idea that it's expensive andit's also transitionally
(32:26):
emotionally tough because youmight have departments of people
who are paid to do somethingwhich the machine will do in
moments.
Speaker 2 (32:33):
Well, if we just come back into this section, so back
to Alan Turing again.
He invented a thing or it wasnamed in his honor called the
Turing test, which is, if amachine can convince a human
it's not a machine, then we'vereached a point where AI, then
we've reached a point where AIis equal to us.
(32:53):
So far, no one's passed theTuring test and some of the apps
that have tried to do it and Ithink of I won't mention who it
is, but a dear relative of minehas been on various dating apps
and he's discovered AI is in thedating apps.
Speaker 1 (33:07):
Oh, yeah, I found that out too.
Speaker 2 (33:09):
And you found that out too Well, this is my bro,
not that he listens, listens.
He might start listening untilI.
Speaker 1 (33:14):
I was on one last year where there was a woman who
her pictures were ai generatedright.
Speaker 2 (33:20):
Well, worse than that .
Simon was talking to a womanwho said she was in oxford and
he got a response about going.
I like doing lots of things.
It just sounded odd, so I saidask her about her favorite part
of oxford, because we both knowoxford extremely well.
If she could name ageographical spot.
We'd know what she went and shecame back saying I like both the
busy areas and the quiet areaswent right.
(33:42):
What's?
This so that was a company thattook money from you for the
amount of interactions you had,but the ai was crap, so that
failed the train test and hecame off that website brilliant.
Speaker 1 (33:53):
well, there we are, so, but then, if you think about
it, that's something asrelatively trite as a dating app
, which you know is utilizingthis technology, and it's an
interesting one because it'sit's also using our human
behaviors for advantage, and Ithink that's an interesting
point.
Speaker 2 (34:13):
Well, that's getting very interesting.
So I think of films.
So there was the other filmAgain.
Often they're short stories byL Ron Hubbard or they're short
stories by Clark.
A lot of the AI work in fictionis now getting closer, but
there's a Willith film wherethese droids come out.
Do you know the film I mean?
And the droids create anupgrade of themselves.
(34:37):
An upgrade of the droids iscreated like these robots, okay,
and they try to kill all theother robots and then take over
and the head droid convincessomeone that this is all
masterminded by a human.
But it's not.
It's masterminded by the droids.
That's a great film with WillSmith.
I have to search it justbecause listeners might want to
(34:59):
see it.
Speaker 1 (34:59):
Well, of course you know, if you look at the
Terminator, it's all about that.
You know the fact that theworld is taken over by computers
and tries to wipe out humanity.
I mean, that's the worst casescenario.
Speaker 2 (35:11):
It's called iRobot and that's probably a short
story.
Isaac Asimov, that's who I'mthinking of.
It's based on a book by Asimovwhich he wrote in 1950.
You see, you talk about yourtimeline.
Asimov was born in Russia in1920.
By 1950, he's creating thesesort of AI envisaging books.
(35:32):
The other author who theycategorize him as?
Robert A Heinlein, of course.
So Arthur C Clarke, heinlein,asimov, you know very scary
science fiction that's gettingcloser to the truth.
So any more disadvantages fromyou, or does that cover that?
Speaker 1 (35:50):
No only except to say there was a very big report
done on AI in 2021 called theAppen State of AI Report, and
the conclusion there was thatall businesses or commercial
businesses have a critical needto adopt AI or risk being left
behind.
Being left behind, and now,having done some experimentation
(36:11):
with some of the AI stuff, I'mcoming around to that feeling as
well, actually.
Speaker 2 (36:17):
What does GPT stand for?
Do we know that?
Speaker 1 (36:19):
No idea, but I do know that it is an extremely
interesting website.
So give me what you got.
Give me what you got.
The first thing I'll say aboutit is it is free, and that's the
thing I can't quite get my headaround that this service is
entirely free, because if youwere to pay a company to do some
(36:40):
of the questions I'm about totell you about, they'd charge
you thousands for this kind ofstuff.
So that's the first thing thisis a free thing.
So I went on this for the firsttime and I asked it first of all
to assess the appeal andeffectiveness of our reality
training website.
Okay, and it did it pretty well.
(37:03):
Now I would say that some of itwas a little bit by rote, so
color, visual stimulation,pictures, calls to action, all
that sort of stuff.
So it's assessed it on thecriteria that that most of we
would if we looked at a website.
But it did pick out some thingswhich we didn't have.
(37:23):
So I thought, well, that'sinteresting.
So then I thought I know whatI'll do.
I'll ask to compare our websiteto one of our competitors in
terms of commercial appeal andlead generation capabilities.
Okay, Now, that's a pretty tallorder If I asked you to do that
, jeremy.
That would be a couple of hoursof your life.
Speaker 2 (37:44):
Well, it would also, based on what you've already
said it would have a bias.
I'd go.
Well, they think they're likethat, you know.
Yeah, I'd go.
Oh, that guy, he doesn't do anywork, he's a figurehead.
She's semi-retired.
She only works the summerholidays.
Speaker 1 (38:03):
Brilliant.
Speaker 2 (38:03):
They're owned by a massive company.
I would chuck in a load ofbiases, even though I'd tell you
I hadn't.
There we are, yeah.
Speaker 1 (38:10):
Good point, good point.
Speaker 2 (38:11):
That's good.
Speaker 1 (38:12):
So then I went broader and I asked the question
how could AI deliver salestraining to companies?
Now, there's lots of ways thatAI could deliver sales training
to companies, but then I askedwhat are the disadvantages of
this?
And of course, that's when itcomes back to this thing we
talked about earlier the idea ofnuance and emotion and
(38:35):
creativity and thinking upthings on the fly and coming up
with responsive language quicklyand naturally.
There is always going to be agap on that.
When you're using a machine, itjust it doesn't have that
reflexive way that the peopletalk currently sorted well, an
ai machine doesn't have adhd no,it doesn't, but it could
(38:58):
probably be programmed.
Programmed to have it, itprobably could.
So then I asked it how could weadapt our site so we compare
more favorably to our competitor?
Speaker 2 (39:10):
and it came up with some really good ideas that I
hadn't thought and not that youneed to say it aloud, but you
can matter on the screen.
You picked one competitor, didyou?
Speaker 1 (39:17):
yes, I did someone we've lost some work to yes,
okay, and I thought it wasreally interesting that, whilst
we're very different companies,there are things that we could
definitely do.
That would certainly show anadvantage over them if we were
to do them.
Speaker 2 (39:34):
But I'm just wondering if AI does the
comparison and makes us a bitmore like them.
Speaker 1 (39:39):
That's a very interesting question, but my
question was how do we adapt oursites so we compare more
favorably To them?
Speaker 2 (39:47):
in the eyes of a buyer.
So it's given us a couple ofthings that this company doesn't
do, that we should do, and theydon't do either Okay.
Speaker 1 (39:55):
Absolutely.
So that was quite interesting.
Then it said as a singlequestion back to me would you
like a detailed plan of how tomake these changes to your site?
So I said, yes, please, that'sall I said yes please.
So I said yes, please.
That's what I said yes, please.
Massive time-bound series oftasks of what we would create,
(40:16):
do change and how quickly we canchange our website.
Enough marketing for months,months, really, well or less.
I mean, you could just followtheir timeline and in two to
three weeks you've got animproved and brilliant website.
Speaker 2 (40:31):
Got to get people to find it.
Got to get people to click.
It doesn't end.
Does it marketing?
No, it doesn't end.
Speaker 1 (40:36):
It doesn't end.
But you see, one of the thingsit suggested that we do.
Rather than us do it, I couldask the chat GPT to do it for us
.
Do what bit?
Sorry, I could get it to writesomething for us.
Oh, I see.
Now let me tell the listenerssomething.
So at the beginning of thisepisode we read an introduction,
a scripted introduction thatI'd asked ChatGPT to create.
(41:00):
I said we're doing this podcastall about the advantages and
disadvantages of AI.
I gave it a few ideas of whatwe were going to be doing.
I gave it a few hints as towhat type of people we were, and
then I asked it to write theintroduction, which it did.
It took four seconds to writethe introduction, the scripted
elements.
I changed about three or fourbits of it to make it a bit more
(41:24):
like things that we would bemore likely to say, but that
task took three minutes.
Speaker 2 (41:29):
So can I?
I ask did it come up with thosecoffee jokes trying to?
Make it a right?
It actually thought a way toconnect to us would be to make
jokes about coffee beans andmachines and tax returns and
yeah, and all I said was it wasuh, light-hearted.
Speaker 1 (41:43):
I said it was going to be a podcast to introduce ai
in layman's terms, to people,and it took all that on board
and created that.
Now I don't think it'samazingly brilliant, but it's
certainly a different approach.
Speaker 2 (41:58):
Well, I don't think we'd expect AI to be a writer,
quite yet.
You know, novelists, goodwriters, still have a place,
whether they self-publish orthey are represented by book
houses, but I don't think peopleare asking AI to tell them
stories yet well, I think we'renot far off it, because we know
(42:19):
that people use AI in education.
Speaker 1 (42:21):
We know people use it for you can write books because
, especially if a factual book,if you've got a non-fiction book
about a subject, I think if youput in enough information about
that subject and some of thethings you were looking for, I
think ChatGPT will come up withsomething pretty damn close to
what you were looking for.
Speaker 2 (42:41):
So, based on what you've just said about writing a
factual book, martha and I lastnight and this is so middle
class, you'll love this we madetwo sticky toffee puddings.
Yes, as you know, now you'rejoking you'd love her to make
you one because you're coming tostay next week.
You'd love to have one beforeshe goes.
Martha and I researchedwebsites together, some of the
(43:05):
classic food recipe websites.
Her sticky toffee pudding isoff the charts.
Good.
I only helped a bit.
She did nearly all of itbecause I took a call halfway.
Could I get chat halfway?
Could I get ChatGTP to write mea cookery book?
But here's my question Would itonly access recipes from books
(43:25):
and information that's beengiven freely?
It wouldn't be able to go intoa published book that you could
only buy.
Speaker 1 (43:30):
I can't answer that.
I don't know the answer to thatquestion.
One thing I do know is, as youknow, I'm trying to sell some
copies of a book I've recentlywritten and I asked chat GPT
just to make sure we're clear onthat how I could sell more
copies of my book.
It gave me a few answers that Iknew, like advertising and
(43:56):
looking at search terms and thatkind of thing.
But it also suggested somedifferent routes that I could
use, based on the title of thebook, to promote it into
websites.
I'd never heard of thatactually review books when
they're launched and that sortof thing.
I had no idea about them.
So there's no doubt that it isconnected.
(44:18):
So there's no doubt that it isconnected.
Chat GPT is connected to everypossible outlet that you could
want on the internet and usesthose connections and I suppose
the trends and the usage ofthose things over years of data
gathering to make that point.
Speaker 2 (44:37):
So your book is about walking the Camino way.
Did it suggest Camino groups tosend it to, or just book review
sites that are generic?
Or did it say these are travelbooks?
How specific was it?
Speaker 1 (44:49):
It looked specifically at vloggers and
social media influencers, but italso went to general book
review sites to say stick it onhere and you'll get the thing
reviewed.
And it's followed by readers.
I don't even know what sitesthey are.
Speaker 2 (45:06):
You've written them down, I'll send them through.
I've got them Because there'sanother book coming out that
we'll need to do this with aswell.
Speaker 1 (45:11):
Exactly this is the key, because what you have here,
if you think about a Googlesearch, you might need to do 20
Google searches to find theinformation you need to create a
marketing plan for a book.
Okay, I think what ChatGPT doesis bring together that entire
task and, in seconds, will giveyou, at the very least, the
(45:36):
basis of that plan, but probablya number of very good, positive
actions that you can take thatwill make this thing work and
that is impressive.
Speaker 2 (45:45):
Tell you what's just gone off in my head.
Hearing this is De Bono's sixthinking hats might be a very
good method to distill it anddebate it once.
Chatgpt has done a lot of thework for you.
Speaker 1 (45:56):
I don't know.
Is that because you don't trustthe information?
Speaker 2 (45:59):
it's going to give you no.
No, it's the group of people.
So I'm thinking of a company,not our company.
I'm thinking of a company witha board yeah people are
reluctant to allow it.
It's agreed that we're going toreview a project where maybe
we're making a component.
I don't know, chat gpt does thework and they assembled it.
The human biases would come inagain against it.
(46:20):
So I'm thinking, if you saidright, what do we feel
emotionally about this?
What's a good idea that wecould build on?
Okay, if we all black hat someof it straight away, what's the
white hat, what's the missinginformation?
It just might be quite a goodway to disseminate.
That's in a way, you're right.
It's if you're not fullytrusting it.
Speaker 1 (46:36):
Yeah, it is, but I think also what's interesting
about that is that that's a bitlike jeff bezos amazon meetings.
This is your starting pointfrom which you build, but I
think what I then realized isthat I can have that as my
starting point and then I can go.
Do you know what?
If I ask it a slightlydifferent question, it will go
(46:57):
deeper.
Yes, and if I give it anotherquestion, it's going to go even
deeper.
So I don't need to keep onusing the starting point to do
more work.
I can actually keep asking thething questions until it gives
me everything that I need.
Speaker 2 (47:12):
Well, that's my very limited experience.
I asked about my middledaughter's university choices
and I asked them to compare twouniversities and it sort of was
quite I don't think it's rightto say sat on the fence, but it
got me to do more work.
It said well, you need to lookat the lecturers.
What are their papers?
What's the employment prospectslike?
What's their alumni?
Like you'll need.
(47:32):
It actually made me go gosh.
Yes, good point.
How can it be so simple?
If I then said who are thenotable alumni of x university
and who are the notable alumniof other, I could have compared
the alumni, but just as you'resaying that would have been me
asking it to do a secondaryquestion.
Speaker 1 (47:47):
So in a way informs you of the questions you've not
even asked it, which is prettyamazing well, I heard a great
thing the other day which wasexplained to me and it's similar
to what we've done, but they'vetaken it further.
So a gentleman said he askedchat gpt to assess his website
against 20 competitors and tocome back to him with ways that
(48:12):
he would have an advantage overall of those competitors if he
changed his offering.
Speaker 2 (48:17):
Okay, offering or design, or off.
Speaker 1 (48:20):
Actual everything.
Everything you know calls toaction, the whole lot.
And that comes back to you andthen you go, wow, now you can
either look at it and go, well,no one's worried about that.
But then you think, hang on aminute, what if one of those
competitors has done the exactsame thing and is currently
changing their website to havean advantage over me at the
moment?
Then it becomes important.
(48:40):
And, yes, we're at the earlystages of this, but it is
changing really, really quicklyand the interface with it is
easy, simple, so simple to use.
And what I think now we and allof our listeners must challenge
ourselves to do is think right,how can I make this work for my
(49:02):
organization, for my company,for my business?
Speaker 2 (49:05):
Even at an early stage, just experiment.
I tested something personalabout university choices.
I also tested about the companyJust start to question it and
build up ways of using it.
But this has been veryenlightening.
Speaker 1 (49:21):
Well, I think we'll leave this subject at this point
and we will return to thissubject, I'm sure, in future
episodes, because it's going tobe a movable feast in the years
to come.
And I'd also like to name checkone other podcast which talks
about this subject.
It's the rest is politicsleaders, and they recently
(49:43):
interviewed nick clegg, formerlythe head of the lib dems and
deputy prime minister, and hewas now, of course, he's a
global president of facebookmeta, and he talks all about
this, all about ai, what therisks are and what the future is
and how exciting it is, andalso, interestingly, about the
type of people who work inSilicon Valley on this sort of
(50:06):
thing.
And if you're interested inthis subject, you could do a lot
worse than listen to thatexcellent interview on the Rest
Is Politics Leaders.
Speaker 2 (50:14):
Well, thank you for some AI discovery.
Thank you for tuning in.
If this is your first listen,you'll have joined us today and
thank you so much.
We will be producing more.
If you've already been alistener to Bob and Jeremy's
Conf Lab, you'll have noticedwe've become the reality of
business and nothing changes.
We are your hosts.
We just thought it was a morereflective name than what we had
(50:38):
previously.
So thank you ever so much,bobby.
We'll be back with you soonwith a new episode okay, take
care.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!