March 14, 2025 • 30 mins
What if the future of work includes team members who aren't human but highly capable AI agents?
In this episode of “What’s the BUZZ?—AI in Business,” host Andreas Welsch engages in a fascinating discussion with Dan Sodergren, a seasoned keynote speaker and tech futurist, on the swiftly changing landscape of AI agents and their implications for the workforce. Learn how AI can be more than just a tool but a collaborator, reshaping how businesses operate.
Dan shares his insights as they explore pivotal questions:
- What does that mean for the people in a business doing the work today if agents take over market research, operational tasks, etc.?
- CEOs are ordering their staff back to the office. At the same time, we see layoffs left and right to “prepare for AI”. Is that the future of work?
- How can people still make a living when they’re either out of a job or only have half a job?
- Will AI agents turn a good portion of the population into gig workers? And will we farm out that work to our agents as proxies?
- What makes you hopeful about that future with AI?
This episode is essential listening for anyone from business leaders to tech enthusiasts looking to navigate the AI revolution in a way that fosters growth and efficiency. Discover the balance between leveraging AI for productivity while retaining human decision-making capabilities.
Are you ready to revolutionize your approach to AI and ensure your business stands out in this new era? Tune in now to turn AI hype into concrete business outcomes!
Questions or suggestions? Send me a Text Message.
***********
Disclaimer: Views are the participants’ own and do not represent those of any participant’s past, present, or future employers. Participation in this event is independent of any potential business relationship (past, present, or future) between the participants or between their employers.
Level up your AI Leadership game with the AI Leadership Handbook:
https://www.aileadershiphandbook.com
More details:
https://www.intelligence-briefing.com
All episodes:
https://www.intelligence-briefing.com/podcast
Get a weekly thought-provoking post in your inbox:
https://www.intelligence-briefing.com/newsletter
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Andreas Welsch (00:00):
Today, we'll talk about agent AI and how it's
reshaping our jobs forever, andwho better to talk to about it
than someone who's activelyworking on that and thinking
about that, Dan Sodergren.
Hey, Dan, thank you so much forjoining.
Dan Sodergren (00:12):
Thanks very much for being here.
And just a quick one, I'mdefinitely Dan Sodergren, and
I'm not an AI nor an AI agent,nor a hologram or any of those
things.
ElevenLabs and HeyGen are doingbrilliant work in this sphere,
but promise, I promise you, thisis me okay.
Just to make it absolutelyclear.
Andreas Welsch (00:28):
And they are doing fantastic work.
I've been experimenting withthem too.
And my, my family and friends,they hardly can tell the
difference if they even can.
But I am real too.
So just to get that out therenow.
Hey Dan, why don't you tell us alittle bit about yourself, who
you are and what you do?
Dan Sodergren (00:45):
Yeah hi.
My name's Dan Sodergren.
I'm a keynote speaker, also acorporate trainer.
Have been for about 20 years andI've been for Ally more than a
decade, been talking aboutdigital transformation and also
AI before I suppose it wascalled artificial intelligence.
It was called a load of otherthings.
And I then had the joy of, wewere just joking off stage that
(01:05):
I had the joy of being called atech futurist.
But I'd rather be called abillionaire philanthropist.
Tech futurist just means I didstuff far too well in the
market.
I had a augmented realitycompany about 20 years ago and
did some pretty cool stuff, butI would've been much, I would've
much happier to have been a lotricher.
So this time I'm riding thiswave, so to speak, but I think
hopefully I've caught it at theright time.
(01:26):
I'm not talking about hologramsand things, but AI agents are a
part of this future of work.
And I wrote a couple of bookslast year, one called The Fifth
Industrial Revolution, which isall about this new thing between
me and you.
I thought we had another yearleft until we had these AI
agents, to be honest with you.
It hasn't necessarily caught meoff guard, but the rapid pace of
change, as has got quicker andquicker.
(01:49):
So yeah, it's even for us Techfuturist, this is a pretty
freaky, freaky time to.
Andreas Welsch (01:54):
That's for sure.
I remember when, what was it,two years ago, we started the
first agent frameworks BabyAGIand what they were called.
And then within a couple monthsor a couple quarters, everybody
and in tech vendors were jumpingon the bandwagon and said, Hey,
we've been doing AI and agentsall along.
And here we are trying to holdon and figure out what is
(02:16):
actually real and what does itmean?
Dan, what'd you say?
Should we play a little game tokick things off?
Dan Sodergren (02:22):
Go on then.
Yeah.
Yeah.
I'm slightly scared by thislittle game.
Okay.
Andreas Welsch (02:25):
Okay.
So let's see.
Here we go.
If AI were a vehicle, what wouldit be?
60 seconds on the clock.
Go!
Dan Sodergren (02:35):
If I was a vehicle, what would it be?
I would have to have lots ofdifferent ways of it being a
vehicle i'd.
I tell you what, it's a funnyone.
I could say something like asubmarine because you don't
really see it coming until it'stoo late.
So maybe a submarine.
I'm gonna go with, if it was asubmarine, I'm gonna go for
submarine.
Does that count as a vehicle?
Andreas Welsch (02:52):
I think it counts.
If I may overgeneralize I hear aBritish accent, so it might be
yellow submarine, something likethat.
Playing to the stereotypes, Ihope you forgive me.
Dan Sodergren (03:04):
Oh to be honest with you, I think the Beatles, I
think it links in, doesn't it?
Didn't the Beatles do an AI tunea bit?
Go and Paul McCartney did.
So there you go.
Yes.
Even yellow submarine people aredoing it right.
Andreas Welsch (03:15):
Great response.
Submarine.
I don't know if I could be stuckunderwater for three months or
something that.
That takes a lot, right?
Yeah.
Think of all the news that you'dbe missing and everything that'd
be happening in, in those threemonths.
Crazy.
Dan Sodergren (03:30):
I always wondered, so they don't have we
must probably get we diverseamount too much.
I wonder if they still have the,they don't have wifi either.
God, what a terrifying idea tobe under the water doing that
for that long.
You'd have to be a certain typeof person.
Andreas Welsch (03:42):
Let's, that's, yeah.
Good question.
I.
I know it.
20, 30 years ago there was nocommunication radio silence.
Oh
Dan Sodergren (03:50):
real.
Of course.
Yeah, of course.
Oh God.
Oh, then AI's nothing likesubmarines.
So that was a terrible thing tosay.
That was a terrible metaphor.
Andreas Welsch (03:57):
No.
But things happen in stealth andthen they come out and we're
surprised.
Dan Sodergren (04:03):
There you go.
Perfect.
Like agentic AI.
There you go.
Andreas Welsch (04:07):
Exactly.
We've been talking about thefuture of work for more than a
decade.
I remember those earlyconversations, 2015, 16 ish of
how awesome things were going tobe.
And not only would we all spendour days at the beach, while
some other things are doing thework for us but even in, in a
more serious context that.
(04:27):
We, we would see work being donedifferently.
And I've always perceived it asa change in how we organize work
and, how the present day worklooks like.
But I feel now with agents,we're turning that upside down
so quickly and, I'm justcuriouswhat you're seeing there, what
does it mean for people in abusiness doing this work?
(04:49):
If these agents can do marketresearch and can take on
operational tasks and a wholeother host of things, what does
it mean in, business?
Dan Sodergren (04:58):
Number one, I think it's most probably the old
classic, I used to start all myall my law essays with this,
with the John Spiro quote, whichis"Without definition there can
be no inquiry." Because youcan't really, one of the big
things we have in AI and I thinkin the tech world is that.
We presuppose that everybodyknows what we're talking about
and that everyone's talkingabout the same thing, right?
Yep.
So we are talking about AIagents for a very long time, but
(05:19):
then people say, oh, of coursewe've got an AI agent, it's a
chat bot, and it's ah, thatmight not be an AI agent, but I
can see where you're going withthat, right?
And you might say, technically,isn't ChatGPT an AI agent.
No, it's not and that's why thedefinition's important.
So if you define it much morearound we dunno if you agree
with me, but adaptability andbeing autonomous and those kind
(05:41):
of things is more important inthis kind of AI en gen world.
Only'cause I did a poll onLinkedIn, I think it's 22% of
people didn't know what an AIagent was.
So I was like, ah, I'm gonnatalk about it.
Let's at least define it.
So I think if you look to thatbeing autonomous and
adaptability but also being ableto do, almost doing an objective
(06:02):
rather than a task, if thatmakes sense.
As I used to say a bit of acliche thing, but jobs may
shift, but but skills uplift.
It is the same thing with AInow, then it's now got a new
skillset.
It's now got a new ability.
Yeah.
Yeah.
So now AI agents have thepotential of doing an objective,
a full objective, not quite, Idon't think quite a full job,
(06:23):
but a full objective, right?
Rather than just a task.
So they are very different tochat bots.
They are very different to GPT.
They are very different based.
And I don't know if you've seen,and you will have seen of
course,'cause you're an expertin it too the stuff that's
coming over from China withManus and all these other
things, and even with operatorand these other things it's now
astounding what AI agents cando.
(06:44):
And even though we talked aboutme coming on the show, I think
it was a couple of months ago,even in those couple of months,
it's fundamentally changed whatI would've said because now
we're at a place which I didn'tthink we'd be at for at least
six months to a year.
So now AI agents are in that.
We did, which is again, anothercliche, AI won't replace you,
but somebody using AI will,which like we used to say, it's
(07:06):
now actually AI might be usingitself to replace you.
And I think that's the thing.
So from a leadership point ofview, it's fundamentally
different.
I talked about this in the bookI wrote earlier this year, which
was the kind like, like a how toguide it was called, Leadership
in the Times of AI.
The point of the end of thatbook, and I was, had the joy of
(07:27):
writing with about 15 other AIand security experts.
The bit at the end of that bookwas talking about that how do we
as leaders.
We have to fundamentally changebecause we've now got potential
of people having, it's almostlike they're the their own
company now.
I used to joke that it was a bitlike, how do you manage the
Avengers?
They've all got superheroes andthey've all got superpowers,
right?
But now it's like, how do youmanage one person who
(07:49):
potentially is five people or ispotentially a company?
In their ability becauseeverybody now almost everybody
now, who has a good computer anda good computer understanding
and has been trained in thisstuff can start potentially
building AI agents, and that'sonly happened in the last couple
of weeks.
Now, that is fundamentallyterrifying from an employee
(08:13):
point of view, from an employerpoint of view, and an employee,
the potential of it, it's likenuclear power.
The potential of it isastounding.
It could save the world, but itcould also destroy the planet.
It's the same, it's not the sameall.
Andreas Welsch (08:28):
That's a pretty broad spectrum.
Dan Sodergren (08:30):
That's for sure.
Exactly.
Of outcomes.
But just like we talked aboutwith the submarine thing now
we're talking about nuclearsubmarines right now.
Andreas Welsch (08:37):
I see a question in the chat from Kerry, and he's
asking, what are your thoughtson transferring or sharing
decision making with models andhow we think about
accountability to ourstakeholders?
I think that fits nicely in,into the business discussion.
Dan Sodergren (08:49):
It's a fantastic question, and I think this is
the biggest thing now for notjust employees, but employers,
right?
It, a month ago it was what'sour AI policy?
Or now it's gonna have to bewhat's our AI agent policy?
It's actually what's ouremployee policy and actually
what's our employee contract.
And we have to, and by the way,I would love that governments
are doing this, but governmentsmove too slowly.
(09:11):
We are now in this world wherewe have to start to question
pretty much everything about ip,pretty much everything around
privacy, pretty much everythingaround anything that you now do.
Yeah.
So you have this big thing nowand I've gotta be careful with
words'cause this is getting abit too futuristic, but it's
only in a couple of months,right?
There will need to become aconversation about who owns the
agent.
(09:32):
Yeah.
If you've built an agent on acompany laptop with company
information, but it was youdoing the work to build the
agent who owns the outcome.
Yeah.
Yeah.
Now this is the nice thing aboutit is these agents will take the
grind.
Yeah.
And then human beings willmanage'em and refine.
So as I'm talking about, asindividuals, we've gotta become
(09:54):
a lot more mindful about what wedo, much more mindful about what
we do, and are actually morestrategic than ever.
Because actually what you aredoing now is gonna echo for a
generation, and you might feelthat silly, but the data that
you are gonna be producing inyour job is that your
intellectual property and thear, the old argument, of course,
is no, but it's weird, isn't it?
(10:15):
This the great question.
It's if I create a great promptat work, what's my incentive to
share the prompt?
Yeah.
And at the moment there's no wayyou don't go, oh, you get a
bonus for a thousand pounds'cause you did a great prompt.
But then we laugh a bit, right?
We laugh, but it should.
It should be when we all workedin factories and someone came up
with a great idea to make thefactory better, they got money
because they came up with agreat idea.
(10:37):
This is how it works.
Right now we've got this bitwhere we're talking about
knowledge.
There aren't the incentives inplace right now to really reward
people.
Now, if you are clever enough tostart coding an agent, which
does your job for you.
Are you still doing that job?
Now?
That's a big question.
Andreas Welsch (10:55):
I think you bring up an excellent point and
I haven't heard anybody elsearticulate it at all.
And definitely not in the sameway.
And that is if you are anemployee at a company, you come
in with a certain knowledge, yougain knowledge.
But once you leave, you takethat experience with you to
whatever other company orventure you, you go into to your
(11:16):
point, if you build an agent andthat agent stays with the
company, what happens to thatknowledge that either you have
transferred into it in buildingthat agent if you leave, if
you're an employee today, youleave there are no agents, then
your knowledge is gone.
It's no longer accessible forthe company beyond what you've
documented.
(11:36):
But if this agent continues todo work for the company, yeah.
You've built in and conceivedand fed with your insights and
your previous knowledge even.
Exactly, right?
Dan Sodergren (11:47):
And this is what we, no, and it is not quite the
same as when it was a humanbeing because, and I used to
joke, and I still think it'strue AI is just automated
interns.
That's what it should stand for,right?
It's, you've got an intern thatyou have to train and that's how
you get better at the AI gettingbetter.
It's not just promptengineering, it's managing the
AI right now with that momentwhen if your apprentice then
(12:08):
moves on and the apprenticebecomes the master.
Like, where's the situationthen?
As human beings, we've gottalearn to manage AI before the AI
manages us, and that's where weare right now.
Now, the question for everybodyis if the productivity increases
are huge, what happens to thefuture of work?
Because if an AI agent can dosome of your job and then starts
(12:30):
doing 50% of your job.
Is your job to manage the agent?
And the answer is yes.
By the way, I do not believethat we should take human beings
out this equation.
I do not believe they should besemi, they could be autonomous,
but they can't not be managed.
They must be managed with humanexperiences, human emotions.
Otherwise we, are in a hugeamount of danger, not just from
(12:50):
as ethically you know, thereason why, because of
hallucinations, et cetera.
But it's a big question we'veall gotta ask ourselves is.
We are managing the machinesthat will take on most of the
work, but we have to manage themachines.
Andreas Welsch (13:03):
That's right.
And, let me build on thatbecause I think we're seeing a
lot of these return to officemandates.
We're seeing CEOs ordering thestaff back.
At the same time we're seeinglayoffs and we need to prepare
for AI and trim our workforce,have more capital to, to invest.
Is that the future of work?
(13:23):
And to your point, right?
I was discussing that with JoeReis a couple weeks ago on the
show.
Will we all have multiple jobsbecause now agents do 50% here,
and we ideally maybe not have50% more time.
Do we need a second job to coverthe other 50% and the other 50%
of the income?
Dan Sodergren (13:43):
Again, in the book the Fifth Industrial
Revolution, I talk about thefour different intelligences
that we are gonna need.
We need artificial intelligencetaken as a given.
I think we need more emotionalintelligence.
Ironically enough, as the hardskills become diminished, the
soft skills go up.
We are gonna need to be morecooperative and nicer to each
other because the people whoaren't very nice will basically
don't have jobs because thepeople who are technically great
(14:04):
lawyers but aren't very nicestpeople, will not be lawyers
anymore because they'retechnically great.
Lawyers are actually the AImachines.
This is the irony of the wholething.
Anyway, that's emotionintelligence.
I lack emotion intelligence.
I'm not saying this because I'mreally emotion intelligence.
I'm not.
I lack emotion intelligence.
I just recognize.
The fact is gonna be important.
The third one was independentintelligence.
(14:25):
And that brilliantly leads intowhat you were saying about this,
which I believe will becomewe'll have more gig economy kind
of work and more fractionalwork.
I.
The job for life is gone, but Iactually think the job for the
year might be very soon upon us.
And again, that becomes amassive thing about who trains
who, if you're not training upwhat happened to loyalty.
But again, remember, these areall concepts.
(14:47):
So in the fifth industrialrevolution, which we're in, a
lot of those concepts willdisappear because actually your
manager will realize that theyare if middle managers still
exist, you know that you've gotan amazing amount of power.
And actually employees are moreimportant.
Highly productive and highlyflexible.
More employees are moreimportant than they thought they
were before right?
(15:07):
Now, the other thing is, whichis just with organizational
intelligence and we've got tobecome cleverer.
We've got to do more with less.
It is likely that we're gonnahave much smaller businesses,
not much bigger, and it's verylikely we're gonna have some
hyper successful businesses,which are tiny.
Which are tiny businesses.
Now, if you look online andyou're massively geeky around AI
(15:30):
stuff, you already know this ishappening'cause of vibe coding,
and you already know this ishappening because people are
making mobile games and makingsilly amounts of money with only
one person.
But you're gonna have Sam Altontalks about the, first business
very soon we'll have a billionturnover company of just one
person.
But that's because they'll have50 agents.
Or a hundred agents.
(15:50):
Or a thousand agents.
Andreas Welsch (15:52):
And, to me that brings up another good point
about middle management, right?
To what extent will we stillhave or need middle management
or will there be orchestratoragents in, in, in that sense
orchestrating the work betweenother agents and people that
supervise them.
Dan Sodergren (16:05):
Absolutely.
We've gotta remember as wellwe've gotta be careful'cause I
imagine there's quite a fewmiddle managers on the call
perhaps.
I don't know.
But it is a job.
Like I've gotta be careful.
Alright, that job won't, thatjob will disappear.
Yeah.
But it will just shift.
So basically we'll shift therole, but we'll keep the goal,
like the goal of middlemanagement will still exist, but
middle managers might not.
Also, we might be able to givethis stuff to lots more of the
(16:27):
AIs and if you have a load ofdata, and this is why I built
something called your flock afew years ago with Michel
Wichnevsky.
And the idea behind that is thatyou could most probably quantify
company culture and help companyculture through data points Now.
That can now be built reallyquickly with AI, which is why we
didn't get a third round offunding because you, it didn't
have a moat anymore because AIcan do it through three or four
(16:48):
different prompts and maybe alittle bit of knowledge base,
and you could build it with acustomizable GPT.
Now we'd spent 300,000 trying tobuild our own system and open AI
just said, oh, have this one forfree, and you could just use a
prompt awesome.
Great for them, not great forus.
Same principle here is if youhave all that data and knowledge
does, do you need middlemanagement as much anymore?
(17:10):
Now, I would think the answer'sno, but you still need a human
being because some of that stuffis too nuanced for a machine,
right?
But what if the machine ismanaging machines?
Now that's different because thenuance therefore isn't needed.
'cause the machine's never gonnabe crossed.
The machine will never say, I'mnot working anymore.
(17:32):
I'm gonna leave because themachine doesn't have emotions.
Then you don't need emotionalintelligence.
You need a lot of artificialintelligence.
And that's where we are now.
We're in this world potentiallywhere, and by the way, this is
terrifying in some respects,that the middle managers are
manage, are machines that aremanaging the machines.
And really you only have thebosses of the machines.
(17:52):
But what if you are the boss ofthe machine?
And that's the point, isn't it?
You know what, if you are asuper powerful accountant?
'cause you've got 10 accountantsin your company, but it's not a
company, it's just 10 agents.
Andreas Welsch (18:02):
To me that's that's both right, like you
said.
Exciting thinking about wheredoes this actually go so you can
properly prepare and also seewhere the opportunities are.
And on the other hand, a littlebit frightening too, because
where do we stay in, in whereremain as people and what does
it mean to our identities to ourknowledge, to how we see
(18:25):
ourselves and how we've beentrained to work how work has
been done for decades or 200years.
And
Dan Sodergren (18:34):
but we have to remember that I'm sure we used
to work in the fields, right?
95% of us used to work 95% ofpeople worked in the fields.
And then when we didn't work inthe field it didn't stop, did
it?
I noticed.
We still work now.
We're still working, right?
Yeah.
But we're just working ondifferent things.
Like we used to work infactories and now not all of us
work in factories.
We used to work in mines, butnow we don't work in mines.
And this is the bit that I don'tget.
(18:55):
This is why I always talk aboutit.
Like the fifth industrialrevolution.
When industrial revolutionhappens, this massive shift
happens.
So the work hasn't gone, it'sjust the tasks move on.
No.
So all you're gonna be doing ismanaging a machine to do more of
your work for you.
Now, the bigger question for usall, and by the way, I think
this is a brilliant question, iswhat happens to that excess
(19:16):
productivity?
Now, that's a great question.
Like where does the money go?
If I get, I literally had aslight argument with someone
this morning on this, andthey're absolutely right to say
it, but I'd never thought of itlike this because I just hadn't
thought about this.
They were like, if I become 400%quicker at marketing and better
at marketing, I don't make 400%more money because my employee
(19:37):
will pay me the same.
They'll just fire three of us.
Andreas Welsch (19:41):
And that's where I think the difference between
corporate employment andfreelancing or running your own
business comes in.
If, you're an independent andyou can do four x the work that
you've previously done, thatmeans you can, or you do it four
times faster, that means you cando four projects.
And even if your rate doesn't goup.
You scale, right?
Dan Sodergren (20:02):
Yeah, absolutely.
And even if it's not times four,it's times two.
Yeah.
It still works still.
It still works.
Yeah.
But, okay.
And then this is the thing, butsame thing, which is why I like
to call AI apprentices automatedinterns.
Same principle, isn't it?
If I had all this and then I gotfive interns, would I have a
better company?
It depends how well I manage theinterns, doesn't it?
(20:23):
It depends how well I do promptengineering.
It also means if I pick theright AI,'cause so many people
use AI and say, I have used chatGPT and I didn't really like it.
It didn't do what I wanted itto.
Therefore it's rubbish.
That's'cause you haven't usedClaude.
That's'cause you haven'tdownloaded reel.
That's'cause you haven't,there's like horses for courses.
Every one of these things isdifferent.
And also you haven't trainedyour AI in the first place.
(20:43):
So it's actually, it's on you.
'cause you have, but also youhaven't been trained how to use
AI.
And that's why I teach peoplenow how to use AI because
actually it's been going soquickly that people have just
had a go.
But I haven't been trained init.
You remember when computersfirst came out, you had to be
trained how to use a computer.
It's thinking you can plan yourway through AI is potentially
(21:04):
quite dangerous, but also it'son your employer to train you in
how to use AI.
That's what we should do.
And you'll be amazed the lowrates for adoption in the UK on
are actually to deal with thefact that there's so many small
businesses.
Most large businesses now havedone AI and they've done a bit
of AI training.
They've got an AI policy.
Were you surprised?
You might be surprised how manydon't actually have an AI
(21:25):
policies.
It's terrifying actually.
Anyway, but a lot of smallbusinesses don't do the AI
training part.
Less than a hundred people thathaven't done the AI training.
And then you, this is what Italk about in the book.
There's this whole thing calledthe secret cyborg that are
turning up and doing the work ona different computer or logging
in with a different browser.
And you but they're doing it onsystems that are free because
(21:45):
you are not paying the 20 poundsa month.
That's really bad.
It's like not giving someone alaptop when they turn up to
work.
I'm not giving them a chair ornot giving them a mobile phone.
Whatever it is, you've gottaspend the money.
If we are, if we've got leaderson the call, I employ you.
Spend the money not onlyresearching and doing some
training on yourself and yourown mindset, but also on what's
(22:05):
the best AI that your people canuse.
Andreas Welsch (22:07):
I see an interesting question from Dita
and I'd love to, to take a stabat it, but also give your
perspective and Dita is asking,what would you recommend to
students who to prepare for thisfuture and.
I teach at the local university.
I teach management informationsystems.
And as part of that, we havebeen talking about AI for
(22:28):
different cohorts and semesters.
And last fall I startedcooperating gen AI and prompt
engineering and a bit of AIliteracy in my class to say,
Hey, look, here are examples ofAI.
Here's how you can try it out.
Google teachable machine.
Here's how you generate animage.
Here's how you use chat GPT andcloud.
And here are differenttechniques, to get students to
(22:50):
warm up to the technology and,use it if they haven't already
done so.
That's another thing then haveused it before.
But then I think the interestingpart where I have conversations
with with fellow professors is.
How do we, evaluate whatstudents actually learn and
retain and do?
And I think it's true that wechange those assessments, right?
(23:14):
It's, not necessarily about theoutput you generate, but if,
you're able to assess, is thisgood?
Is this specific?
Why is it good?
Why is it not good?
How can you make it better in.
What was your thought processgoing into it?
Yes.
So to me that is something weneed to change and to teach.
What's the thought process?
Dan Sodergren (23:31):
I'll give you a great example there.
cause I use it all the time.
I also have a non-for-profit,which I run with a friend of
mine who's a teacher and it'scalled the AI teacher course.
The AI marketing course I runand my training company I run,
but the AI g course is anon-for-profit, and we go around
trying to help teachers use AI,right?
But one of the biggest thingsactually is the mind shift here.
And I always use this examplelike years and years ago, I'm
(23:53):
this old that you didn't havecalculators.
It took 13 years for calculatorsto come inside UK schools, but
teachers literally wanted to banthem forever.
And then of course, we then hadthis revolution where they
realized actually it's no, nobad thing.
They're bringing a piece oftechnology.
Actually, it's great.
So maths changed in the way itwas taught because you had to
show your workings and had towork out other things.
So the calculator was a tool,but it wasn't such a tool that
(24:15):
it destroyed the education point'cause it couldn't have been.
Yeah, it's same thing here.
It's like AI is a tool, it's nota rule, like it's not that
powerful that it should overtakeeverything.
But if you are just assessingpeople on their ability to write
an essay, that's fundamentallydangerous because the AI can now
write the essay.
But we shouldn't be doing thatanyway.
That's why education has tochange now.
(24:36):
But to answer the gentleman'squestion, it's, I say it's the
soft skills, not the hardskills.
We've gotta be teaching the nextgeneration, because actually
it's the questions and thecreative question and the
analytical thinking behind thisAI is brilliant, but you've
gotta know how to ask the rightquestions of it to get the right
responses.
Now that critical questioningthat is something that I don't
(24:57):
think we teach enough inschools.
We also don't teach enough inschools things like
entrepreneurship or how to benice or how to budget or just
life skills stuff.
And I actually know that beingnice is gonna be more important
because.
The, that, the grunt work thegrinding work is got what the
AI's gonna do.
As I always say, we've gotta bemore mindful, less grind.
(25:19):
We've gotta not be thinking, I'mgonna work really hard on this
because actually the machinewill do the hard work for you.
Yeah.
And that changes everything.
Whether are we ready as asociety?
As businesses, we've got to be,because that's how businesses
work and make money.
And we've got, we can only blameourselves and the leadership if
we don't change our mindsets.
(25:39):
But as a society, are we readyfor a new world where you don't
you can love what you do andtherefore because I, as I talked
about in my TedEx talk thefuture of work isn't what you
think.
It's most probably what youlove.
But are we ready for that insociety?
Are we ready for people thathave to be engaged and love
their work to actually get paid?
I'm not sure we're at, I hope weare because that's the fifth
(26:00):
industrial revolution.
But yeah, I think to take a,hold an advantage of that, we've
gotta fundamentally change howwe educate people, including
ourselves.
Andreas Welsch (26:08):
Now then let me ask you this before we wrap it
up for today with all thischange coming and all that
uncertainty and, again, rapidchange or orders as a magnitude
of bigger change than what we'veused been used to.
What makes you hopeful aboutthat future with AI and for the
future of work?
Dan Sodergren (26:28):
The reason why I'm so excited by it is because
now we have a potential wherethe either kind of odd way, the
means of production and now inthe hands of the workers.
Which sounds a bit weird, butbasically we can start creating
our own destinies.
We can do things now with AIthat we could, that are in a
madman's dreams.
Before we could have not doneall this dumb stuff before.
My concern would be thatgovernments and slightly bigger
(26:51):
people in power might start tocorrupt some of that or even
start to delimit it and say,actually you can't have AI
anymore and take it offers.
But right now, you've got theability to take an LLM and put
it into your own machine and toactually build your own agents
and build your own AIs rightnow.
Now that's very exciting'causethat means that a lot of the
(27:12):
cognitive load that you wouldnormally have to do for your job
can be done by a machine, whichmeans that the 50% of your time
can be done to do either morehigh value work, which I think
inherently is more fun for humanbeings.
Or even better have more timeout to see your family and
friends.
Because actually society's gonnaneed us to actually be more
(27:32):
present with our loved ones insociety than it is gonna be
present at work.
But you can only do that if theAI does your work for you and
you can manage the AI.
And that's why I'm excited bythe future of work.
It's not because I'm excited byAI, I'm excited by what AI can
do for human beings to make theworld and hopefully their own
lives a better place.
But I am a massive hippie, sowho knows?
(27:54):
I might just be a technooptimist.
But honestly I see it as beingsomething that changes society
in a very positive way.
And if I didn't I wouldn't dothe job that I do.
Andreas Welsch (28:05):
Yeah.
That's wonderful.
Dan, maybe before we wrap up,what are the three key takeaways
for our audience today from ourconversation that
Dan Sodergren (28:14):
you'd like to I would think that they have to
start to retrain, to retain.
So really start thinking abouthow do you train your people up
and including yourself.
Because at the end of the day,that evolution is really to do
with your own mindset.
So how are you gonna train yourpeople up?
How are you gonna train yourselfup and change the mind shift?
(28:35):
And then the big one for mearound this whole fifth
industrial revolution point is,how are you going to manage the
resources in the business so itcan fully take advantage of this
moment?
I think a lot of people at themoment are thinking AI can just
do stuff without time, money,and energy.
It can't.
It cannot.
But if you put the time in andif you put the energy and you
(28:55):
put some money into it, you willfundamentally change your
business.
And then hopefully change theworld for your, for yourselves
and your family and everyoneelse.
So as long as we're mindful thatit's not just, it's not a magic
bullet and we go, shim sham andthe agents will do it all I
think we'll be okay.
Andreas Welsch (29:11):
Wonderful.
Alright, Dan, thank you so muchfor joining us and for sharing
your experience with us.
I really appreciate it.
Dan Sodergren (29:17):
Thanks for having me on.
Andreas Welsch (29:19):
Alright, and for those of you in the audience,
thanks for, joining us as well.
Today, we'll talk about agent AI and how it's
reshaping our jobs forever, andwho better to talk to about it
than someone who's activelyworking on that and thinking
about that, Dan Sodergren.
Hey, Dan, thank you so much forjoining.
Dan Sodergren (00:12):
Thanks very much for being here.
And just a quick one, I'mdefinitely Dan Sodergren, and
I'm not an AI nor an AI agent,nor a hologram or any of those
things.
ElevenLabs and HeyGen are doingbrilliant work in this sphere,
but promise, I promise you, thisis me okay.
Just to make it absolutelyclear.
Andreas Welsch (00:28):
And they are doing fantastic work.
I've been experimenting withthem too.
And my, my family and friends,they hardly can tell the
difference if they even can.
But I am real too.
So just to get that out therenow.
Hey Dan, why don't you tell us alittle bit about yourself, who
you are and what you do?
Dan Sodergren (00:45):
Yeah hi.
My name's Dan Sodergren.
I'm a keynote speaker, also acorporate trainer.
Have been for about 20 years andI've been for Ally more than a
decade, been talking aboutdigital transformation and also
AI before I suppose it wascalled artificial intelligence.
It was called a load of otherthings.
And I then had the joy of, wewere just joking off stage that
(01:05):
I had the joy of being called atech futurist.
But I'd rather be called abillionaire philanthropist.
Tech futurist just means I didstuff far too well in the
market.
I had a augmented realitycompany about 20 years ago and
did some pretty cool stuff, butI would've been much, I would've
much happier to have been a lotricher.
So this time I'm riding thiswave, so to speak, but I think
hopefully I've caught it at theright time.
(01:26):
I'm not talking about hologramsand things, but AI agents are a
part of this future of work.
And I wrote a couple of bookslast year, one called The Fifth
Industrial Revolution, which isall about this new thing between
me and you.
I thought we had another yearleft until we had these AI
agents, to be honest with you.
It hasn't necessarily caught meoff guard, but the rapid pace of
change, as has got quicker andquicker.
(01:49):
So yeah, it's even for us Techfuturist, this is a pretty
freaky, freaky time to.
Andreas Welsch (01:54):
That's for sure.
I remember when, what was it,two years ago, we started the
first agent frameworks BabyAGIand what they were called.
And then within a couple monthsor a couple quarters, everybody
and in tech vendors were jumpingon the bandwagon and said, Hey,
we've been doing AI and agentsall along.
And here we are trying to holdon and figure out what is
(02:16):
actually real and what does itmean?
Dan, what'd you say?
Should we play a little game tokick things off?
Dan Sodergren (02:22):
Go on then.
Yeah.
Yeah.
I'm slightly scared by thislittle game.
Okay.
Andreas Welsch (02:25):
Okay.
So let's see.
Here we go.
If AI were a vehicle, what wouldit be?
60 seconds on the clock.
Go!
Dan Sodergren (02:35):
If I was a vehicle, what would it be?
I would have to have lots ofdifferent ways of it being a
vehicle i'd.
I tell you what, it's a funnyone.
I could say something like asubmarine because you don't
really see it coming until it'stoo late.
So maybe a submarine.
I'm gonna go with, if it was asubmarine, I'm gonna go for
submarine.
Does that count as a vehicle?
Andreas Welsch (02:52):
I think it counts.
If I may overgeneralize I hear aBritish accent, so it might be
yellow submarine, something likethat.
Playing to the stereotypes, Ihope you forgive me.
Dan Sodergren (03:04):
Oh to be honest with you, I think the Beatles, I
think it links in, doesn't it?
Didn't the Beatles do an AI tunea bit?
Go and Paul McCartney did.
So there you go.
Yes.
Even yellow submarine people aredoing it right.
Andreas Welsch (03:15):
Great response.
Submarine.
I don't know if I could be stuckunderwater for three months or
something that.
That takes a lot, right?
Yeah.
Think of all the news that you'dbe missing and everything that'd
be happening in, in those threemonths.
Crazy.
Dan Sodergren (03:30):
I always wondered, so they don't have we
must probably get we diverseamount too much.
I wonder if they still have the,they don't have wifi either.
God, what a terrifying idea tobe under the water doing that
for that long.
You'd have to be a certain typeof person.
Andreas Welsch (03:42):
Let's, that's, yeah.
Good question.
I.
I know it.
20, 30 years ago there was nocommunication radio silence.
Oh
Dan Sodergren (03:50):
real.
Of course.
Yeah, of course.
Oh God.
Oh, then AI's nothing likesubmarines.
So that was a terrible thing tosay.
That was a terrible metaphor.
Andreas Welsch (03:57):
No.
But things happen in stealth andthen they come out and we're
surprised.
Dan Sodergren (04:03):
There you go.
Perfect.
Like agentic AI.
There you go.
Andreas Welsch (04:07):
Exactly.
We've been talking about thefuture of work for more than a
decade.
I remember those earlyconversations, 2015, 16 ish of
how awesome things were going tobe.
And not only would we all spendour days at the beach, while
some other things are doing thework for us but even in, in a
more serious context that.
(04:27):
We, we would see work being donedifferently.
And I've always perceived it asa change in how we organize work
and, how the present day worklooks like.
But I feel now with agents,we're turning that upside down
so quickly and, I'm justcuriouswhat you're seeing there, what
does it mean for people in abusiness doing this work?
(04:49):
If these agents can do marketresearch and can take on
operational tasks and a wholeother host of things, what does
it mean in, business?
Dan Sodergren (04:58):
Number one, I think it's most probably the old
classic, I used to start all myall my law essays with this,
with the John Spiro quote, whichis"Without definition there can
be no inquiry." Because youcan't really, one of the big
things we have in AI and I thinkin the tech world is that.
We presuppose that everybodyknows what we're talking about
and that everyone's talkingabout the same thing, right?
Yep.
So we are talking about AIagents for a very long time, but
(05:19):
then people say, oh, of coursewe've got an AI agent, it's a
chat bot, and it's ah, thatmight not be an AI agent, but I
can see where you're going withthat, right?
And you might say, technically,isn't ChatGPT an AI agent.
No, it's not and that's why thedefinition's important.
So if you define it much morearound we dunno if you agree
with me, but adaptability andbeing autonomous and those kind
(05:41):
of things is more important inthis kind of AI en gen world.
Only'cause I did a poll onLinkedIn, I think it's 22% of
people didn't know what an AIagent was.
So I was like, ah, I'm gonnatalk about it.
Let's at least define it.
So I think if you look to thatbeing autonomous and
adaptability but also being ableto do, almost doing an objective
(06:02):
rather than a task, if thatmakes sense.
As I used to say a bit of acliche thing, but jobs may
shift, but but skills uplift.
It is the same thing with AInow, then it's now got a new
skillset.
It's now got a new ability.
Yeah.
Yeah.
So now AI agents have thepotential of doing an objective,
a full objective, not quite, Idon't think quite a full job,
(06:23):
but a full objective, right?
Rather than just a task.
So they are very different tochat bots.
They are very different to GPT.
They are very different based.
And I don't know if you've seen,and you will have seen of
course,'cause you're an expertin it too the stuff that's
coming over from China withManus and all these other
things, and even with operatorand these other things it's now
astounding what AI agents cando.
(06:44):
And even though we talked aboutme coming on the show, I think
it was a couple of months ago,even in those couple of months,
it's fundamentally changed whatI would've said because now
we're at a place which I didn'tthink we'd be at for at least
six months to a year.
So now AI agents are in that.
We did, which is again, anothercliche, AI won't replace you,
but somebody using AI will,which like we used to say, it's
(07:06):
now actually AI might be usingitself to replace you.
And I think that's the thing.
So from a leadership point ofview, it's fundamentally
different.
I talked about this in the bookI wrote earlier this year, which
was the kind like, like a how toguide it was called, Leadership
in the Times of AI.
The point of the end of thatbook, and I was, had the joy of
(07:27):
writing with about 15 other AIand security experts.
The bit at the end of that bookwas talking about that how do we
as leaders.
We have to fundamentally changebecause we've now got potential
of people having, it's almostlike they're the their own
company now.
I used to joke that it was a bitlike, how do you manage the
Avengers?
They've all got superheroes andthey've all got superpowers,
right?
But now it's like, how do youmanage one person who
(07:49):
potentially is five people or ispotentially a company?
In their ability becauseeverybody now almost everybody
now, who has a good computer anda good computer understanding
and has been trained in thisstuff can start potentially
building AI agents, and that'sonly happened in the last couple
of weeks.
Now, that is fundamentallyterrifying from an employee
(08:13):
point of view, from an employerpoint of view, and an employee,
the potential of it, it's likenuclear power.
The potential of it isastounding.
It could save the world, but itcould also destroy the planet.
It's the same, it's not the sameall.
Andreas Welsch (08:28):
That's a pretty broad spectrum.
Dan Sodergren (08:30):
That's for sure.
Exactly.
Of outcomes.
But just like we talked aboutwith the submarine thing now
we're talking about nuclearsubmarines right now.
Andreas Welsch (08:37):
I see a question in the chat from Kerry, and he's
asking, what are your thoughtson transferring or sharing
decision making with models andhow we think about
accountability to ourstakeholders?
I think that fits nicely in,into the business discussion.
Dan Sodergren (08:49):
It's a fantastic question, and I think this is
the biggest thing now for notjust employees, but employers,
right?
It, a month ago it was what'sour AI policy?
Or now it's gonna have to bewhat's our AI agent policy?
It's actually what's ouremployee policy and actually
what's our employee contract.
And we have to, and by the way,I would love that governments
are doing this, but governmentsmove too slowly.
(09:11):
We are now in this world wherewe have to start to question
pretty much everything about ip,pretty much everything around
privacy, pretty much everythingaround anything that you now do.
Yeah.
So you have this big thing nowand I've gotta be careful with
words'cause this is getting abit too futuristic, but it's
only in a couple of months,right?
There will need to become aconversation about who owns the
agent.
(09:32):
Yeah.
If you've built an agent on acompany laptop with company
information, but it was youdoing the work to build the
agent who owns the outcome.
Yeah.
Yeah.
Now this is the nice thing aboutit is these agents will take the
grind.
Yeah.
And then human beings willmanage'em and refine.
So as I'm talking about, asindividuals, we've gotta become
(09:54):
a lot more mindful about what wedo, much more mindful about what
we do, and are actually morestrategic than ever.
Because actually what you aredoing now is gonna echo for a
generation, and you might feelthat silly, but the data that
you are gonna be producing inyour job is that your
intellectual property and thear, the old argument, of course,
is no, but it's weird, isn't it?
(10:15):
This the great question.
It's if I create a great promptat work, what's my incentive to
share the prompt?
Yeah.
And at the moment there's no wayyou don't go, oh, you get a
bonus for a thousand pounds'cause you did a great prompt.
But then we laugh a bit, right?
We laugh, but it should.
It should be when we all workedin factories and someone came up
with a great idea to make thefactory better, they got money
because they came up with agreat idea.
(10:37):
This is how it works.
Right now we've got this bitwhere we're talking about
knowledge.
There aren't the incentives inplace right now to really reward
people.
Now, if you are clever enough tostart coding an agent, which
does your job for you.
Are you still doing that job?
Now?
That's a big question.
Andreas Welsch (10:55):
I think you bring up an excellent point and
I haven't heard anybody elsearticulate it at all.
And definitely not in the sameway.
And that is if you are anemployee at a company, you come
in with a certain knowledge, yougain knowledge.
But once you leave, you takethat experience with you to
whatever other company orventure you, you go into to your
(11:16):
point, if you build an agent andthat agent stays with the
company, what happens to thatknowledge that either you have
transferred into it in buildingthat agent if you leave, if
you're an employee today, youleave there are no agents, then
your knowledge is gone.
It's no longer accessible forthe company beyond what you've
documented.
(11:36):
But if this agent continues todo work for the company, yeah.
You've built in and conceivedand fed with your insights and
your previous knowledge even.
Exactly, right?
Dan Sodergren (11:47):
And this is what we, no, and it is not quite the
same as when it was a humanbeing because, and I used to
joke, and I still think it'strue AI is just automated
interns.
That's what it should stand for,right?
It's, you've got an intern thatyou have to train and that's how
you get better at the AI gettingbetter.
It's not just promptengineering, it's managing the
AI right now with that momentwhen if your apprentice then
(12:08):
moves on and the apprenticebecomes the master.
Like, where's the situationthen?
As human beings, we've gottalearn to manage AI before the AI
manages us, and that's where weare right now.
Now, the question for everybodyis if the productivity increases
are huge, what happens to thefuture of work?
Because if an AI agent can dosome of your job and then starts
(12:30):
doing 50% of your job.
Is your job to manage the agent?
And the answer is yes.
By the way, I do not believethat we should take human beings
out this equation.
I do not believe they should besemi, they could be autonomous,
but they can't not be managed.
They must be managed with humanexperiences, human emotions.
Otherwise we, are in a hugeamount of danger, not just from
(12:50):
as ethically you know, thereason why, because of
hallucinations, et cetera.
But it's a big question we'veall gotta ask ourselves is.
We are managing the machinesthat will take on most of the
work, but we have to manage themachines.
Andreas Welsch (13:03):
That's right.
And, let me build on thatbecause I think we're seeing a
lot of these return to officemandates.
We're seeing CEOs ordering thestaff back.
At the same time we're seeinglayoffs and we need to prepare
for AI and trim our workforce,have more capital to, to invest.
Is that the future of work?
(13:23):
And to your point, right?
I was discussing that with JoeReis a couple weeks ago on the
show.
Will we all have multiple jobsbecause now agents do 50% here,
and we ideally maybe not have50% more time.
Do we need a second job to coverthe other 50% and the other 50%
of the income?
Dan Sodergren (13:43):
Again, in the book the Fifth Industrial
Revolution, I talk about thefour different intelligences
that we are gonna need.
We need artificial intelligencetaken as a given.
I think we need more emotionalintelligence.
Ironically enough, as the hardskills become diminished, the
soft skills go up.
We are gonna need to be morecooperative and nicer to each
other because the people whoaren't very nice will basically
don't have jobs because thepeople who are technically great
(14:04):
lawyers but aren't very nicestpeople, will not be lawyers
anymore because they'retechnically great.
Lawyers are actually the AImachines.
This is the irony of the wholething.
Anyway, that's emotionintelligence.
I lack emotion intelligence.
I'm not saying this because I'mreally emotion intelligence.
I'm not.
I lack emotion intelligence.
I just recognize.
The fact is gonna be important.
The third one was independentintelligence.
(14:25):
And that brilliantly leads intowhat you were saying about this,
which I believe will becomewe'll have more gig economy kind
of work and more fractionalwork.
I.
The job for life is gone, but Iactually think the job for the
year might be very soon upon us.
And again, that becomes amassive thing about who trains
who, if you're not training upwhat happened to loyalty.
But again, remember, these areall concepts.
(14:47):
So in the fifth industrialrevolution, which we're in, a
lot of those concepts willdisappear because actually your
manager will realize that theyare if middle managers still
exist, you know that you've gotan amazing amount of power.
And actually employees are moreimportant.
Highly productive and highlyflexible.
More employees are moreimportant than they thought they
were before right?
(15:07):
Now, the other thing is, whichis just with organizational
intelligence and we've got tobecome cleverer.
We've got to do more with less.
It is likely that we're gonnahave much smaller businesses,
not much bigger, and it's verylikely we're gonna have some
hyper successful businesses,which are tiny.
Which are tiny businesses.
Now, if you look online andyou're massively geeky around AI
(15:30):
stuff, you already know this ishappening'cause of vibe coding,
and you already know this ishappening because people are
making mobile games and makingsilly amounts of money with only
one person.
But you're gonna have Sam Altontalks about the, first business
very soon we'll have a billionturnover company of just one
person.
But that's because they'll have50 agents.
Or a hundred agents.
(15:50):
Or a thousand agents.
Andreas Welsch (15:52):
And, to me that brings up another good point
about middle management, right?
To what extent will we stillhave or need middle management
or will there be orchestratoragents in, in, in that sense
orchestrating the work betweenother agents and people that
supervise them.
Dan Sodergren (16:05):
Absolutely.
We've gotta remember as wellwe've gotta be careful'cause I
imagine there's quite a fewmiddle managers on the call
perhaps.
I don't know.
But it is a job.
Like I've gotta be careful.
Alright, that job won't, thatjob will disappear.
Yeah.
But it will just shift.
So basically we'll shift therole, but we'll keep the goal,
like the goal of middlemanagement will still exist, but
middle managers might not.
Also, we might be able to givethis stuff to lots more of the
(16:27):
AIs and if you have a load ofdata, and this is why I built
something called your flock afew years ago with Michel
Wichnevsky.
And the idea behind that is thatyou could most probably quantify
company culture and help companyculture through data points Now.
That can now be built reallyquickly with AI, which is why we
didn't get a third round offunding because you, it didn't
have a moat anymore because AIcan do it through three or four
(16:48):
different prompts and maybe alittle bit of knowledge base,
and you could build it with acustomizable GPT.
Now we'd spent 300,000 trying tobuild our own system and open AI
just said, oh, have this one forfree, and you could just use a
prompt awesome.
Great for them, not great forus.
Same principle here is if youhave all that data and knowledge
does, do you need middlemanagement as much anymore?
(17:10):
Now, I would think the answer'sno, but you still need a human
being because some of that stuffis too nuanced for a machine,
right?
But what if the machine ismanaging machines?
Now that's different because thenuance therefore isn't needed.
'cause the machine's never gonnabe crossed.
The machine will never say, I'mnot working anymore.
(17:32):
I'm gonna leave because themachine doesn't have emotions.
Then you don't need emotionalintelligence.
You need a lot of artificialintelligence.
And that's where we are now.
We're in this world potentiallywhere, and by the way, this is
terrifying in some respects,that the middle managers are
manage, are machines that aremanaging the machines.
And really you only have thebosses of the machines.
(17:52):
But what if you are the boss ofthe machine?
And that's the point, isn't it?
You know what, if you are asuper powerful accountant?
'cause you've got 10 accountantsin your company, but it's not a
company, it's just 10 agents.
Andreas Welsch (18:02):
To me that's that's both right, like you
said.
Exciting thinking about wheredoes this actually go so you can
properly prepare and also seewhere the opportunities are.
And on the other hand, a littlebit frightening too, because
where do we stay in, in whereremain as people and what does
it mean to our identities to ourknowledge, to how we see
(18:25):
ourselves and how we've beentrained to work how work has
been done for decades or 200years.
And
Dan Sodergren (18:34):
but we have to remember that I'm sure we used
to work in the fields, right?
95% of us used to work 95% ofpeople worked in the fields.
And then when we didn't work inthe field it didn't stop, did
it?
I noticed.
We still work now.
We're still working, right?
Yeah.
But we're just working ondifferent things.
Like we used to work infactories and now not all of us
work in factories.
We used to work in mines, butnow we don't work in mines.
And this is the bit that I don'tget.
(18:55):
This is why I always talk aboutit.
Like the fifth industrialrevolution.
When industrial revolutionhappens, this massive shift
happens.
So the work hasn't gone, it'sjust the tasks move on.
No.
So all you're gonna be doing ismanaging a machine to do more of
your work for you.
Now, the bigger question for usall, and by the way, I think
this is a brilliant question, iswhat happens to that excess
(19:16):
productivity?
Now, that's a great question.
Like where does the money go?
If I get, I literally had aslight argument with someone
this morning on this, andthey're absolutely right to say
it, but I'd never thought of itlike this because I just hadn't
thought about this.
They were like, if I become 400%quicker at marketing and better
at marketing, I don't make 400%more money because my employee
(19:37):
will pay me the same.
They'll just fire three of us.
Andreas Welsch (19:41):
And that's where I think the difference between
corporate employment andfreelancing or running your own
business comes in.
If, you're an independent andyou can do four x the work that
you've previously done, thatmeans you can, or you do it four
times faster, that means you cando four projects.
And even if your rate doesn't goup.
You scale, right?
Dan Sodergren (20:02):
Yeah, absolutely.
And even if it's not times four,it's times two.
Yeah.
It still works still.
It still works.
Yeah.
But, okay.
And then this is the thing, butsame thing, which is why I like
to call AI apprentices automatedinterns.
Same principle, isn't it?
If I had all this and then I gotfive interns, would I have a
better company?
It depends how well I manage theinterns, doesn't it?
(20:23):
It depends how well I do promptengineering.
It also means if I pick theright AI,'cause so many people
use AI and say, I have used chatGPT and I didn't really like it.
It didn't do what I wanted itto.
Therefore it's rubbish.
That's'cause you haven't usedClaude.
That's'cause you haven'tdownloaded reel.
That's'cause you haven't,there's like horses for courses.
Every one of these things isdifferent.
And also you haven't trainedyour AI in the first place.
(20:43):
So it's actually, it's on you.
'cause you have, but also youhaven't been trained how to use
AI.
And that's why I teach peoplenow how to use AI because
actually it's been going soquickly that people have just
had a go.
But I haven't been trained init.
You remember when computersfirst came out, you had to be
trained how to use a computer.
It's thinking you can plan yourway through AI is potentially
(21:04):
quite dangerous, but also it'son your employer to train you in
how to use AI.
That's what we should do.
And you'll be amazed the lowrates for adoption in the UK on
are actually to deal with thefact that there's so many small
businesses.
Most large businesses now havedone AI and they've done a bit
of AI training.
They've got an AI policy.
Were you surprised?
You might be surprised how manydon't actually have an AI
(21:25):
policies.
It's terrifying actually.
Anyway, but a lot of smallbusinesses don't do the AI
training part.
Less than a hundred people thathaven't done the AI training.
And then you, this is what Italk about in the book.
There's this whole thing calledthe secret cyborg that are
turning up and doing the work ona different computer or logging
in with a different browser.
And you but they're doing it onsystems that are free because
(21:45):
you are not paying the 20 poundsa month.
That's really bad.
It's like not giving someone alaptop when they turn up to
work.
I'm not giving them a chair ornot giving them a mobile phone.
Whatever it is, you've gottaspend the money.
If we are, if we've got leaderson the call, I employ you.
Spend the money not onlyresearching and doing some
training on yourself and yourown mindset, but also on what's
(22:05):
the best AI that your people canuse.
Andreas Welsch (22:07):
I see an interesting question from Dita
and I'd love to, to take a stabat it, but also give your
perspective and Dita is asking,what would you recommend to
students who to prepare for thisfuture and.
I teach at the local university.
I teach management informationsystems.
And as part of that, we havebeen talking about AI for
(22:28):
different cohorts and semesters.
And last fall I startedcooperating gen AI and prompt
engineering and a bit of AIliteracy in my class to say,
Hey, look, here are examples ofAI.
Here's how you can try it out.
Google teachable machine.
Here's how you generate animage.
Here's how you use chat GPT andcloud.
And here are differenttechniques, to get students to
(22:50):
warm up to the technology and,use it if they haven't already
done so.
That's another thing then haveused it before.
But then I think the interestingpart where I have conversations
with with fellow professors is.
How do we, evaluate whatstudents actually learn and
retain and do?
And I think it's true that wechange those assessments, right?
(23:14):
It's, not necessarily about theoutput you generate, but if,
you're able to assess, is thisgood?
Is this specific?
Why is it good?
Why is it not good?
How can you make it better in.
What was your thought processgoing into it?
Yes.
So to me that is something weneed to change and to teach.
What's the thought process?
Dan Sodergren (23:31):
I'll give you a great example there.
cause I use it all the time.
I also have a non-for-profit,which I run with a friend of
mine who's a teacher and it'scalled the AI teacher course.
The AI marketing course I runand my training company I run,
but the AI g course is anon-for-profit, and we go around
trying to help teachers use AI,right?
But one of the biggest thingsactually is the mind shift here.
And I always use this examplelike years and years ago, I'm
(23:53):
this old that you didn't havecalculators.
It took 13 years for calculatorsto come inside UK schools, but
teachers literally wanted to banthem forever.
And then of course, we then hadthis revolution where they
realized actually it's no, nobad thing.
They're bringing a piece oftechnology.
Actually, it's great.
So maths changed in the way itwas taught because you had to
show your workings and had towork out other things.
So the calculator was a tool,but it wasn't such a tool that
(24:15):
it destroyed the education point'cause it couldn't have been.
Yeah, it's same thing here.
It's like AI is a tool, it's nota rule, like it's not that
powerful that it should overtakeeverything.
But if you are just assessingpeople on their ability to write
an essay, that's fundamentallydangerous because the AI can now
write the essay.
But we shouldn't be doing thatanyway.
That's why education has tochange now.
(24:36):
But to answer the gentleman'squestion, it's, I say it's the
soft skills, not the hardskills.
We've gotta be teaching the nextgeneration, because actually
it's the questions and thecreative question and the
analytical thinking behind thisAI is brilliant, but you've
gotta know how to ask the rightquestions of it to get the right
responses.
Now that critical questioningthat is something that I don't
(24:57):
think we teach enough inschools.
We also don't teach enough inschools things like
entrepreneurship or how to benice or how to budget or just
life skills stuff.
And I actually know that beingnice is gonna be more important
because.
The, that, the grunt work thegrinding work is got what the
AI's gonna do.
As I always say, we've gotta bemore mindful, less grind.
(25:19):
We've gotta not be thinking, I'mgonna work really hard on this
because actually the machinewill do the hard work for you.
Yeah.
And that changes everything.
Whether are we ready as asociety?
As businesses, we've got to be,because that's how businesses
work and make money.
And we've got, we can only blameourselves and the leadership if
we don't change our mindsets.
(25:39):
But as a society, are we readyfor a new world where you don't
you can love what you do andtherefore because I, as I talked
about in my TedEx talk thefuture of work isn't what you
think.
It's most probably what youlove.
But are we ready for that insociety?
Are we ready for people thathave to be engaged and love
their work to actually get paid?
I'm not sure we're at, I hope weare because that's the fifth
(26:00):
industrial revolution.
But yeah, I think to take a,hold an advantage of that, we've
gotta fundamentally change howwe educate people, including
ourselves.
Andreas Welsch (26:08):
Now then let me ask you this before we wrap it
up for today with all thischange coming and all that
uncertainty and, again, rapidchange or orders as a magnitude
of bigger change than what we'veused been used to.
What makes you hopeful aboutthat future with AI and for the
future of work?
Dan Sodergren (26:28):
The reason why I'm so excited by it is because
now we have a potential wherethe either kind of odd way, the
means of production and now inthe hands of the workers.
Which sounds a bit weird, butbasically we can start creating
our own destinies.
We can do things now with AIthat we could, that are in a
madman's dreams.
Before we could have not doneall this dumb stuff before.
My concern would be thatgovernments and slightly bigger
(26:51):
people in power might start tocorrupt some of that or even
start to delimit it and say,actually you can't have AI
anymore and take it offers.
But right now, you've got theability to take an LLM and put
it into your own machine and toactually build your own agents
and build your own AIs rightnow.
Now that's very exciting'causethat means that a lot of the
(27:12):
cognitive load that you wouldnormally have to do for your job
can be done by a machine, whichmeans that the 50% of your time
can be done to do either morehigh value work, which I think
inherently is more fun for humanbeings.
Or even better have more timeout to see your family and
friends.
Because actually society's gonnaneed us to actually be more
(27:32):
present with our loved ones insociety than it is gonna be
present at work.
But you can only do that if theAI does your work for you and
you can manage the AI.
And that's why I'm excited bythe future of work.
It's not because I'm excited byAI, I'm excited by what AI can
do for human beings to make theworld and hopefully their own
lives a better place.
But I am a massive hippie, sowho knows?
(27:54):
I might just be a technooptimist.
But honestly I see it as beingsomething that changes society
in a very positive way.
And if I didn't I wouldn't dothe job that I do.
Andreas Welsch (28:05):
Yeah.
That's wonderful.
Dan, maybe before we wrap up,what are the three key takeaways
for our audience today from ourconversation that
Dan Sodergren (28:14):
you'd like to I would think that they have to
start to retrain, to retain.
So really start thinking abouthow do you train your people up
and including yourself.
Because at the end of the day,that evolution is really to do
with your own mindset.
So how are you gonna train yourpeople up?
How are you gonna train yourselfup and change the mind shift?
(28:35):
And then the big one for mearound this whole fifth
industrial revolution point is,how are you going to manage the
resources in the business so itcan fully take advantage of this
moment?
I think a lot of people at themoment are thinking AI can just
do stuff without time, money,and energy.
It can't.
It cannot.
But if you put the time in andif you put the energy and you
(28:55):
put some money into it, you willfundamentally change your
business.
And then hopefully change theworld for your, for yourselves
and your family and everyoneelse.
So as long as we're mindful thatit's not just, it's not a magic
bullet and we go, shim sham andthe agents will do it all I
think we'll be okay.
Andreas Welsch (29:11):
Wonderful.
Alright, Dan, thank you so muchfor joining us and for sharing
your experience with us.
I really appreciate it.
Dan Sodergren (29:17):
Thanks for having me on.
Andreas Welsch (29:19):
Alright, and for those of you in the audience,
thanks for, joining us as well.
Popular Podcasts
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
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Therapy Gecko
An unlicensed lizard psychologist travels the universe talking to strangers about absolutely nothing. TO CALL THE GECKO: follow me on https://www.twitch.tv/lyleforever to get a notification for when I am taking calls. I am usually live Mondays, Wednesdays, and Fridays but lately a lot of other times too. I am a gecko.