Why I Don’t Use AI to Write — Oliver Burkeman

So, let's talk about AI. Um, I mean, just to give you a little context first, I the the goal of this series is really to, you know, I mean, I know, you know,

there's so much conversation around AI.

It's basically all anyone can talk about and a lot of it is like pontificating, you know, just broad stroke. This is the end days, this is the best days, blah

blah blah. And then there's the reality

blah blah. And then there's the reality in the middle of like how are actual cool creative people using it and and how

will we all kind of use it or not or think about it or not in the coming months, years, decades. Um so this this the goal of this interview is to kind of

really dive into that part of things like how are you actually using it if at all? How are you thinking about it these

all? How are you thinking about it these days? Uh, let let me tell you what I

days? Uh, let let me tell you what I what I understand your your vibe on AI to be and you tell me if I'm right or wrong. Um, from what I can tell, you

wrong. Um, from what I can tell, you seem to be old man yells at Sky.

Yeah, a little bit. But more just like [laughter] this isn't great. It's not for me and I'm a [snorts] little annoyed that other people are talking about it [laughter] all the time.

Which how because maybe that's also some of what I have in me. So, I'm I might be projecting, but is that how would you No, I mean, yeah. Well, first of all, you're uh

yeah. Well, first of all, you're uh tapping into a an accurate sort of allpurpose commod aspect of my

personality that that I that I don't um I don't disavow. Um

yeah, I think I I've been through various sort of phases of viewpoints like there was a time when I I mean a brief time when I

felt myself very very convinced by the kind of real science fiction doomer arguments about uh AI and the sort of instantaneous

eradication of all humanity one day next next month kind of um kind of thing.

Uh but this was sort of that that was a little earlier prior to the the recent chat GPT releases and that was probably

several years ago. Now I just sort of find myself thinking that yeah I what's the way to put this? I I I feel

like I at the most extreme end I sort of feel like I'm sometimes going crazy because it seems so clear to me that certain ways in which people are

thinking about LLMs and integrating them into their lives are kind of not honest not authentic like it's not what's really going on. There's an this

the thing that I found specifically bothering me is this the degree the readiness that people have to kind of uh

believe or proceed as if there's a consciousness there right not a not a position on whether AIs are conscious not a kind of philosophy of mind thing

but just the sort of the basic idea that you are in some kind of a relationship with this uh code and these this this generated output. And and then also when it comes

output. And and then also when it comes to sort of the creator economy, and we can talk about this in more detail, but like [snorts]

a a growing sense that um there could well be an argument from there could well be an important reason not not to go too far into these use of

these tools in order to retain a sort of a a razondet in an economic sense to have a sort of competitive edge. there could be

a real benefit to um sort of doubling down on the the purely human instead of allowing oneself to be shaped by these tools in the way that I think I see

people being shaped by them. But I

should probably say at the same time, right, I'm not coming at this from the perspective that like there's no possibility that this has any

merit or that there aren't going to be already great sort of um innovations in certain corners of scientific research, for example, that are really um transformed by it. So I'm not a total

lite.

Yeah. I mean, it's interesting.

[clears throat] One thing I I clocked early on in all of this, you know, AI stuff is the language around it. You

know, like when we say Google, we we say I Googled something, right? Like I'm the guy doing the

right? Like I'm the guy doing the googling, but with chat GBT, you refer to it as chat GPT did it, you know? So like we gave agency pretty

you know? So like we gave agency pretty quick.

Yeah. Um, and I think that's really hard to like claw back even if even if we wanted to. And I think most people

wanted to. And I think most people don't even realize that they are, you know, kind of giving it soul. They

because they assume it either has it or they're just not thinking much about that. Um, but why do you think that B

that. Um, but why do you think that B Yeah, I think it's interesting. Uh what I'm about to

it's interesting. Uh what I'm about to say may cause people to like completely dismiss me as being worth listening to on any level, but I have been uneasy about

something similar like all the way back to the advent of like Alexa and smart speakers and you know that kind of voice activated technology. [snorts]

technology. [snorts] I have I'm the kind of person who always claims I'm not going to adopt this or that technology and then I and then I do like everybody else and come to depend on it. But actually the voice activated

on it. But actually the voice activated stuff is an exception to that rule. Um

we we don't have it in our house and I'm probably the main person the the main cause of that. Um, and I've always felt

that there's something really just kind of creepy about the idea of a technology that that invites me to treat it as if there's a

a person there while at the same time sort of existing to do my to to obey my every command. Um

because like you know it is perhaps a sort of hyperbolic um analogy but but um

we we recognize that you know the the while the while while the primary reason that like um slavery or indentured servitude is a moral uh abomination is

is what it does to the um people who are indentured or enslaved. There is

something kind of deeply corrupting about being someone who treats someone else as property. And I think probably, you know, people who know more about cantian ethics than me would would would bring that in here and say like that

that whole concept of treating a person as nothing but a means to your end is is like depraved in some way. And although

it's obviously not the same thing when it's just Alexa, there is that sort of like it's not that anyone intellectually thinks there's someone inside the smart speaker,

but it's that sort of I don't even know what the word is, some sort of orientation that you have in your just your interactions. Um, and you know, it's years ago now that people

were noticing that Alexa made turn their kids into [ __ ] because they trained them to just it trained them to just sort of um bark commands and then they had to be reminded that that they were allowed to

bark commands at at that lady in the box, but they weren't allowed to bark commands at their mom or their dad.

So, I think this that is one of the pieces of this, right? this kind of I I don't I really don't think anything good comes from pretend from from sort of convincing

oneself that there's a person where there isn't and it and it that it probably damages one's relationships with with real people and then I guess separately there is this whole idea that

like there's something sort of well then then there's to do with like yeah what what what the output is and what what what the importance of of of being the

person who does the writing is and all that, but that's kind of a slightly separate uh dimension to it, I suppose.

But I like that that I think that we let's go there a little bit. Like what

how do you feel about that? Um and I mean do you use it at all or are you not using it at all?

I mean I I um I I mean firstly I suspect all of us are using it in lots of ways that we can't control, right? in terms

of be it being behind other services and platforms that we're using. Other than

that, I really am I saw somebody classify like what the different gener how the

different generations interacted with with um uh AI and uh it said that you know for Gen X it's a slightly glorified

search engine and that's that and I'm just pure Gen X in that in that in that regard. So I do use um perplexity

regard. So I do use um perplexity because it's way better than Google is now for finding links to [snorts] existing sources, right? And so yeah,

that will be kind of it seems kind of it would be uh dumb on some level to refuse to do that because because I think it's chat GPT, isn't it? Is the

would be the um uh is is the the sort of motor behind that. But that's really

that. But that's really pretty much it. Um,

like I have definitely sort of experimented here and there with with more. I I one of the things that raises my hackles is

seeing the claim that, you know, the only reason not to be in awe of and fully dedicated to these technologies is that you haven't used them enough to see what they're capable [snorts] of. I

think it's certainly true that people who use chat GPT an awful lot in their lives come to think it's really great, but that cuts both ways. I'm not sure what the I'm not sure what that says.

Whether that says that whether that's because they're discovering its extraordinary depths or because they're sort of mesmerizing themselves into the understanding that that that is the way

to to work. Um, and I'm not saying this won't change at all. I don't think I'm sort of putting a stake in the ground to say I will never

extend the my use of these tools in the research phases of of writing. But it seems totally clear to me that I don't want

either to use it to create first drafts of things I'm writing because first drafts are not just kind of uh amateurish stabs at

things. they are where the the concepts

things. they are where the the concepts are connecting or the structure of the piece is being sort of uh imperfectly

built and then nor really do I want to use it for the later stages where my voice which

as a writer which you know a person may may rate or not my voice but it's kind of all I've got I feel like in terms of making what I do distinctive and and

worth reading That's that's the stage at which that is is put in and you know you can always find edge cases right and do

I object to running people running their newsletter post through LLMs to catch grammatical mistakes like that would be

really stupid to sort of object uh on principle uh to that I suspect but but um but still like I really feel like the

thing that I'm doing when I'm bringing ideas together and trying to create some piece of writing is it yeah I mean I'm I'm open to the

possibility that I'm that I'm just a out of touch old man at this point but it's like it the the conscious engagement with those

processes is the point um and every sort of way in which one could skip those steps would seem to me to

reduce the meaning and value of the process.

Does that make sense?

Mhm.

I think so. Yeah. Yeah. I mean

[laughter] [gasps] I it makes perfect sense to me. I I

Do you care if other people do it or does that not bother you?

So, so I think the point I've made in little bits of writing here and there is something like this. It's that when I read a piece of writing and we'll let's stick

to writing for now because I guess because it's the one I know the best and seems to be the sort of canonical

case. When I read a piece of writing

case. When I read a piece of writing that someone else has written, then in most cases, I think that it is critical to the value

of that experience that I am reading the words that emerge from their own conscious emoting thinking sensibility,

right? It's like it's not just that

right? It's like it's not just that the points uh made are good ones or the facts involved are accurate ones or even that the writing is stylish or

something. It's the fact that it came

something. It's the fact that it came from uh another consciousness

a bit like mine. So, it's a it's a it's an instance of a relationship.

And I think that's true with, you know, uh possibly not all forms of writing, but when I'm reading a novel, there's something very important about the fact that for me that this emerges

from the conscious sensibility of the novelist. And so, you know, there was

novelist. And so, you know, there was this it's sort of gone away now, but there was this kind of back and forth between AI skeptics and promoters uh a

while ago which sort of hung on whether AI would ever be able to write a novel to the same standard as a human novelist. And some people said, "Oh,

novelist. And some people said, "Oh, this never never could happen." Uh, and then clearly it would be pointed out to them that it effectively already had happened. That this was all sort of

happened. That this was all sort of yesterday's news. And [snorts] my

yesterday's news. And [snorts] my argument was always not that an LLM that what's important about a novel

written by Jane Austin is is not that it it's unimaginable that an LLM could create something of

equal merit because humans are just super magic or something. It's the fact that it was written by Jane Austin is a big part of what gives it its value.

So, so it's so it's not just about the sort of end end result. It's about it's about where

end result. It's about it's about where the end result comes from. And to that end, yes, I think that if you um I think I do care that the people I'm

reading, I think I do care that they don't um that they aren't heavily or overwhelmingly um relying on

LLMs to generate what I'm reading.

Uh at this point, someone will say, "Well, you but you can't tell." And I'm sure you've already been tricked many times already and or I'm sure you've already read lots written by LLM that you didn't realize was written by

Absolutely. Don't deny that at all. The

Absolutely. Don't deny that at all. The

point I'm making is that, you know, when I do find out, uh, I think something shifts in my experience of the product completely,

right?

I mean, I'm repeating something I put I'm repeating something I put in a new Sorry, just very quickly. I mean, I

I I um I submit to sound like a a lawyer, you know, that if um if we had here an interesting conversation and you were somewhat interested in the ideas

that I was communicating and I was interested in your responses and your thoughts about these topics and you left the conversation feeling like, okay, that was a useful use of a bit of my time. That was interesting. It

was not completely meaningless. you

know, there was something worth doing about that. And you then discovered that

about that. And you then discovered that all along you'd been Yeah.

uh interacting with the simulation, even if the ideas were so somewhat interesting cuz they would still be there, right? something really important

there, right? something really important would be sort of would be sort of sapped from that experience like right it would

like it something would be gone because it turned out that there wasn't a a person there well and because you would have deceived

me like I would have felt lied to right like like someone I think it's that kind of personal like I was tricked you know into believing one thing and wasn't

true. And I think people really don't

true. And I think people really don't like that feeling of like that's just like not okay to be tricked like that.

Yeah, I think I think that's I think that's absolutely right and I think it's sort of quite a deep point because it's it exists on that very obvious level of like

you know if I present my Substack newsletter as being the musings of a person but in fact there is no such person and it's just I'm doing I'm writing 50 of these by putting uh by

prompting uh generative AI but there's also this kind of subtle level in which

almost any kind of writing that comes from a perspective that that that implies that the author has certain thoughts and feelings and priorities in

life and opinions like is already implying that it's human created even if even if you were sort of technically honest about the fact that it was an AI.

So I I'm not sure I'm conveying this properly, but this sort of like I've if I if I read something and no claim is explicitly made that it's by a human being. So there's no lying in

that sense. But like the very act of

that sense. But like the very act of sort of telling me what you think about things is to assert that you are a conscious sensibility. And if you're not, then the whole thing is just yeah, what what are we doing here? Now there

are types of writing that this doesn't apply to, right? I mean, I think it's clear that I don't care if the instruction manual for my new dishwasher or something has been created in that way. I think we got to be honest about

way. I think we got to be honest about that. But for for writing where it

that. But for for writing where it matters that it came from a person, not because I can't not because an LLM won't be able to write and can't already write things like better than I can write them by whatever standard we're using for

better, but because there's no one there.

So yeah that I in in an earlier interview this writer David Prell brought up the kind of this point that like he thinks that there's going to be or there there will be two types of

writing right the kind that like puts the reader in the driver's seat which is like informational like I need to know where to go in Brooklyn and it's going to like give you a piece that's LLM

created just for you. Nobody wrote it.

And then there's the, you know, ride along style, the Joan Diddian, the like here's me and you get to like come on my journey. Um, I mean, do you what do you

journey. Um, I mean, do you what do you think about that? Do you

even as I say it, I'm like, yeah, but I sometimes want the first one to also be a person, you know? I don't know. What

do you think?

Right. Right. No, I I mean I think um I love uh David Pell's um work and his uh long form interviews

with other writers. I think he's right about the distinction. I don't know whether the I don't know how big the set of things where

there being no human involved isn't a problem. I don't know how big that is.

problem. I don't know how big that is.

Right. So I think that you know absolutely in those old rough guides and lonely planet you know um physical tourist guide books which are

repositories of information that could and in many ways I know are being taken on by AI these days. I mean

that sort of role even there there's often very much a real sense that you know it matters that that that was put together by a group of people going to those places and seeing them and sharing

them with you. So that's a sort of a that's a harder case. the um the sort of technical writing for how to use an appliance feels like the the kind of the

the the the other extreme partly just because we're probably already so accustomed to poor quality product in that regard. Now

I'm starting to think it would be fantastic to have kind of really beautifully authored dishwasher. I was

just gonna say in instructions we [laughter] should Joan Diddy inviting like how to make your how do you make your dishwasher be incredible.

No, exactly. And even if that's silly and not going to happen like that incredibleness that tells you something.

The fact that that's a cool thought tells you something. Um

and yeah, you know, I I I do think that um you know, one thing that people have

said, I thought that was where you were going with David PL's observations there is that, you know, those of us who want to either produce and sell or consume

human written writing are going to be engaged in a sort of a a sort of premium you know, it's like micro breweries and

and uh premium uh coffee roasting, right? It's like sort of artisal

right? It's like sort of artisal uh sector for whatever proportion of the population hasn't had their jobs taken away by AI and can afford to do such things. And I think that's probably

things. And I think that's probably there might be some truth to that.

Already being a already working as an author is being in a weird little niche industry. I mean, it's not um it's uh

industry. I mean, it's not um it's uh it's already it's already pretty small

and so uh I can see that and you know maybe the market for you know maybe like maybe most people don't think like me and they'll settle for something that I

can that I turn my nose up at and that's cuz I'm a snob in some sense and there's always a market for the cultural products that please snobs but I do think there's something substant ive

here, right? It's not just kind of a

here, right? It's not just kind of a kind of a [snorts] nostalgic commitment to the human for the sake of it. There's something very

very real about the importance of there being a a human author. The question is whether enough people care to make it

financially viable, I suppose.

Well, [clears throat] yeah, that that's what I was going to say earlier. It it

it seems like right now the open question of [snorts] culture is um is it okay to lie about this sort of thing,

right? Like because because we've like

right? Like because because we've like the the the lying as a problem has is an age-old thing, right? Like I was talking to my friends about um you know the million little pieces dude who wrote

that memoir and then it was on Oprah and then it turned out that he didn't write it and it was a huge scandal because it was like dude you lied to us you know and it's like

I think the question now is like if if that and and there were probably 10 cases where someone lied and we didn't find out but it's like whatever in if if someone were to get outed now

would it be a problem? I think the answer is yes. You know, I I don't I I think already it's kind of as like that you mostly wrote something

with AI, right? Like if if your next book it like came out that Oliver Burkeman like barely touched that thing.

Yeah.

Would that be bad, you know, or or would that be okay? Cuz with a ghostriter, it's fine, but that's mainly for celebrities. I don't know. It gets real

celebrities. I don't know. It gets real weird. It gets muddy.

weird. It gets muddy.

Yeah. And a lot of a lot of contracts I think also I think mine cuz my contracts are fairly standard have their sort of there's definitely you know publishers

are definitely uh requiring that authors that what authors submit to them in return for their advanced money is not um is not generated uh interesting

to any significant degree by um by by AI. Um, yeah. I mean, it gets us back to

AI. Um, yeah. I mean, it gets us back to this question of the lie, right? I mean,

the it should definitely be a problem if you if you claim [clears throat] to have if you claim or strongly imply that that

the book has been written largely by you thinking things up and going and finding them out and actually a significant proportion of it was written by AI. the the more sort of

tricky question comes when like is there any problem at all with that if it's all just open and honest, right? Like if

there's an author's note at the beginning saying that, you know, most of this I did this way, is that a problem? And I, you know, I'm not saying people shouldn't be allowed to do it if they're being open and

transparent about it, but I do think there's a problem in terms of like, yeah, I mean, I just I in most cases, I'm just not going to

want to read that. I'm just not going to want to consume that that writing. Uh,

some I saw someone online put it, you know, if you can't if you can't be bothered to write it, I'm not sure why I should be bothered to Yeah.

to read it. This actually goes to this question like I wrote this whole well two books really on um the the finitude of human life and how

the fact that we have the sort of finite time is a terrifying thing but also kind of gives value to to what we do. And

there is something important there too, right? It's like it's like

right? It's like it's like you're going to pay whatever it is for a copy of a book.

[snorts] Something important is about there's something important about the fact that like I used up a bunch of my time to

create that and I gave that time to that when I could have given it to something else and I thereby created some kind of value that I'm that

I've got the tmerity to ask you to to to pay for. Um,

pay for. Um, yeah, and I thought about it and I used up some of my time doing that thinking and I'm asking you to use up some of your time doing the reading. And this all

just takes place within the world of human limitation where scarcity generates value. And you know, that's

generates value. And you know, that's why it's worth your while paying for and dedicating time to a book. And

something's gone wrong with that if it actually took me none of my valuable time at all.

Yeah.

Um because then like what what's going on in that transaction? Why should you pay? I guess many people would say that

pay? I guess many people would say that the value of books in that context would uh you know tend towards zero. Um and

then why should you why should you give your time to it, right? And and I think already all of

right? And and I think already all of this is kind of true in that like you know there's all those apps that like summarize a book for you in 10 minutes and it's like nobody likes [snorts] that

stuff. That stuff sucks. like like we

stuff. That stuff sucks. like like we all kind of know it sucks. You know what I mean? Like it's there's there's just

I mean? Like it's there's there's just sort of these like laws of physics almost when it comes to like good stuff and it's like I don't

know. I I find that I I try to rest in

know. I I find that I I try to rest in those when I'm when I'm freaking out about this stuff or I'm just like I think it's just going to be fine. I

think people like good stuff. They know

when it's shitty, you know? [laughter]

Like I I don't know. There's not like depth to that maybe, but it's people can just tell. I I I find a lot of solace in

just tell. I I I find a lot of solace in that that people can just tell.

People can just tell. I think the the models are nowhere near good enough for people to be completely unable to tell all the time. But again,

even if we get quite soon, which I suppose we might, to a point where people can't just tell, like even that is not doesn't like I

think you're right to be sort of somewhat optimistic about that particular aspect of it. But it's like it will it

won't what am I trying to say?

If I'm right that it matters that something is created by a human emoting conscious sensibility or not, then it will still matter just as much,

right? when people can't necessarily

right? when people can't necessarily tell that the difference is there. The

point is once I find out and so trust becomes this incredibly yes uh important thing. You need to be able to trust that the person you're who

whose newsletter you're subscribing to is is um is is is producing it uh at least to a significant degree

themselves. And you know, people

themselves. And you know, people sporadically propose kind of symbols and logos that people are going to display on their on their newsletters to prove it. And of course, it won't prove

it. And of course, it won't prove anything because you could use it.

Yeah.

On an AI generated uh newsletter. And

then when the AIs are setting up the Substack accounts themselves, they will obviously want to use those uh those those logos to give credibility to their output. But there's something

their output. But there's something maybe there's something maybe admirable in that.

in that basic sort of the sort of motivation behind that kind of kind of thing. I don't know whether it'll work or not.

Yeah. One of the other writers who um Venitas Ralph, I don't know if you know him, he does a lot of like online Yeah.

So he he's very like pro- AI uses it in his writing.

Says like right now he has it in a different section of his Substack, but he says soon he'll just kind of combine it all. And you know, he's still doing a

it all. And you know, he's still doing a ton of work, but it's very right.

It's very, it's not like he's just like blah blah blah. He's he's doing he's working with it.

Um, but he had this this take that that I wanted to get your thoughts on, which is just that like the very idea of like

an individual being responsible for a piece of work is fairly recent.

um like of the last 500 years, a 500 blip in history I think he said. Um

where before that like artists did not at all feel like they were the ones that produced something. It was much more

produced something. It was much more like, you know, God spoke through them or something and they were the vessel.

Um, and that there was like a linguistic commons that or or an artistic commons that you know people just kind of moved stuff through them.

Um, and that maybe AI is a return to that.

What do you make of that?

It's such an interesting um thought and partly I want to sort of turn the tables and say it's exactly that kind of um

idiosyncratic thought that uh has made me value the writing that I've come across from from

him. So, it's like, you know, uh, put

him. So, it's like, you know, uh, put that in your pipe and smoke it. But I

[laughter] guess, you know, the the the question is, I mean, I'd need some time to think about this, but I think that uh, you know, I'm sure that that's an accurate

account of of history. There are sort of extreme versions of this according to which people just only thought about creativity as the gods speaking through them until sometime a few thousand years

ago. and and uh certainly all sorts of

ago. and and uh certainly all sorts of kind of workshopbased art production in the you know in the Renaissance people um works carrying the impremature of somebody

without them necessarily having touched it and the ways in which kind of people like Andy Warhol brought that into the

into the um contemporary era.

I'm just spitballing here. I haven't got a this is not a perfect LLM style

response but I would suggest that there's maybe there's the idea that there's a sort of modern or the romantic idea I guess of the sort of individual writer who is the

sort of or the individual artist who is a sort of genius in in single self-contained genius who's putting these miraculous things out into the world

and Then there's the idea that in fact what that person is doing is channeling something from beyond their own simple egoic

consciousness that it's the collective unconscious or it's the sort of all the sort of mythic resources of that culture.

Now those two things seem in some ways just to be different ways of talking about the same thing. Maybe the latter is sort

of truer somehow.

The kind of role of aggregate and generic um writing and thought that that that seems

to me to be at the core of how the LLM's work seems to be like a separate thing.

Right? So I'm not I'm uncomfortable with that idea that colle because the collective consciousness or the sort of mythological world is is collective that therefore just because LLMs are based on something

collective too that they're the same thing. It seems like

thing. It seems like it's one thing to talk about the specific way in which one writer or group of writers say expresses tales that have been in the culture in broader ways for for forever or they

might experience themselves as giving voice to um ideas that are moving through them rather than coming from them.

and something else entirely to just take a whole collection of individuals and the work that they all produced uh as individuals in the normal way we think about it anyway. Most of them probably

thinking of them their work in that way.

Uh train large language models on that stuff and then have them put out. I mean

I'm at the limits of my technical understanding here for certain but I do think that as I understand how those uh

how these AIs um create their outputs. there is

something sort of deeply and fundamentally tending towards the generic towards the towards the most sort of um the sort of

averaging out of perspectives and viewpoints and opinions. Right. So I I'm not I know the

opinions. Right. So I I'm not I know the sort of um probabilistic next word generator critique is is kind of contested in various ways, but at some

in some fundamental way it seems to me that there's something kind of antique creative going on

in that process because it because it can only possibly be uh exerting a sort of tendency towards making what you what

you do with them generic and um and generic is not the same as collective necessarily. Wow. Yeah, I

collective necessarily. Wow. Yeah, I

don't I need to think about that. Big

thoughts. [laughter]

I I I I think I get what you're saying.

Um and I also got what he was saying. I

mean I I find that whenever I'm whoever I'm talking to, I kind of just am like I bet that person's mostly right. But but

I also feel like um I I don't know my where where I kind of land on it is like do I actually feel good when I'm using it if I'm using it for creative work?

And right now for me the answer is like not really.

I just don't feel good. It just doesn't feel you. I guess that that's where I

feel you. I guess that that's where I keep coming back to is like the feeling of it. But it's just like

of it. But it's just like I feel a little gross. I don't feel very creative. I I don't know maybe I'm a

creative. I I don't know maybe I'm a kaggen too then but like that's just kind of right and tell I don't like how it makes but that bring you know [clears throat]

right and there's two things to I want to say about that like one of them is sure maybe that is a kind of intuitive pointer towards the attitude you should

be taking towards these things but also just that feeling in and of itself is like that's that's why we do these

things right It's like um it's um on some level to spend your large parts of your life doing this stuff is in order to feel, you know, in some sense good about doing it. Not that you're

going to be cheery and happy at every single stage of the creative process or no one I know anyway, but that but that you know if you don't feel good about it

persistently, then that's its own message. like why why do that thing uh

message. like why why do that thing uh with your life?

Well, yeah. And I I think the fact that the very fact that I don't you you're not like I don't expect to feel cheery and happy. I actually expect the

and happy. I actually expect the opposite and then I expect maybe to like kind of get to the other side of that for a little bit, you know? But but it

but and and I think with LLM it kind of just like cuts out that part where it's like a mess and I kind of hate myself and you

know like that's that's that's half the thing [laughter] for me is to like get to the other side of that.

Yeah. Like

is maybe Yeah. No. Yeah.

Yeah. No. Yeah.

For a brief moment.

Yeah.

You know.

Yeah. No, I think that's a really good way of putting it. Yes. I mean, when you say it doesn't feel good to use uh AI, you're not saying that when you're not

using AI, you feel just tiptop at every uh every point in the process. But I

think what you're feeling, I would say, is something like a sort of authentic engagement with reality, right? You're

sort of you're in life. Um and

that can absolutely be um uh that can absolutely be a very unpleasant experience, but it's an alive experience.

Yes. Yes. And it's and it's the recognition of it that I I find kind of the best, you know, [snorts] like even when it's bad,

you know, especially when it's bad. I think yeah, our society is really not great at um being okay with that part, you know. So,

I find that art that does that, I'm like, "Yes, thank God." you know, and I mean maybe what we're coming to

here is some sort of more consensus kind of conclusion that, you know, if you have a clear sense of what gives you the feeling that what

you're doing is meaningful, if you have a clear sense of what it is, then maybe there are people for whom it absolutely

isn't this process of um, you know, thinking the thoughts and manually and consciously bringing the ideas into

connection with each other and trying to or doing all the things that that that LLM's take off your hands. And I guess

maybe that's that's fine. I I struggle to see what the what the um meaning is if it isn't that. But

you know quite possibly there are I don't know pe people who find other parts of the overall process of kind of deciding what to write about doing the writing putting

it out into the world to be the fulfilling parts. Um and I think the

fulfilling parts. Um and I think the really important thing is to have a yeah cleareyed sense of what you find um meaningful in what you're doing and not

um uh sort of outsource precisely the bit that gave it the meaning which is something we're prone to do in lots of areas of life I think

right for the often the reasons that you write about a lot which are like I got to save time you know like and I got to make Right. Um I need to I need to figure out

Right. Um I need to I need to figure out how to make money doing this thing I love. So let me like fast track it.

love. So let me like fast track it.

Yeah. Right. Right. And I mean clearly that's what's driving a lot of it.

Right. It's a sense that you got do this or you'll get left behind.

Um and then you won't be uh able to make a living. I

a living. I I have a hunch and it's not much more than a hunch although we've discussed some good arguments for it here I think but like that that it might be the other

way around that it might be that my best chance of continuing to make a living doing this thing that I find meaningful is to is to um is to not join the race to do the thing that everyone else is doing.

therefore thereby making myself the same as everyone else and not worth buying the books books from because what's the difference?

Mhm. Yeah, that's kind of my last question was around like you know advice for writers and creatives either who are a little bit further along or who are

just getting started who like you know don't want to fully [snorts] use AI but are worried about whether that's possible.

Yeah. I mean, I would say that if you're I if if you feel that you have something to say or a perspective, a viewpoint,

however you define that, to to communicate to the world.

Um, you know, it's a matter of experiment. And these days,

experiment. And these days, experimenting is is easier than it's it's ever been. But it's

to sort of give up on the chance of that resonating with people before you've even tried would seem to be uh uh a tragedy. So I would I would definitely

tragedy. So I would I would definitely encourage anyone who feels um that sense of dissonance between what they're being asked to do and what what

they think they really bring uh to the world to focus on the what they really bring to the world for at least a bit to see how people respond because I think

that people are going to continue and ever more vigorously respond positively to the sort of marks of the the signs of

the human uh and the the very best way to do that is just to be a human doing it. Uh rather than um try desperately to

it. Uh rather than um try desperately to create prompts for LLM so that they'll make it sound like like like it was done by a human.

Yeah. I mean I think a lot of great art that's going to come out is actually going to be maybe about that dissonance, right? That that that that conflict.

right? That that that that conflict.

Yeah.

between what you feel like basically the economy wants you to do and what you want to do, you know, like the new version of that is going to be around AI.

Um, yeah. No, absolutely. And I've and I've

yeah. No, absolutely. And I've and I've always felt on these this kind of issue of like how much do you compromise what you want to be doing

in order to make it financially viable. There will always be people who

viable. There will always be people who basically suggest that the answer is is you compromise everything, right? You

just you just jump on the bandwagon and go with the herd. But there a certain point in that process and they very rarely go this far in their arguments.

at a certain point in that process like retrain as a plumber. And that might be a brilliant thing to do, by the way. I'm

not against that in any way. But like

why why continue to why insist on continuing to travel under the banner of I'm a writer.

If you are automating away all those aspects, like if you're going to go that far, then maybe that's just not the thing for for you to be doing that would bring you alive in life. And like that's

great. It is for me, but everyone's

great. It is for me, but everyone's different. That sort of desire to cling

different. That sort of desire to cling on to a certain kind of label while actually not doing any of the things that are involved in it um seems like the only the only the only wrong answer to

Yeah.

to the question.

Yeah.

Okay, cool. So, but we there's two more little parts. They're very quick though,

little parts. They're very quick though, so we're almost done. One is just a a recipe part. Um I don't know if we if

recipe part. Um I don't know if we if you have any to share of like how you use it. You talked a little bit about

use it. You talked a little bit about like per perplexity but I'm curious if there are any specific kind of use cases

of how you are finding it useful AI um in your creative process even if that is just perplexity but like about something specific or whatever.

Yeah. Yeah, I mean basically the only uh use case here is I use um uh Plexity to uh generate

uh lists of other sources pretty much all of which are you know written by individual

human beings. So I I'm drawn to the way

human beings. So I I'm drawn to the way that Perplexity, you know, footnotes and links the sources uh in its in its output. There's

something about the search engine field that doesn't um you know, I'm use natural language for the for the queries, but I'm not sort of engaging in

uh back and forth dialogue as if with a human. Uh, I'm just typing in, you know,

human. Uh, I'm just typing in, you know, I feel quite free not to use please and thank you or Mhm.

question marks. It's just how I would use Google. Honestly, it feels like I'm

use Google. Honestly, it feels like I'm just using it because of the famous um decline in the quality of uh of of Google as search in the last in the last

few years. So, honestly speaking, that's

few years. So, honestly speaking, that's um that that's where I'm I'm at with it.

Uh and you know, maybe I'll maybe this will just seem completely I mean, for some people it will already seem completely absurd, but maybe this will just seem completely absurd in a few years time to

everybody, even to me.

I still think it's a that's probably a price worth paying for, yeah, continuing to find the writing, the experience of writing as meaningful,

albeit infuriating and maddening. Uh,

Yeah, that's maybe the point is that it's infuriating.

Right. Right. Right. Right. On some

level. Yeah. No, absolutely. Uh yeah,

I'm not trying to smooth the edges away out of that process completely because because that's what makes it worth doing.

Yeah. All right. Last last quick thing is a speed round. So these are just quick questions you can answer with a few words, a sentence, and then we will be done. We will have done it.

be done. We will have done it.

Go ahead. So, how would you describe your relationship with AI? Is it an assistant, a partner? Someone described

it as like their nurse, any kind of How would you describe that relationship you have with it, if if one at all, with a with a person, I guess.

Um, yeah, I I I don't think in relationship terms um for when it comes to when it comes to AI. I don't think I don't think you only have to have relationships with living beings, right?

I feel like you can have a relationship with a landscape or a place. So it's not like the word is completely um

inapplicable but um no I try quite hard to keep it the kind of relationship that I have with like my my MacBook or my um

or the pens I use right it's a it's a tool and I don't want to forget that.

Yeah. What do you wish AI could do but fear that it never will?

Again, I'm not sure that I uh want in in creative work. I'm not sure

that I that I want it to do um uh anything that it uh doesn't yet do.

Obviously, um, you know, if there are specialist domains where it can cure cancer, I'm not going to be sitting here saying don't do that because I'm not completely

insane. But, uh, no, I I sort of start

insane. But, uh, no, I I sort of start from a place that makes me not want to answer that question, I guess.

Fair.

Everyone is making predictions about AI.

Let's make a prediction about humans.

Where do you think humans will be in 10 years?

I think it can be very hard to see in times that feel very full of change and volatility and human and civic society falling apart and civilization collapsing. I think it could be very

collapsing. I think it could be very hard to see the truth of that very well-known observation that by far the best gamble you can make about how we'll be

in 10 years is is how we are now. Um,

yeah.

And I think that sort of like messy at each other's throats, boundlessly creative, constantly on the verge of

like destroying ourselves but not actually doing so. That just seems to me like by far the most um plausible scenario 10 years from now.

Yeah.

I mean, it'd be nice if it was better than that, but that's my that's my prediction.

Yeah. What question about AI and creativity isn't being asked enough?

Well, I don't know how to phrase this, but it is that question of like how much of the value of creativity of the of creative

products is connected to the sensient emoting, thinking awareness of the of

the creator. I think this we talk a bit

the creator. I think this we talk a bit about the conscious, whether AI is or could be conscious and what that would mean, but but I think we could do with talking a bit more about how human

creators are conscious and why that matters so much to what they produce.

Cool. And last question, what do you wish people talked about more instead of AI?

[laughter] almost anything. Um, I I I um

almost anything. Um, I I I um I think that um I I I I don't know what the name for

this topic is, but I think we need to be thinking and talking so much more than we do about uh well, I guess cognitive empathy is one

of the phrases that have been used in this area, right? sort of talking about trying to get inside the minds of people radically different from us. Not

in order to sort of um be convinced by them or to compromise our views somewhere between ours and theirs, but but just to sort of just to spend more

time thinking about what it's like to come from inside another person's consciousness. I think that is

consciousness. I think that is right at the heart of a lot of our civic and political issues at the moment.

Are you are you working on anything new?

Anything we can we can sneak peak to people?

I am working on a new book. It's very

hard to exactly say what it's about at this point. I'm not being koi. I'm just

this point. I'm not being koi. I'm just

struggling for the language. But it is something closely connected to this idea of like aliveness or what it is that gives human

life that feeling of being worth living.

And so AI is definitely implicated in this discussion. Although I'm not going

this discussion. Although I'm not going to write a book on AI because there are enough of those.

I can't wait to read it. Oliver, thank

you so much. This was awesome.

Seriously, this was really It's a pleasure. Thanks for your enormous patience in uh setting this up.

I've appreciated that.

No problem. It was well worth it.

All right. Thank you. Thanks, Alex.

All right. Cheers. Take care at