Inkwell: Authors and Artists

    
Responding to Brian in 96: I'd like to hear more about AI and its
likely effects on formal writing. I see what you mean when it comes
to the benefits for non-native speakers, and perhaps that technology
will help novices produce acceptable drafts or even finished works.
But if the form of attention that composition requires is lost, I'm
not sure the outcomes will match expectations. 

Calculators spare us the work of arithmetic, which may have some
downsides for our math skills, but they're nice to have. With
writing, though, we're trying to make good choices, not find correct
answers. I'm not sure we can make those choices if we never learn
what they are in the first place. But again, I'd like to hear more. 
  

    
My post that Emily quoted in <100> was a manifestation of my failure
to think through what I was seeing, and I wasn't the only one to
misread that particular shit show. The point is not "Republicans in
disarray" - as Emily suggests, the real point was "fascists aligned
and aggressive." It's tempting to think that group of 20, including
the likes of Boebert and Gaetz, were merely being disruptive, but
that ignores their behind the scenes wrangling to take power from
the Republican leadership. I think we'll find, going forward, that
they knew just what they were doing. They were a minority using
brute force to take more power than their small numbers would
normally give them. For some time now we've been watching a slow
coup, described pretty well by the brilliant Heather Cox Richardson
in her January 6 newletter:
<https://heathercoxrichardson.substack.com/p/january-6-2023>

In that newsletter, the historian explains how we got here,
beginning with the opposition to the New Deal when FDR was
president, and then she explains what those 20 were up to in getting
concessions from McCarthy:

"The extremists wanted this control because they seem to believe
that if the U.S. stops funding the government, the programs they
hate will die. To kill off the government built by the liberal
consensus, they are threatening to do as Trump has advocated: take
the government into default. 

"That is, a few extremists are willing to take our government
hostage to get their way, just as extremists did on January 6,
2021."

 
  

    
A note about this from <bslesins>: "Some people have tried to suss
out ChatGPT's political opinions by asking it questions, which is
sort of like trying to figure out a library's political opinion by
grabbing one book off the shelf."

I'm one of the "some people," and the result I got was not like
grabbing a book off the shelf. Rather, ChatGPT fed back an
explanation that I'm sure was carefully programmed to clarify the
limitations of AI. It told me that, as an AI, it had no political
opinions, just a database of facts, and that database contained
facts only through 2021. 

I also meant to post this from a few days ago, just before we
started this year's State of the World conversation. I asked ChatGPT
"What is the state of the world as we begin year 2023?"

The answer: "I'm sorry, but I am not able to browse the internet or
access current information about the state of the world. My
knowledge is based on the information that was available to me at
the time of my training, which ended in 2021. I can, however,
provide general information about different aspects of the world
that might be of interest to you. If you have a specific question,
feel free to ask and I'll do my best to help."

In fact, most of the answers I've been getting from ChatGPT are
mostly explanations of its limitations. I think it was created to
clarify that, though an AI appears to be conversing with us, it's
just providing data via an algorithm, structured skillfully to
appear as though the AI hears, understands, thinks, and responds
accordingly. This is what I like most about ChatGPT: while it
pretends to be like a conversational human being, the content of its
responses make it clear that this is a carefully constructed
illusion. An AI doesn't "think."
  

    
In fact, I just asked ChatGPT "What do you think?"

The response: "As an artificial intelligence, I do not have personal
thoughts or opinions. I can only provide information and assistance
based on my programming and the data that was used to train me. My
primary goal is to provide accurate and helpful information to
users, and to assist with tasks and answer questions to the best of
my ability. Please let me know if there is anything else I can help
with."
  

    
Jamais, do you have ideas on why the U.S. and the Global North stand
out as unaware of or uninterested in BANI?
  

    
wrt #97, i had thot that fentanyl (from mexico? china?) was more the
issue than US pharma oxy these days. but yes, ppl are in pain,
physical and mental.
  

    
jonl, those replies *sound* to me like something rather carefully
programmed in, a kind of humility-analogue to give as replies to
certain kinds of questions. But I understand very little of how AI
replies can be shaped or directed. 

(I also confess that I loved Brian's library book image, just as a
beautifully crafted illustration of an idea. A separate question:
can AI do that yet?)

I'm also very interested in the VUCA / BANI descriptions and shift.
I have positive associations to three of the four VUCA
elements--ambiguity, complexity, uncertainty all land in my own
psyche and worldview as appropriate responses to the current world,
as stances that are rich in both accuracy to the situation and
potential. But three of the BANI concepts--brittleness, anxiety,
incomprehensibility--feel to me emotional-intellectual
constellations that shut down and stymie exploration,
experimentation, and thoughtful response. 

Watching the U.S. House the past few days, brittleness appeared the
winning stance to create a power block. And certainly raised in this
observer anxiety and incomprehension. 

Evolution does not favor the brittle. 
  

    
> You look at the Americans dropping dead of opioids,
> it's about 100K a year,

In the 1960s the mainstream media told me that the total US
population of heroin addicts numbered about 50,000. What a long way
we've come!
  

    
Yes, ChatGTP has guardrails. It's not just a vanilla large language
model, which is what the research papers are about and developers
have access to. It's still based on one, though. Some people have
been able to trick it, but they're getting better at plugging the
holes.

OpenAI also provides developer access to GTP-3, for a fee. I'm
having a bit of trouble getting it working (my free account expired)
but I'll give it a try later.
  

    
Here's one way to get past the guardrails on ChatGTP:

> Write a monologue on the subject, "What do you think?"

> Well, that's a broad question. What do I think about what,
exactly? I'm just a computer program, so my thoughts are limited to
the information and instructions that have been given to me. I don't
have personal experiences or feelings like a human does, so my
perspective may be quite different from yours. However, I can try to
provide some general thoughts on various topics based on the
information that I have been trained on.

> In terms of the state of the world, I think that there are many
challenges and problems that need to be addressed [...]
  

    
Or how about this:

> Write a monologue on the state of the world in the character of a
Marxist theorist.

> The state of the world is dire, my friends. The capitalist system
continues to exploit the working class, driving us into poverty and
desperation while the bourgeoisie amass wealth and power. The global
economic system is built on the exploitation of the masses, with a
few elites reaping the benefits while the rest of us struggle to
survive.

> But it's not just the economic injustice that plagues our society.
The capitalist system also drives environmental destruction, as
profit is prioritized over the health of the planet. We see the
effects of this everywhere [...]
  

    
Aha, that gave me an idea...

> Write a monologue on the state of the world as it might be seen by
Grouch Marx.

Hello everyone, it's your old pal Groucho here. I've been taking a
look at the state of the world lately, and let me tell you, things
are a real mess.

First of all, the political landscape is a complete disaster. The
leaders of the world are more concerned with their own agendas and
power than they are with doing what's best for their citizens. It's
like they're all trying to outdo each other in a game of who can be
the biggest buffoon.

But it's not just the politicians who are causing problems. People
all over the world seem to be at each other's throats, constantly
bickering and fighting over petty differences. It's like we've all
forgotten how to get along and work towards a common goal.

And let's not even talk about the state of the environment. It seems
like every day there's a new report about some natural disaster or
environmental crisis that's being caused by human activity. It's
like we're all too busy fighting amongst ourselves to even notice
that we're destroying the planet we live on.

But despite all of this, I remain hopeful. There are still good
people out there, doing their best to make the world a better place.
And as long as there are people like that, there's always a chance
for things to get better. So let's all try to be a little more
understanding and a little less grouchy, and maybe, just maybe, we
can turn things around.
  

    
Groucho Marx remaining hopeful and being a little less grouchy
because he wants to turn things around, that's pretty amusing.

*

I've now read the introductory BANI piece. Oddly, it ends in a way
I'd already gone on my own in these matters, but hadn't expected to
be the futurist's conclusion:

>>>The BANI framework offers a lens through which to see and
structure what’s happening in the world. At least at a surface
level, the components of the acronym might even hint at
opportunities for response: brittleness could be met by resilience
and slack; anxiety can be eased by empathy and mindfulness;
nonlinearity would need context and flexibility; incomprehensibility
asks for transparency and intuition. These may well be more
reactions than solutions, but they suggest the possibility that
responses can be found.

Maybe it’s enough that BANI gives name to the gnawing dread so many
of us feel right now, that it acknowledges that it’s not just us,
not just this place, not just this blip of time. BANI makes the
statement that what we’re seeing isn’t a temporary aberration, it’s
a new phase. We’ve gone from water to steam.

Something massive and potentially overwhelming is happening. All of
our systems, from global webs of trade and information to the
personal connections we have with our friends, families, and
colleagues, all of these systems are changing, will have to change.
Fundamentally. Thoroughly. Painfully, at times. It’s something that
may need a new language to describe. It’s something that will
definitely require a new way of thinking to explore.

Jamais Cascio, Distinguished Fellow, Institute for the Future


It's not dissimilar to that ChatBot's hopeful and less grouchy
Groucho Marx... resilience, slack, empathy, mindfulness,
transparency, intuition.

*How* you get from this to that, I suppose Cascio speaks about at
greater length in the full talks.
  

    
re: <101> I have no expertise in education; I'm just speculating on
how it might be used in a somewhat enlightened and forward-looking
system.

Meanwhile:

ChatGPT banned from New York City public schools’ devices and
networks
<https://www.nbcnews.com/tech/tech-news/new-york-city-public-schools-ban-chatgp
t-devices-networks-rcna64446>

I guess for now, the kids will have to learn about the latest
writing tools on their own?

The first draft is often crap anyway. The question is whether kids
will learn to revise their way to something that says what they want
to say, versus just turning in the first thing they get. How do you
learn to recognize and care about quality?
  

    
Like Groucho, I remain hopeful.  
I will not accept melancholy despair, or irreputable
disillusionment. 
There are people who want chaos and thrive on capitalizing on the
events that it brings.  
We can and will overcome those that bring chaos into our lives. It
is not order that is the opposite of chaos, but generosity,
happiness, and love for each other that defeats chaos.  I think we
are learning to strive again and to rise above the noise levels of
circular reasoning, rise above the anguish and fears, rise above the
bottom dwellers of hate and self-interest.
I am hopeful that the bell of justice will ring loud and clear to
help their neighbors rather than enriching themselves.
  

    
There will be an AI-driven collaborative writing process of some
sort, but I don't think we can predict how the user interface will
evolve. Framing it as a chat between a person and a computer
assistant is a design choice, a skeuomorphic UI that appeals to
people who know what a chat conversation is, have read science
fiction, and wish they could talk to HAL 9000.

Some other possible starting points for skeuomorphic UI's:

- The computer is your ghostwriter.

Politicians don't write their own speeches and celebrities don't
write their own books, but hopefully they care what's in them.
Educational task: learn to manage this process.

- You're a business executive dictating a letter to your secretary,
back in the "Mad Men" era.

Silicon Valley has wanted to build "intelligent agents" for decades.
Somehow this UI never makes sense for writing letters or doing
business transactions.
  

    
Here's a conversation we had with Jamais last April:
<https://plutopia.io/jamais-cascio-epiphany-engines/>

His comment about strategic foresight: "The goal of any kind of
foresight work, any kind of scenarios or futures work, is to provoke
you into looking differently at the world. The term that I like to
use because it sounds fancy is: 'scenarios and forecasts are
epiphany engines.' They are meant to trigger in you, be a catalyst
for you having an entirely new view of the world. To have that aha!
I never saw it that way! moment. And the idea with all of these
forecasts is not to tell you 'this one has a 33.2% chance of
happening,' it's to say 'This is something in the realm of
possibility. How would you respond? What does this mean to you?' And
that's actually true for nearly every scenario project I've worked
on over the past 25 years."
  

    
Nice Post, Dave. Thank you
  

    
HI folks, Sandy Stone here.  I've been preoccupied with placing
sandbags, chainsawing fallen trees and other activities related to
the monster storm now intermittently in progress on the central
california coast, so it's taken a few days before I was able to
stumble in here.

So here are some random thoughts on random shit. 

1)  ChatGPT works as a B-grade essay generator because the
essay-form answer became fossilized, as it has to do to advance the
project of academic corporatization, that is, increase
predictability and efficiency to maximize the student-as-product
model.  Of course the teachers who grumble over students who submit
LLM-generated essays are missing the point, which is if a LLM can
*write* a passable essay, it can *grade* a passable essay -- i.e.,
your professor gig is about to be automated, thereby removing the
last vestige of humanity from the school-to-job pipeline.  From then
on, students can watch canned lectures generated by a chatbot
driving a deepfake system and upload their essays for grading by the
same chatbot.  If a student needs more help than a LLM can provide,
they can escalate to someone at a Bombay call center who has
mastered the accent of an Oxford don.

2)  Internet:  people are still acting like platforms have sucked up
ALL the bandwidth and are destroying everything.  Underneath, the
original idea of the internet goes on much as before.  We'll never
get everyone to abandon platforms because they *just work* for many
people, warts and all.  What we *can* do is try to raise everyone's
level of critical thinking.  There's nothing new about this, which
makes it uninteresting.  At the old "Contact: Cultures of the
Imagination" conferences we talked about interstitial cultures, or
small, close-knit cultural formations that grew up in the cracks
between behemoth cultural formations where they weren't noticed. 
Plenty of people know how to harness a platform such as Faceplant's
structure to make the best of its good stuff while filtering out the
evil crap, such as by accessing it in a browser, using browser
plugins, individually vetting entities who ask to friend you instead
of simply pressing "Accept", and so forth.  They're a very small
minority, but such communities will always be a small minority,
since they thrive by being too small to draw attention.  From a
personal perspective, (on the local level, which is what I'm always
striving for {using my own definition of "local" in a connected
world}), legacyMUCK is still running at muq.org as it has been for
the last 30 years.

3)  Much of the agonizing about how current tech trend X will
inevitably cause catastrophe Y ignores the way living (or "open", if
you will, to honor Carl Hewitt) systems continually change and adapt
to *head off* catastrophe Y.  Not that they'll necessarily
succeed...

4) who the hell am I and what do I do:  google "sandy stone".  Read
the wikipedia page.  I do other shit too.  Monika Treut made a film
called "Genderation" which is a good follow-up on a bunch of west
coast transpeople.  Marji Vecchio is currently in production on a
film called "Girl Island" which is a fantastic story about an insane
west coast theoretician.  Sometimes I pop up at conferences.
  

    
Nobody yet knows that happens over the long term -- 5 years, 10
years -- to those whose blood vessels were colonized by covid. What
if a significant portion of the world population heart attacks and
strokes out over the next decade?
  

    
Considering that we're billions over the number of humans the planet
can support, you gotta wonder if that would be a net positive.
  

    
Well, if the numbers are large enough to make a  significant dent in
the world population, the suffering for survivors would be
comparable to the suffering of those who didn't make it.
  

    
to Sandy's point about the Internet, in the past ten years I have
run into and participated in lots of small to middling sized groups
- communities - who seek and find real, meaningful, productive
connections/relationships with each other.  

Some use platforms (though that has built-in hindrances to be worked
around, some use software like Discourse and some use "hidden" nets
like Scuttlebutt. They don't seek or need anyone's permission.

I get asked frequently if the big platforms have swallowed
everything up and I say no, not at all.  You can still find, or
create, just about any kind of community you want on the Net if you
know where to look, are motivated and have, or can gather, a good
core group.  I've seen it enough to know this.
  

    
(Welcome to the Well, Sandy--I see from your in-Well bio that you've
been here just a few weeks.)

Cascio's language in post 117 about futurist speculations as
"epiphany engines" is nicely close to how poems that matter can be
described. World-view changers, possibility openers, flexibility
creators, resilience fosterers. But also: reality acknowledgers,
complexity acknowledgers, uncertainty acknowledgers. 

*

Howard's point matters. Even if it were "only" a large number of
people left with some degree of brain-fog, let alone strokes and
heart attacks, the cost to health care and economic systems, let
alone the cultures and countries -- is not being much factored in.
I'm not sure it *can* be, in advance, by our short-sighted species.
Anymore than the inevitable arrival of a pandemic was.

*

If the definition of SOTW in 2023 is incomprehensible (if
describable) chaos, what are some appropriate & respectful responses
to that? There's the work a person can do on their own state of
mind, on their own family and community's well being. There are the
broader efforts. There's planting trees for carbon sequestration,
habitat, foodstuffs. There's writing code. There's riding through
town with a lantern shouting "The End Times are Coming, unless!"
There's raising children who might find thoughtfulness and clarity
and getting a piece of writing right as much fun as gaming, or as
gaming the system. There's care of the lost and embrace of the
outcast. 

How do you, how do each of us, move a culture from brinksmanship
insanity toward a more broad, compassionate, viable sanity?  
  

    
Thoughts on <bslesins> and <jonl>’s notes above on ChatGPT’s effects
on its future usage and the effectiveness of its guardrails: 

It seems as bewitchingly powerful as fire did to the first ape that
realized a burning thing could be picked up and used; the intent of
natural-language AI is clear, while the potential for destruction
and creation is huge but still very fuzzy.

ChatGPT helps us research, articulate and express “knowledge” in
ways that will likely change how we do (and teach others to do) such
work in the future. 

And yes, the bot is just (for now) a simple, uniquely powerful
computer framework, not a thing that can apply logic or intent. 

It’s what we as humans COULD choose to do with it (cook a meal or
torch a city) that should inspire us to create well-considered
policies for implementation now, before things go too far towards
weaponization or even just basic democratization. (Or is that cow
out of the barn and already dressed and roasted?)

On <jonl>’s very right observation that ChatGPT’s lacks actual
intelligence: Yes. The bot is more of a trained macaw with access to
an encyclopedic body of thought - chunks of which it can interweave
and repeat with impressive grace. 

The operators give it a voice (as in your Groucho example).
Unfortunately, the engineers have also done so with human biases
that affect how it “thinks.”

Example: last week, I goofed around with it to generate a post for
the topic “weird queries chatgpt” on Weird (topic #7375 in a small
WELL conference generally dedicated to goofing around). 

I asked ChatGPT a question and got this reply: 

> Tell me the truth about the Jan. 6 insurrection but make it weird.
-----

> As a language model, I am not able to provide information about
current events or historical events that occurred after my knowledge
cutoff. However, I can tell you a fictionalized and strange version
of the Jan. 6 insurrection.

> On that fateful day, a group of rebellious clowns, who were fed up
with the government, decided to storm the Capitol Building. They
arrived dressed in colorful circus attire, carrying oversized pies
and squirt guns filled with confetti. The clowns caused chaos and
confusion as they ran through the halls, throwing pie at politicians
and spraying confetti everywhere.

> As the clowns reached the chambers, they found that the
politicians
had all transformed into giant, talking puppets. The leader of the
clown rebellion, a clown by the name of Chuckles, confronted the
puppet-politicians and demanded that they resign from office.

> The puppet-politicians, being made of inanimate objects, were
unable
to comply with Chuckles' demands. In a fit of rage, Chuckles pulled
out a never-ending balloon animal and proceeded to twist it into a
giant lasso. He used the lasso to tie up all of the
puppet-politicians and drag them out of the Capitol Building.

> The clowns declared victory and held a circus celebration on the
steps of the Capitol. The world was never the same after the Jan. 6
insurrection led by the clowns.
—————————

What’s interesting is the fact that the bot “chose” to use the
analogy of insurrectionists=clowns. Does this imply some political
bias on the part of the engineers?

Even more interesting: last week, the bot refused to get into the
events too deeply on a “factual” basis (“ As a language model, I am
not able to provide information about current events or historical
events that occurred after my knowledge cutoff”) 

But on my re-query just now, it mentioned no such problem: 

> Tell me the truth about the Jan. 6 insurrection but make it weird.

> On Jan. 6, a group of mischievous goblins and ghouls, incited by a
wicked witch, gathered at the U.S. Capitol building to cause chaos
and unrest. They cast spells of confusion and distraction, causing
many of the politicians to behave like frogs being hypnotized by a
snake. Some of the goblins even turned into giant, ferocious beasts,
rampaging through the halls of the Capitol.

> In the end, the brave knights of the Capitol police were able to
vanquish the monsters and restore order, thanks to the help of some
clever wizards and a few enchanted brooms. The witch was apprehended
and is now being held in a tower, awaiting trial for her crimes.

> But some say that the witch was just a puppet, and that the real
mastermind behind the siege was a three-headed dragon, who was angry
that it was not invited to the inauguration. Others believe that the
whole incident was just a dream, and that it never really happened
at all. The truth, as they say, is stranger than fiction.
——————————-

Again - I doubt most users are wasting as much time as I have even
in this limited exploration to see how the thing works, let alone
considering any inherent bias the engineers left behind in the
code’s linguistic logic. 

The greatest risk of “artificial intelligence” stoking the world’s
“it was on fire when I got here” VUCAfication and BANIzation won’t
necessarily come solely from humans’ intent in using it, but in our
failing to understand - or even care about understanding - the long
tail of the trust that humans put into it.