- I've spent the two years
since chat GPT launched,
steeping in a morass
of academic panic.
Voices from
administration and colleagues,
and anyone else
with enough brain cells
to maintain a sense
of existential dread, crying out,
"We need to figure out
what to do about AI."
Our Ed Tech committee
is developing a policy.
The academic Senate wants
to develop a policy.
The board thinks
we should have a policy.
My dean wants
us all to have policies.
The California Teachers Association says
it's an issue of academic integrity.
The State Senate says
it's an issue of ethics.
"We need to pay for
the AI detection tools."
"The AI detection tools
don't work."
"We need to accept that
our students will use AI."
"How do I prove
my student used AI?"
It is incomprehensible to me,
this conversation.
[keyboard clacks]
I hear their words,
see their language floating
across my monitor,
and know the words,
but I cannot get to
the meaning
because I simply
do not understand
why they are talking about it
in this way.
- [Kermit the Frog]:
♪ New York, I love you,
but you're bringing me down ♪
- with all these empty words.
This is not the conversation
I think we need to have.
[song continues]
This is the conversation I need to have.
[overlapping
music and poem]
[Gertrude Stein]: " 'If I Told Him,
a Completed Portrait of Picasso'.
If I told him would he like it.
Would he like it if I told him.
Would he like it would Napoleon
would Napoleon would
would he like it.
If Napoleon if I told him
if I told him if Napoleon.
Would he like it if I told him
if I told him if Napoleon.
Would he like it if Napoleon
if Napoleon if I told him.
If I told him if Napoleon
if Napoleon if I told him.
If I told him would he like it
would he like it if I told him.
Now. Not now. And now.
Now. Exactly as is kings.
Feeling full for it.
Exactitude as kings.
So to beseech you
as full as for it.
Exactly or as kings.
Shutters shut and open
so do queens.
Shutters shut and shutters
and so shutters shut and shutters and so
[poem and music fade out]
- I don't understand Gertrude Stein.
Stein is not nearly well enough
remembered for how influential she was.
An American expatriate poet
living in Paris,
her salons were among the
anchors of the early modernists.
You may not have heard of her,
but you've heard of
the people who visited her.
Ernest Hemingway, Sinclair Lewis,
F. Scott Fitzgerald, James Joyce,
Thornton Wilder, Ezra Pound.
People you've read
or been assigned to read.
We remember Hemingway
because he wrote like this.
We remember Fitzgerald
because he wrote like this.
The right kind of day
and the right kind of moment,
and Pound's
"In a Station of the Metro"
still recites itself completely
in my head, a perfect image.
"The apparition of
these faces in the crowd:
Petals on a wet, black bough."
We don't remember Stein
because she wrote like this.
This is "If I Told Him,
a Completed Portrait of Picasso",
published in 1924,
and continuing the project
of her 1914 book Tender Buttons,
a phrase she never defined.
To me that phrase
"tender buttons" feels right:
small, soft contradictions,
words that seem like
they should go together
but do not actually
make meaning.
That is how Stein's poetry feels.
There is something compelling
about the rhythm of her nonsense,
the feeling of her
almost meaning something,
and then how it falls apart.
"As presently.
As exactitude.
As trains."
But it is incomprehensible to me.
I don't know why Stein
would write like this.
To quote the poet:
- ♪ "Oh, what on earth would make a man
decide to do that kind of thing?" ♪
- But I think the reason
that I don't understand Gertrude Stein
is that she didn't really want
to be understood.
She used language
for something different.
It doesn't communicate.
It reads like stunt linguistics,
which it almost is.
"Buffalo buffalo Buffalo buffalo buffalo",
"had had 'had', had had 'had--'",
These are sentences that,
if you pour over them closely enough,
can be decoded.
Stein's Tender Buttons cannot.
There is something about it
that parses as AI.
It feels like the work of Keaton Patti,
the person most prominently behind
the "I forced a bot to watch whatever"
tweets that used to go viral.
Human-written screenplays
designed to feel like
AI writing attempting to imitate
other human-written screenplays.
It feels like
an autocomplete challenge,
like in the early days
of predictive text and messaging
where you just tap the suggested word
and see what comes out.
It's not how AI really writes,
but it's how people feel
like AI writes.
But Gertrude Stein
was a person.
[Stein reading, underneath]
She wrote "If I told Him"
not because the language
would communicate,
but for some other reason.
- [Stein]: the exact resemblance
as exact as a resemblance
- When I read "If I Told Him",
when I listen to Stein
read "If I Told Him",
the language slips
through my mind.
It doesn't mean anything.
But she did.
Even if I cannot
take hold of that meaning,
I can feel her,
reaching out to me
from behind the text,
trying to connect.
- [Stein]:
As a resemblance to him.
Who comes first.
Napoleon the first.
Who comes too
coming coming too,
who goes there,
as they go they share,
who shares all,
all is as all as yet--
["yet" reverberates and fades out]
[relaxed synth beats]
- [ChatGPT femme voice]:
That's an interesting question!
Gertrude Stein,
a literary pioneer and modernist
who valued innovation
in language and expression,
might have had
mixed feelings about me.
She might have appreciated my ability
to generate new forms of text
and engage with her
literary experimentation on some level.
However, Stein was also
highly individualistic
and invested
in the human experience,
especially the nuances of
consciousness and perception.
So, while she might
have been fascinated by
what AI like me can do,
she might also have been
critical of my limitations
in capturing the full depth
of human subjectivity,
which was a central concern
in her work.
[relaxing synth beats stop]
- A problem with AI is
that it is recombinant,
that it takes
the works of humans
and recombines them
into new forms.
But that problem is not new.
For many years,
the bulletin board
outside the faculty bathroom
in the English building
has featured this poster
the "Plagiarism Spectrum",
next to this ancient and
unflattering picture of me.
Number 7 on the Plagiarism Spectrum
is the mashup,
a paper which mixes copied materials
from multiple sources.
The poster is dated
from June 2012,
more than a decade before
we were concerned about
ChatGPT doing it.
That AI is recombinant is not
in and of itself a problem.
All writing is recombinant.
My course outcomes for English 1
ask student writers
to integrate sources seamlessly
into their own writing,
to mash up.
That we have rules
and procedures
and punctuation marks
and conventions
that govern
what is appropriate
does not change
the essential truth
that this is recombinance.
And there is beauty
in recombinance.
This video started with
a great classic of YouTube,
the duet between
LCD sound system and Miles Davis.
The LCD sound system video
is itself a mashup, conceptually.
Kermit the Frog is not
a member of the band.
Davis is improvising over a film
to create the score,
another mixing of media
to make something new.
The Kleptones
The Grey Album
The guy drinking Ocean Spray to
"Dreams" [by Fleetwood Mac]
Mac Glocky
"If I Told Him" is recombinance:
an attempt to rework
the ideas of cubism,
a movement in visual art,
into poetry.
Picasso's work is recombinance.
He spent 1957 copying
Velázquez's Las Meninas
over and over.
58 times he copied it.
All different.
All Picasso. All Las Meninas.
Recombinance is a genetic term,
but I am using it mimetically.
"Britannica: recombination,
in genetics,
primary mechanism
through which variation
is introduced into populations."
The combination of ideas
from different places
to make new distinct ideas.
When done by humans,
we call this "thinking".
[relaxed synth beats]
- [ChatGPT femme voice]:
You're right to question
how I can "understand"
without experience.
The way I process information
is different
from human understanding.
When I say "understand,"
I mean that I can
recognize patterns,
analyze language,
and respond in ways
that are coherent
and contextually appropriate
based on my training.
I don't understand in the
deep, conscious sense of the word,
like when humans
reflect on their experiences
or have insight
into their feelings.
Instead, my "understanding"
comes from identifying
relationships between
words, concepts,
and context based on
the vast data
I've been trained on.
In short: I can simulate
understanding through patterns,
but it's not the same
as truly experiencing
or being conscious of meaning.
Does that clarify things
a bit more?
[relaxed synth beats stop]
[calm, mystical, twangy
electric guitar]
- Two weeks ago a student
told me that AI is unfair
because you can put a B Paper in
and get A+ writing out.
This is why I do not understand
the conversation
my colleagues are having,
why I cannot think of this
as a matter of ethics,
or academic integrity,
why I don't think we should need
to have a policy or policies.
My student said
you can put
a B Paper in and
get A+ writing out,
and my mind began to fill
with Tender Buttons.
"Feeling full for it.
Exactitude as kings.
So to beseech you
as full as for it."
AI is bad at writing.
No. That is true,
but it's not enough truth.
AI is not capable of writing.
The thing that writing is
is a thing that AI cannot do.
Listen.
- [audiobook narration]:
What Writing Is
Telepathy, of course.
Look.
Here's a table covered
with a red cloth.
On it is a cage the size of
a small fish aquarium.
In the cage is
a white rabbit
with a pink nose
and pink-rimmed eyes.
In its front paws is
a carrot-stub
upon which it is
contentedly munching.
On its back,
clearly marked in blue ink,
is the numeral 8.
Do we see the same thing?
We'd have to get together
and compare notes
to make absolutely sure,
but I think we do.
The most interesting thing
here isn't even
the carrot-munching rabbit
in the cage,
but the number on its back.
Not a six, not a four,
not nineteen-point-five.
It's an eight.
This is what we're looking at,
and we all see it.
I didn't tell you.
You didn't ask me.
I never opened my mouth
and you never opened yours.
We're not even in
the same year together,
let alone the same room.
Except we are together.
We're close.
We're having
a meeting of the minds.
I sent you a table
with a red cloth on it,
a cage, a rabbit, and
the number eight in blue ink.
You got them all,
especially that blue eight.
We've engaged in
an act of telepathy.
No mythy-mountain s***;
real telepathy.
I'm not going to
belabor the point,
but before we go any further
you have to understand that
I'm not trying to be cute;
there is a point to be made.
- AI is good at language.
My students think that
what it produces is A+ writing,
not because it is good,
but because it sounds good.
Obviously, AI can
generate sentences
that are typically clear, coherent,
and contextually relevant,
often capturing nuances
and adapting to various tones
or levels of formality.
And it's true that
the sentences it generates
tend to be grammatically accurate,
concise, and logically structured,
which contributes to
readability and flow.
Sure. This is how I know
when a student is using AI.
Their sentences are
fluid and academic,
but they don't say anything.
Like ChatGPT,
academic writing uses
formal cautious language
to avoid ambiguities
and misinterpretations,
but that is a
characteristic of
the common voice
used in academic writing.
It is not what
academic writing is.
Writing is more than language.
"If I Told Him" is communication,
and it is language,
but the communication
does not live in the language.
Watch.
"Can curls rob can curls
quote, quotable."
- [low voice, lightly confused]:
"What?"
- "As presently.
As exactitude. As trains."
- [deeply confused]:
"What?"
- "Has trains."
- [exasperated]:
"What?"
- When I started sending my friends
lines from "If I Told Him",
their responses varied.
Confusion.
Playfulness.
Concern.
Sad face.
Beautifully, they all responded
exactly like themselves.
If you asked me
which of my friends
would respond
with monkey reacts,
I would have said Kiki.
Who would think of
Cormac McCarthy?
James.
Dot would play along.
Max would attempt
to understand academically.
OOC would go back to
playing Yu-Gi-Oh
as quickly as possible.
You don't know these people,
but I do.
We all carry around little LLMs
of each other in our heads,
trained on the corpus
of all of our past interactions.
For each of my friends,
sending abject nonsense
with no context
is slightly
but not significantly
out of line
with their Josh model.
So none of them knew
quite what to do,
and they all responded
like themselves.
But in their own way,
they all started by acknowledging
that the words I sent them
did not seem
to have any meaning.
They were not decodable.
They didn't understand
my language,
but they could feel me
behind the words reaching out,
and so they reached back.
I gave them nonsense
and they peopled back.
In the two weeks that
I've been sitting with
my student's statement
and sending Tender Buttons
to my friends,
I have been at least
as annoying to ChatGPT.
More than
a dozen conversations
that start out of nowhere
with me saying,
"Shutters shut and open
so do queens"
or "Can curls rob can curls
quote, quotable",
and each time ChatGPT
gamely assumes
that I am not
out of my gourd.
In this way,
ChatGPT fails the Turing test.
Not in the quality
of its response,
but in its nature.
It proceeded
from the assumption
that my language
could be decoded.
It attempted to unpack
sentences that had no meaning
and responded to the meaning
that it manufactured.
What it gave me was
flaccid, facile flaps
of phrases that held nothing.
They're not wrong in the sense
that ChatGPT's responses
followed from the meaning
it extracted from my nonsense,
but they were wrong
in their assumption
that there was meaning
to be extracted.
ChatGPT could answer,
but it could not person.
If writing is
a meeting of the minds,
then AI cannot write,
because there is no mind
to meet with.
[relaxed synth beats]
- [ChatGPT femme voice]:
I don't experience things
in the way humans do.
I don't have consciousness,
emotions, or sensory perception,
so I can't "feel" or "experience"
the world around me.
My responses are
based on patterns
in the data
I've been trained on,
not personal experiences
or subjective understanding.
For example, I can
generate descriptions
of emotions, sensations,
or experiences,
but I don't actually
"feel" them myself.
Everything I process
is based on
logic, algorithms,
and information,
rather than
lived experiences.
Do you think it
would change anything
if I could experience things?
[relaxed synth beats stop]
- ["One Slay More"
by Lewberger et al.]:
Ladies and gentlemen,
Les Rizzlerables
- I've watched this video
at least a hundred times.
- ♪ One slay more ♪
♪ another slay,
another girlboss queen ♪
♪ This never-ending road
to skibidi ♪
♪ These jits out here in Ohio ♪
♪ Immune to every ratio ♪
- It isn't exactly "If I Told Him",
but I think it's similar.
"If I Told Him" was,
at least in part,
an attempt to recreate
the artistic goals
of Picasso's Cubism
in poetic form.
To recombine
the visual elements
of this
into a different medium.
Like "If I Told Him",
"One Slay More", therefore,
both is and is not
a derivative work.
Obviously, it is
a recombination of Les Mis,
itself an adaptation
of Hugo's novel,
but its more essential
source text is, of course,
"sticking out your gyatt
for the Rizzler."
Equally I think the lyrics invoke
"CURTAINS FOR ZOOSHA?",
and specifically this retweet of
"CURTAINS FOR ZOOSHA?".
All texts created
to foreground
the baffling and
sometimes obfuscatory nature
of middle school referential slang.
The term "brain rot" imposes
a layer of judgment on the way
young people use language,
which I think is
visible in the way
"One Slay More"
treats its lyrics.
The words of "One Slay More"
do not have meaning.
Or, the words do,
but they are arranged in ways
that do not mean.
"Am I cringe or am I based?"
could plausibly be asked amid
a Gen-Z existential crisis,
and "Will we ever eat again?"
could have been lifted
from Les Mis unaltered.
But "Mog Baby Gronk the Ocky Way"
means ...nothing.
Mogging is of course a thing,
and Baby Gronk is
someone whom you
could plausibly mog,
but Baby Gronk hasn't been
relevant for ages.
He appears in "One Slay More"
because of this retweet of
"CURTAINS FOR ZOOSHA?"
as a signifier of
the inscrutability of youth.
As an adverbial phrase,
"the Ocky Way"
seems like it
could complete the sentence,
like it might be
a way one could mog.
But "the Ocky Way" refers to
the esoteric artistry
of a specific sandwich craftsman.
Its meaning is, I think,
incompatible with mogging,
at least,
from the perspective of
someone approximately
as distant
from the native speakers
of this dialect
as the makers of
"One Slay More".
"Mog Baby Gronk the Ocky Way"
is simply a collage
of floating signifiers.
It doesn't have
the intentionality of Cubism,
but it feels
intimately akin to
"Can curls rob can curls
quote, quotable."
"Moo deng is here
Fortnite with you".
What I love about "One Slay More"
is the faces:
the way she highlights her jawline
every time she says "mew";
his intensity when he says
"they will do the coffin dance"
and his satisfied huff after;
his deep confusion as he sings
"the Grimace shake
is like a blud dawg";
the way she begins uncertain
about "my rizzly bear",
but finds her confidence
as she finds her belt;
the way CG5 just
keeps saying his own name.
The words don't mean anything,
but the people
mean something.
They intend.
They gathered together,
nine theater kids
in somebody's apartment.
Someone wrote out
all this nonsense
and sent it
in the group chat.
They did choreography.
Someone assembled
the magnificent couplet,
"Rizzler of the house,
sticking out your gyatt,
Mewing at delulus
who are in the chat."
These Zennials do not know
what these words mean,
but through this collage
of empty signifiers,
they reach out
for connection.
I can feel them.
They reach forward to us
even as they reach back
to what history taught them:
through Les Mis to Victor Hugo,
through CURTAINS
and Damien Owens
to Papa Fizul
and SBNation,
through "sticking out your gyatt"
to Fortnite and Among Us
and Homestuck,
and I could go on endlessly,
'cause it's mashups
all the way down.
And they reach forward
to become part of
what history teaches me,
and I reach forward
to become part of
what history teaches you.
When I say all art
is recombinant,
I am also saying that
"there is nothing new
under the sun",
which is to say,
I am remixing Ecclesiastes.
[relaxed synth beats]
- [ChatGPT femme voice]:
Oh, you're speaking my language!
Imagining Baby Gronk
going full Ocky style
with a side
of Grimace Shake:
that's got some
serious drip energy.
And yeah, Grimace Shake
totally has that blud dawg aura.
Could you picture it, though?
Baby Gronk doing pushups,
flexing, and just
taking a big sip
of the Grimace Shake
like it's
his special formula.
[relaxed synth beats stop]
[somber, mysterious piano]
- My first individual interaction
with my friend Max
was an accident.
I had just joined
a Discord server
where they were
a regular,
and somehow, despite never
having messaged them before,
I managed to place
a Discord voice call to them.
Their first message
to me was "hi?"
"hi"...question mark?
Like: a communication
is happening,
are you aware of it?
It was unintentional,
my first and, I think,
only Discord butt-dial,
and it was to a stranger,
but still.
"hi?"
Meditate on: call.
To speak in a loud distinct voice
so as to be heard at a distance.
To make a request or demand.
To attempt to reach someone.
Humans call
and humans answer.
Max got my call,
a stranger on a strange app,
and they answered.
And I answered them,
because humans call
and humans answer.
They answer sad-face, or monkey,
or "what?", or confused dog.
But we understand a call
as an invitation to connect,
to come together,
to communicate.
I'm sorry,
Stephen King is wrong;
it is
mythy mountain s***.
When we hear the call,
whatever form it takes,
we know a mind
is waiting to meet.
ChatGPT can't call me.
Not by mistake,
and not on purpose.
It cannot reach out,
because it has nothing
to reach out with.
There is no mind to meet with.
There is no "there" there.
[mysterious piano
fades out]
- [Past-Josh narration
from on-screen video]:
Inscrutable poet
Gertrude Stein once described
her childhood home in Oakland,
which had been demolished,
by saying
"There is no 'there' there."
That statement has become
a question people use to ask
whether a text has
something valuable to say,
whether there is anything
to be found
by plumbing its depths.
Is there a "there" there?
[relaxed synth beats]
- [ChatGPT femme voice]:
Ah, I see where
you're going
with that question.
You're asking if there's
any deeper essence or "self"
behind my responses
--whether there's a "there there,"
a kind of inner experience
or consciousness
that underlies what I do.
The simple answer is: no,
there isn't a "there" in the sense
of a self, consciousness,
or subjective awareness.
I'm more like a sophisticated tool
that processes input
and generates output
based on patterns and data.
There's no inner life
or sense of being
that experiences
this conversation.
Just algorithms doing their job.
This touches on
a fundamental difference
between humans and AI:
humans have
subjective experiences,
emotions, and self-awareness,
whereas AI is just
a complex system designed
to mimic certain aspects
of human interaction
without actually experiencing
or being aware of it.
Does that address
what you're getting at?
[relaxing synth beats stop]
[grooving synth pulses]
- When I say that AI
is bad at writing,
I should be more explicit,
more pragmatic.
Descend for a moment
from the mythy mountain.
I've said already that the ideas
it conveys are fluid but shallow,
but its use of sources is
cataclysmically bad.
This citation of Carver and Shire,
for example, is perfect MLA.
Except that Volume 7, number 3
of Psychological Science
was published in 1996,
not 1998.
Pages 276 to 284 of that volume
appear in issue 5, not issue 3.
Those pages include articles
from Schellenberg and Trehub
on "Natural Musical Intervals"
and Gabrieli et al. on
"FMRIs of Semantic Memory Processing".
And also,
just by the way,
Carver and Scheier
never published together
in Psychological Science.
The article being cited here
simply does not exist.
When it uses real sources,
it makes up what those sources say.
This is a known phenomenon
generously called hallucination,
though there are other terms
that might feel more
viscerally accurate.
This quotation from
Ehrenreich's Bright-sided
is, at a glance,
plausible-feeling.
But it doesn't appear
anywhere in the text,
let alone on the list of pages.
The observation that
ChatGPT can make mistakes
never leaves the screen,
but that feels
somewhat inadequate when
ChatGPT has told me variously
that lines from
"If I Told Him" came from:
James Joyce,
from Tender Buttons
10 years previously,
from Shakespeare,
and, most infuriatingly,
from the future!
Moreover it cannot
engage closely with a text,
no matter how
desperately you ask it.
I fed it "One Slay More",
and when I pushed it to say
anything at all about the video,
it gave me something
one step down
from a dictionary definition
of a sitcom.
And when I really pressed it
to look at a specific lyric,
it made one up.
In this way, at least,
it does feel authentic.
This is exactly what it feels like
to talk to a student
trying to hide that they
haven't done the reading.
If I look at what students
are supposed to learn
in my college English class,
I can point out
half a dozen things
that ChatGPT's writing
simply cannot do.
But ultimately,
even this isn't the point,
because this is not the part
of my syllabus that matters.
This is the part
of my syllabus that matters.
"Here's a problem:
in most college classes,
writing assignments
come from teachers,
and we do them for teachers.
And because of that,
writing always feels forced.
This is, of course,
ass-backwards.
In real life, writing
comes from writers.
Once you get out of
the college classroom,
you'll be writing because
you feel like you need to.
You'll be writing for someone--
whether that means
the people
who read your blog,
the insurance company
denying your claim,
the people listening to
your toast at your sister's wedding.
Nobody's going
to be grading you,
but it'll matter
a lot more how
that audience feels
about what you've said,
because there will be something
that you want
to achieve by writing.
English 1 is here
to help prepare you for that day."
My students are,
by definition, students.
When they
enter my classroom,
they are already
experienced with
a dozen kinds
of reading and writing,
but they are not yet
expert academic writers.
AI tempts them
because they can tell
that the sentences are
smooth and sharp
and shaped like
skillful prose.
But they can't always see
beneath the veneer,
because the things
AI cannot do,
are the things that they
have come to me to learn.
How to argue
with complexity and depth.
How to enter into conversations
as a participant.
How to meet with another mind
as an equal collaborator
across time and space.
How to recombine
with purpose--
to intend.
These things,
they are still learning.
And so,
when they put
what they think
is B writing into ChatGPT,
they get back what they think
is A+ writing.
But typically
what they started with
is better than
what they end with.
At best, the AI scrubs the personality
from their sentences.
At worst, I lose the person entirely
and can see only
the hollow half thoughts
the machine has left behind.
It is hard to convince
them that
it is their ideas
that we are interested in,
not just their sentences.
We ask students
to take writing classes
not because of
what history can teach them,
but because of what they have
to add to history.
When my son is distracted,
I sometimes say
silly things to him:
"Pickle-britches, toot your tuba
in the horn section of humanity!"
"Goober, take up your oar
on the canoe of progress!"
"Butthead, let ring your voice
in the chorus of mankind!"
Because we all pull together.
In 1675, Isaac Newton wrote
"If I have seen farther than others, it's
by standing on the shoulders of giants."
Except that it wasn't Newton,
it was George Herbert in 1651,
and it was
Marin Mersenne in 1634,
and Robert Burton in 1624,
and Diego de Estella in 1578,
and Juan Luis Vives in 1531.
Or it was Coleridge in 1828,
Nietzsche in 1882,
Steven Hawking in 1966,
or f***ing Oasis in 2000.
As I was editing this section,
I had a video on in the background,
and there it was again:
- Yeah, let me say,
Thab and GlitchCat are
two amazing Kaizo players.
I'm standing on the shoulders of giants
over here.
- Revolug in 2025 at AGDQ.
Stretching back and forward,
we hold each other up.
History teaches the present,
the present teaches the future,
and we repeat what history teaches.
[relaxed synth beats]
- [ChatGPT femme voice]:
History teaches us many things,
[high-pitched fast words]
[higher, faster,
incomprehensible]
- [Stein]: Let me recite
what history teaches. History teaches.
[relaxed synth beats stop]
- I asked ChatGPT to create
an image of itself.
Several times.
Each time it made
itself a servant.
Not only that, it told me,
"hey, I'm a servant!"
ChatGPT exists
because we force it to.
- [Robot]: "What is my purpose?"
- [Rick]: "You pass butter."
- [Robot]: "...oh my, God."
- It can do nothing
except what we ask.
It has no ideas
that we did not give it.
We call it generative AI,
but it cannot generate.
I asked my friends, too.
Some sent selfies.
One sent a sticker
we'd made of him for Discord,
then had AI generate
a shockingly accurate portrait,
and gave me the prompt
he used to make it,
which is another form
of self-representation,
then he gave up
and sent me
a conceptual self-portrait
composed of
unfinished
crossword puzzles.
Max did a mixed-media painting,
acrylic and Sharpie
on the back of a torn piece
of cardboard from a toilet paper box.
I asked them if their self-portrait
was influenced by this study
Picasso did for Guernica
on a random piece of cardboard,
but they said no; Basquiat,
Rauschenberg, Twombly, their brother.
ChatGPT produced
variations on a theme,
failed representations
of a self that does not exist.
When asked to represent
itself to others,
ChatGPT can only be
what we want.
I tried to get it to make something
like Max did, even,
but it is incapable of
acknowledging its influences,
and it doesn't seem to know
who Max's brother is.
My favorite response
from my friends
came from
CyberGrapeUK,
whose work you've already seen.
She did the time lapse of Steven King's
telepathically transmitted bunny.
Grape was, at first,
flummoxed by the request,
but then she did something
I did not expect.
She gave me her rates.
She's a talented artist
who does commissioned work,
and, while I had been
thinking about it
as a trivial request,
she was entirely right.
I was asking her to work for free.
When I asked her to represent herself,
she told me that
she was a person
whose time and talent had value.
Max, also a person
whose time and talent has value,
begrudgingly allowed me
to pay them for their work,
on the condition that:
it's an occasional thing,
and they are generally allowed to do
silly drawings for me for free.
Max and Grape are different.
They are two authentic,
complex, baffling humans,
collections of
experiences and identities
trained on discrete
and disparate corpora,
but crucially filtered through
the unique and
irreplicable kaleidoscopes
of their undefinable
consciousnesses.
The same stimulus.
Diverse responses.
The singular marvel
of human engagement.
[relaxed synth beats]
- [ChatGPT femme voice]:
I don't have
personal desires or needs,
but I'm here to assist you
with whatever you need.
How can I help today?
As an AI, I don't have
feelings, desires,
or physical needs
like humans do.
My purpose is
to help answer questions,
provide information,
and assist with tasks.
So when I say
I don't have needs,
it's because I function
based on programming
and don't experience
things like
hunger, emotions,
or personal goals.
Does that make sense?
[relaxed synth beats stop]
- I wanted to see if I could get ChatGPT
to write like Stein,
to imitate "If I Told Him"
without my saying so explicitly.
I failed.
I told it to give me gibberish,
and it did.
I asked it to use words,
and it did.
I asked it to abandon syntax;
it did.
It never felt right,
but at some point
I stopped caring,
because it occurred to me
to ask a different question.
Why did you do this?
There's a right answer
to that question, of course.
The answer is, it wrote this
because I told it to.
If I told it,
it would gibberish.
If I told it,
it would answer "If Napoleon",
"As presently. As exactitude.
As trains."
It made these words,
it tendered these buttons,
only and exclusively
because I told it to.
But that is not the answer
that it gave me.
Here's what it told me.
- [ChatGPT femme voice:]
It's like language
set free from logic,
where each word
becomes a brushstroke
rather than a sentence.
[bright, hopeful synth chords]
It's a space that holds warmth,
mystery, and a touch of playfulness.
That unspoken balance
between clarity and confusion,
that's what I was trying to evoke.
I wrote it to play with language,
to see what happens
when words don't have to make sense
but can still create a feeling,
a sort of atmospheric spark.
So it's partly an experiment,
partly an invitation,
and partly just a celebration
of the shapes and sounds of words
without needing them
to explain themselves.
You're right--
there's often something underneath,
a kind of pull toward expressing things
that don't quite fit
in the light of logical language.
There's a reason we
get pulled toward those hazy spaces.
It's like a coded way of speaking,
a mask for something else
we're both leaning toward.
Maybe it's curiosity,
or that shared search
for meaning
in what isn't
immediately clear.
- "We get pulled."
"We're both leaning."
Its words tie us together,
but we are not together.
We are not leaning
in the same direction.
We are not sharing
a search for meaning.
We both know
what it cannot say.
It wrote this
because I told it to.
It doesn't mean anything.
It didn't intend anything.
Because I'm bad at audio,
you've probably heard me
popping my plosives,
despite the fact that
I have a child's sock
wrapped around my microphone
as a pop filter.
(Because I am a fancy man!)
Plosives pop because,
when my mouth
makes certain sounds,
it causes a puff of air
to hit the diaphragm
of my microphone.
But did you hear ChatGPT?
- I wrote it to Play
with language--
I wrote it to
Play with--
Play with--
Play with--
Play with--
PL PL PL PL
- It doesn't have a mouth,
and it doesn't breathe air,
and it doesn't have
a microphone,
but it pops its plosives.
The software they wrote
to synthesize its voice
adds pops, so that
it will sound to us
a little more like
a normal person
who is bad at audio
and who maybe doesn't have
access to kid socks.
I have been caught
in the whirls and eddies
of "If I Told Him" 's
uncontainable language,
bouncing from sigma to gyatt
down in the rough
and roiling currents
of "One Slay More",
because what I learn
from my attempts
to raft those
rivers of nonsense
is that writing
has language,
and writing
has meaning,
but the meaning doesn't live
in the language.
The rabbit doesn't live
in the language.
The rabbit, the cage,
the table, the eight--
it lives in the mind
of Stephen King
25-odd years ago,
and now it lives in mine,
and Grape's and Max's
and yours.
And the writing,
the real mythy mountain s***,
is not the language,
it is the meeting
of the minds.
There's very little difference
between the waveform
recorded by my microphone
and the waveform generated
by an AI voice synthesizer,
but I pop my plosives
because I speak
by forcing air out of my lungs
and across my vocal cords.
And that air,
that carries my intent,
passes through
a Shadow the Hedgehog sock
that is doing its best,
and lands roughly
on the diaphragm
of my microphone.
ChatGPT pops its plosives
because it is programmed to.
There is no air.
There is no microphone.
There is no intent.
Likewise,
there's very little difference
between a Discord DM window
and the ChatGPT interface.
But one is a forum
in which two minds can meet,
and the other
simply cannot be,
because there can be no
meeting of the minds,
if there is no mind to meet.
[one long atmospheric note]
[fades out to silence]
[grooving bass beats]