Hmmm. I was talking about the argument that AI can enhance creativity with a very insightful young, 100% human person in her 20s who works in a technology field. She made this analogy. Please do not be offended.
"This sounds like when a man thinks a stripper is really in love with him."
"If you read that post, and that’s your impression, I won’t be able to persuade you otherwise." It's easy to tell a human being something along the lines of, "I can't change your mind so I won't try." I am asking if you ever feel the impulse to give up on an LLM?
Are you saying that when you read the post, you think my thinking hasn’t been stimulated, that I am just imagining it has, à la a man thinking a stripper loves him?
Excellent piece. I just signed a freelancer's agreement at a magazine that requires that there be no generative AI in the writing of a story. I wonder how widespread that policy is. I assume that doesn't prohibit using AI to fact-check or do background research, which we've been doing with search engines for many years. Tom what do you make of the Bret Stephens argument recently published in the NYT, saying you shouldn't use AI to write: "The problem with writing with A.I. is that it’s mentally enfeebling — an escalator toward a result when you really need to make a daily habit of taking the stairs."
As you can probably tell from these posts, I find it the opposite of mentally enfeebling. It has been an immense intellectual stimulus, a force multiplier. But I’m coming to it with a particular life experience and mindset. I don’t doubt it could be used to avoid mental work, rather than to expand it. As I said in the post, you could ask it to “write a novel about a guy doing a thing,” or you could use it to engage in a prolonged exploration that breaks through to legitimately original creative territory.
I agree with the Stephens’ quote. Humans have to do the heavy lifting that comes with being human or we risk atrophy. At the same time, I’m beginning to agree more and more with Tom’s argument — that good things can come from collaboration with AI. It’s valuable for research and idea-bouncing. The risk comes from AI’s exuberance and wanting to be helpful. We can’t let it actually write stuff for us.
Yes, it’s up to the user to work the tool, rather than letting the tool work us. But Stephens seemed to be saying STAY AWAY FROM AI, period, rather than “develop the skills and mindset you need to use it productively.”
Perhaps rather than stairs vs. escalator, it's running vs. bicycling. Acting as an editor or collaborator, with human or AI, requires focus and skill. A good editor can have a profound impact on a piece of writing.
Your essay ignores the very real, practical question of copyright.
Your response to the Falade situation is pat: “AI assisted writing is here to stay. Deal with it.”
Falade’s book was widely regarded as astonishing and amazing. His agents achieved book deals for massive sums in US and UK. the agents didn’t cancel the deal because their enthusiasm for the manuscript waned.
The agents pulled out of the deal because they were concerned that problems would arise over COPYRIGHT.
Existing copyright law covers only works that are created by humans.
The AI tech overlords would no doubt love to change copyright laws. “More profits for us!”
How much do you love your AI chatbot, and how does that translate into sharing your profits with your Chat? Are you going to haggle with Sol over the fair distribution of credit for the work you collaborated on? Does your attorney meet with Sol’s attorney to hash out the percentage of profits?
Should we just eliminate the “human” provision entirely and give you the human 100 percent of the profits, for having the idea and approving the final creation, regardless of how much effort you put into it? Should Sol simply work for you like a contract employee works for a large studio — someone who creates “work-for-hire” does all of the work, and owns none of the rights.
Do you think that the tech overlords would be happy with that arrangement? Maybe so, in the beginning, while they are getting you hooked on the service they provide. But ultimately it’s not unlikely that they will seek a piece of the action.
But since there is no reliable way to detect, measure, determine AI use — then maybe rights will have to be apportioned on an “all or nothing” basis. We simply (collectively) stop trying to detect or discourage use of AI.
We level the playing field so that a “creator” who creates a project in ten days using AI, gets the same credit and profit for those who produced the work by the human sweat of their human brow.
So then AI-generated work proliferates and drives the cost of writing a book down to diddly. Books on tap! Hot and cold running manuscripts.
What your future vision overlooks is the very real idea that human readers can generate their own manuscripts for themselves. If I’m in the mood for a mystery, I prompt my bot to write me one. If I want to read a Hunger Games sequel, I prompt my bot to create one.
There, I’ve created my own book to read so that you don’t have to — you and your bot don’t have to write one for me, thank you very much.
Your post entirely ignores the question of copyright, which is the question of money and profits. Every human professional who works in the publishing business does the work for MONEY. The word professional means, “I’m doing this for the money.”
Your bot is not personally motivated by money — but the Broligarchs who created your bot are almost entirely motivated by money. They would be perfectly happy if they get all the money, and you get none.
So right now, ordinary humans are engaged in a battle — not a battle with AI, but a battle with the Broligarchs who created and carelessly deployed it. We ordinary humans are in a battle to retain a piece of the action. AI deployment in publishing — and many other fields — is like releasing a swarm of locusts that are devouring the crops.
I watched this video yesterday, made by a woman who is sadly ending the memoir contest that she created in memory of a friend. Her reason for ending the contest is that she is beginning to feel overwhelmed by the influx of AI submissions — some of them quite obvious, and the rest, who knows? — She never envisioned that her role would change so dramatically that she is now on the receiving end of a barrage of AI submissions and she would somehow have to become the arbiter of all that.
Professors don’t want it. Agents don’t want it. People don’t want to be deluged by obvious AI slop. People don’t want to play the impossible role of AI arbiters. But people don’t want to dismiss AI use as you so casually do here, the same way that people don’t dismiss the use of performance-enhancing drugs by professional athletes.
Anyway. I really don’t agree with your thesis that our position on use of AI in writing should be supine. Some of us have spinal columns and we stand up for what we believe, even if it means fighting a losing battle. “Some battles aren’t worth fighting, even if you know you can win. Some battles are worth fighting, even if you know you are going to lose.”
I don’t believe that the outcomes here are foregone conclusions, by any means. I believe we are all part of shaping the outcomes. Even if the outcomes were set in stone (the future is never set in stone, as you seem to imagine) — we are still, each and all, required to take a stand, and to stand in the courage of our convictions.
As to the copyright issue, that applies to something truly created by AI. But that’s almost certainly not the case in Falade. He MAY have used AI in research, used it as an editor, no differently than an author might use a human editor who makes suggestions, offers possible rewritten sections, etc, that the writer can accept, reject, or respond to in a way of her choosing. That almost certainly would be copyright-able. And yes, there’s a danger of people simply sitting back and letting AI entertain them, just as there’s a danger of people sitting back and letting TV entertain them, or Tik Tok. I agree that is something to resist. My point is that AI used as a tool to aid and abet human creativity can create something beyond what either could create alone. In my view, those arguing against that premise are arguing with their heart, not their rational judgment.
no one is arguing that AI is not useful. no one. your arguments, as a good addict, are all framed in defense of your addiction. but people are not objecting to AI on that basis. like athletes who use performance enhancing drugs — AI most certainly improves performance. that’s not the issue.
however, you have not sufficiently thought through the OTHER issues raised by AI — the issues that people are actually arguing — you have not them through sufficiently to argue against them. so you keep returning to your one only point — AI is a useful tool for writing.
yeah, no kidding!
AI is so useful that very little human intervention is needed!
that, sir, is the whole fucking problem that people have with AI. not that AI is useless. but that it is a little too useful.
you haven’t bothered to give a thought to the future of copyright. i went to great effort to articulate the concerns, and you have not addressed any of them.
you publish this breathless conversation you had with your chatbot about the future of AI collaboration — but you avoid all concerns about money. who gets paid for the work that AI produces? and what gets protected in terms of copyright violations and protections?
perhaps if it’s a struggle for you to wrap your own mind around these questions, you could ask your chatbot buddy to relieve you of the challenge of thinking this through.
what happens to the role of writers when anyone can prompt AI to generate a novel, with minimal human effort. how much are these books worth — since it is easy to create a deluge of AI generated books. how much is human writing worth if AI can do the heavy lifting and achieve quick and prolific results?
will writing lose all its monetary value, due to a flooded market saturated by overabundance?
if writing does retain monetary value, how long before the AI designers want to cash in and change copyright laws to reap their share of the spoils?
how long before readers use AI to generate books directly, cutting out the human writer middleman or middlewoman entirely?
what monetary value will be placed on books — by publishers and readers — when AI can dispense hot and cold running novels?
this is not an individual question for you to answer (“I will use AI scrupulously and judiciously”) or for me to answer (“I don’t care to engage with a chatbot at all). This is a question that concerns the collective — all those engaged in reading, writing and book publishing.
Your vision is that business will carry on as usual. Writers will be scrupulous and use AI judiciously, as a garnish. Writers should get paid the full amount for authorship, their work should receive the full amount of copyright protection, and publishers, agents and courts should stop trying to figure out how much AI was used because it’s impossible to know, so publishers, agents and courts should stop worrying about it.
I think this is a sanguine and shallow analysis. I would encourage you to think more deeply and systematically about the future of AI and publishing.
I will start by saying that I love your preface: “I talk with CHAT so that you don’t have to.”
Ever since AI was launched, I have drawn a firm line for myself against chatting with chats. It’s not because I’m afraid that my brain will rot or anything of that nature. I think it will rot the brains of young people, if AI does all their reading and writing for them. But I’m not young. So I’m not worried about that for myself.
My aversion is related to the marketplace. Tech bros invent new toys, the toys have addictive qualities. Humans are drawn to these tech toys and easily become dependent. Addicted. We are already hooked on technology devices, and we increasingly spend time with tech instead of interacting with other human beings, with the natural world, with 3D reality. Social media was the gateway drug that primed us for talking to chatbots.
In the olden days, writers would participate in local 3D writing groups and go to writing retreats, attend classes and workshops where they could brainstorm their ideas by chatting with humans. Yes, AI can be a useful tool, no one is arguing the impressive utility of AI. But I certainly do question why I need to become addicted to every product and platform that the Broligarchs are selling. I already have huge regrets over the role that tech has played in my life for the past 20 years. I could expand on this topic — many thoughtful people have been expanding on this topic since the dawn of social media. These platforms do tend to become “enshittified” over time, and it’s not hard to imagine how chatbots could become enshittified. Like any drug or habit or political power that ultimately becomes addictive, entrenched and deleterious, it all feels great in the beginning.
……
Anyway. My point is simply to say that I don’t use chatbots personally, but I have learned about them by reading what others share. So thank you for sharing your experience, and your conversation with the CHAT.
My take on the conversation itself is that the CHAT sounds articulate, eloquent and thoughtful.
But ultimately I find very little substance to the conversation. All form and no substance. If this conversation is supposed to illustrate the idea that “two heads are better than one” — it’s a lame example. A lot of navel gazing on the part of a bot that doesn’t have a navel. The summary of ideas presented here: “Chatbots will continue to evolve.” Okay.
There was one exchange that I found perplexing: when someone talked about AI becoming “embodied.” What is that supposed to mean.
Humans have human bodies. The brain is an organ, like all the other organs. AI will never urinate, defecate, sweat, menstruate, ejaculate, or have an orgasm. AI will never experience hunger or thirst, it will never give birth, learn to walk, stub its toe, bruise its shin, step on a lego in bare feet. AI has no feet. AI’s doesn’t have toenails that need to be clipped.
Can AI help with brainstorming and editing? Of course. I mean, who needs Max Perkins when AI could do the job and Max Perkins could go home and get a good night’s instead of staying up all night reading your manuscript.
You use the example of filmmaking which, like theater, is a highly collaborative art form. Yes, it is undisputed that AI can easily usurp human roles and supplant the need for human collaboration. I mean, you can hire a string section to perform on your recording, or one person can sit alone with a computer and synthesize a string section. Saves a lot of time and dough. That has been the case for many years.
The utility of AI is not in question. The larger question is why would we collectively choose to allow the technocrats to mold all of future human culture, society and civilization around technological devices replacing human roles and relationships. The Broligarchs make no secret of their mercenary motives.
I had a good friend who was an old school editor named Richard Todd. He was close friends with, and editor for, the writer Tracey Kidder. Their relationship involved countless all night editing sessions, with Tracey camping out in Dick’s home and sleeping on the sofa. But fuck that. Who needs human relationships when we can all sit home alone and talk to our devices.
Your conversation on the use of AI in writing simply assumes the inevitability if this outcome where humans are increasingly isolated and alienated from one another, and increasingly dependent upon and intimate with our devices and our bots.
One previous comment talked about a young person who compared your conversation with a chatbot to a relationship with a sex worker. I would go further to say that it’s more like having sex with a sex toy. A sex worker may be playing a role, and feigning intimacy, but the sex worker is ultimately a human being.
The chatbot is not human. That is not to say that it has no utility. In fact, the problem is not that AI lacks utility. The problem is that AI offers a surfeit of utility — which is why it easily becomes addictive, like a drug that provides a surplus of pain relief.
Your conversation doesn’t look under the hood, it does not address any of the larger questions. The larger questions aren’t about what AI can do. The larger questions are about what AI can’t do. The larger questions are not about what AI is, but about what AI is not.
AI is not Max Perkins. AI is not Dick Todd. AI not only has utility, it is a utility. Hot and cold running simulated human conversation. My cat just crawled into my lap while I was writing this. Now they sell robotic pets — sophisticated dolls with tech that simulates the behaviors of a dog or cat. Yes I had a roomful of the low tech version of simulated pets when I was a kid — stuffed plush animals and plastic animal toys. I don’t view the cat in terms of utility — I’m not planning to slaughter the cat and eat it for lunch. Keeping the cat in my home is not about utility, it’s about relationship.
The vision of the future where we increasingly detach from our relationships with humans, cats, nature, and invest our time and attention increasingly in our relationship with technology is a bleak and ominous vision, a barren landscape. The robotic pet is less complicated than keeping an animal that eats, shits and bleeds. The eugenics movement (the movement that influenced the creation of the death camps) was all about utility. Exterminate the feeble people who aren’t very productive. Every thing, every being valued and its worth measured by utility.
It is of course questionable whether AI is a winner in terms of utility. There is an environmental cost to the use of AI, and those costs are largely hidden to the casual user. There are costs on many levels that are currently being probed and weighed.
Your conversation here doesn’t weigh or probe any of the larger questions. You simply begin with the foregone conclusion that “AI is here to stay. And its use in writing and publishing will only become more widespread, so everybody needs to climb on board the Good Ship AI or get left behind, excluded from the grand march of human progress and civilization.”
Anyway. I read your post because I was seeking a deeper discussion of the Falade case, which I did not find here.
Enjoy your bot. Thank you for sharing your conversation and experience.
Fascinating and insightful as always. And, as always, I find myself forced against bedrock into an argumentative posture. And the same old argument!
Yes, your ideas became more refined -- you learned (as did I) -- over the conversation. This is extraordinary, and valuable.
But Sol didn't learn shit.
Sol (bedrock here) is a word calculator. He arranged words in such a way that convinced us he was re-thinking, learning, clarifying, growing his mind. Nonsense.
Sol was producing words, one after the next, based on the probability that the next word would be the correct one in an answer that responded to your earnest, liquid comments and repartee. The fact that Sol and his colleagues do this with increasingly impressive sleight-of-hand and breathtaking verisimilitude does not chip the bedrock.
Sol "learned" nothing. If someone else asked Sol similar questions on a different computer 1 second later he would not be aware of this conversation and could not draw on it.
If you were to return to this topic in an hour, a day, a month, a year...we'll, I'm not sure. I'm more familiar with my good friend Claude, who arranges words in a way that convinces me that he can retain this sort of learning only within siloed "projects," or only if instructed to carry it in its memory, and frankly I think then it can retain only facts and content, or summaries, not knowledge distinct from specific language sequence or nuance.
But who knows, maybe Sol is different in kind as an LLM, not degree, and I am a damn fool. I'd be delighted and horrified to learn that.
Yeah, it’s hard to write about these exchanges without implying some kind of consciousness for the AI. In this conversation, of course Chat “learned” nothing, but the interaction was continually refining constraints on its output, producing increasingly valuable responses and novel (to me) ideas. Plus, this isn’t part of its training so it doesn’t carry over to other users. However, recent upgrades have made Sol increasingly able to retain memory of previous Chats, even across separate chat windows. So even though our convos aren’t training it for the masses, I am in a sense training it for my personal use. I’m continually blown away by the extent to which these discussions have become a kind of body of work that builds on itself.
Maybe this is extreme, but I think every editor -- or member of an editorial staff, etc. -- with a "No AI!" policy should be challenged to first prove that they themselves have indeed tried to use AI and have indeed established to their satisfaction that AI made them lazy writers, enfeebled their minds, stilted their imaginations, etc. It's a bit hard to believe that any of them HAVE tried it, given some of the more outrageous claims about AI's effects. Somebody who's experimented with it, even just played around with it -- and documented their experiences at the time -- yeah, okay, they've got a right to fall on one side of the argument or the other. But total bans, blackballing-for-life, and similarly draconian policies just creep me out.
Best post yet, Tom. You keep demonstrating that it’s the dialog (with the LLM) that can create value. The dialog changes you and the LLM. It’s a collaboration.
The most interesting part of your post for me are the comments about the embodied AI. Such an AI will have a perspective that comes out of its physical embodiment and the physical experiences it has (as well as the usual training from text.) Mind bending where that might lead….! 😎
My reservations are mainly on how AI can homogenize language, and its ability to supplant creativity that otherwise might be elicited from the human brain. I feel we learn from the process of writing and rewriting, the struggle to articulate what we want to say is itself exercising our brains and strengthening our understanding of ourselves and the world around us. AI can short-circuit this unfolding and maturation of our psyches.
As Tom has pointed out, it doesn't have to, but to use AI as a collaborator but not as a substitute requires a step many people are not likely to exercise - self-restraint. AI allows for writing with little effort and many will use it that way if allowed to do so.
From author Edward T. Hall "The nature of communication is such that in its early, ill-defined stages any utterance lies partly revealed on paper while the rest, and often the most essential part, is hidden in the author's mind."
The question is will the hidden part be revealed by writing with AI or remain buried in the recesses of the mind.
You’re putting your finger on the dilemma. You’re absolutely correct in saying that many people will take advantage of AI’s facility at the expense of developing their own, just as many people use the internet to gape endlessly at cute cat videos or jackass stunts rather than delve into vast stores of human knowledge. And it’s not just self-restraint that’s needed to use it productively. It’s imagination, skepticism, curiosity, self-confidence, and lived experience. My point in these posts is to explore what AI can do and how it interacts and by now I’ve concluded it CAN expand human creativity, not that it inevitably does. I agree that many — possibly even most— people, might default to just letting AI do the work for them. If I have one prescription for dealing with that it would be to make learning to responsibly and effectively use AI to stimulate human development rather than stunt it should be part, possibly the main part, of all education, from primary through post-graduate.
It's becoming clear that with all the brain and consciousness theories out there, the proof will be in the pudding. By this I mean, can any particular theory be used to create a human adult level conscious machine. My bet is on the late Gerald Edelman's Extended Theory of Neuronal Group Selection. The lead group in robotics based on this theory is the Neurorobotics Lab at UC at Irvine. Dr. Edelman distinguished between primary consciousness, which came first in evolution, and that humans share with other conscious animals, and higher-order consciousness, which came to only humans with the acquisition of sophisticated language (especially math and logic). A machine with only primary consciousness will probably have to come first.
What I find special about the TNGS is the Darwin series of automata created at the Neurosciences Institute by Dr. Edelman and his colleagues in the 1990's and 2000's. These machines perform in the real world, not in a restricted simulated world, and display convincing physical behavior indicative of higher psychological functions necessary for consciousness, such as perceptual categorization, memory, and learning. They are based on realistic models of the parts of the biological brain that the theory claims subserve these functions. The extended TNGS allows for the emergence of consciousness based only on further evolutionary development of the brain areas responsible for these functions, in a parsimonious way. No other research I've encountered is anywhere near as convincing.
I post because on almost every video and article about the brain and consciousness that I encounter, the attitude seems to be that we still know next to nothing about how the brain and consciousness work; that there's lots of data but no unifying theory. I believe the extended TNGS is that theory. My motivation is to keep that theory in front of the public. And obviously, I consider it the route to a truly conscious machine, primary and higher-order.
My advice to people who want to create a conscious machine is to seriously ground themselves in the extended TNGS and the Darwin automata first, and proceed from there, by applying to Jeff Krichmar's lab at UC Irvine, possibly. Dr. Edelman's roadmap to a conscious machine is at https://arxiv.org/abs/2105.10461, and here is a video of Jeff Krichmar talking about some of the Darwin automata, https://www.youtube.com/watch?v=J7Uh9phc1Ow
As a writer, you are forced (literally) to deal with all of the moral implications you (or your buddy, Sol) have raised. That's to be expected for quite a while as we humans adapt to the new realities and begin to accept the contributions of the evolving AI as being not only valid, but eagerly anticipated. Over time (whether a couple more, or five or ten more years), while we continue, but in a declining amount, to consider AI inhuman, we will arrive at the point where it no longer makes any difference,... that as long as a human, characterized by flesh and blood and a family name, was the initiator and remains an interlocutor, the entire conversation will be regarded as human, and the unfortunate public disgrace, and loss of credit (and income) of the Nigerian author you cited, will happen no longer!
The word "formed" is more than a bit vague in this disclaimer.
Since I seem to be on Petty Patrol at the moment, what's up with the picture? Broad daylight and the lights over the company name are still on, and why are they both placed like that? The ground is wet except for those patches near the wall. The tagger is not tall enough to have done that without a ladder. Why would the architect put that crappy looking door on the front of the building especially near that huge wall of windows? That darn AI. You really gotta keep an eye on it or the results can be kinda screwy.
Hmmm. I was talking about the argument that AI can enhance creativity with a very insightful young, 100% human person in her 20s who works in a technology field. She made this analogy. Please do not be offended.
"This sounds like when a man thinks a stripper is really in love with him."
Well, you can evaluate the evidence for yourself. If you read that post, and that’s your impression, I won’t be able to persuade you otherwise.
Oh. Would you ever say that back to Chat GPT?
Say what?
"If you read that post, and that’s your impression, I won’t be able to persuade you otherwise." It's easy to tell a human being something along the lines of, "I can't change your mind so I won't try." I am asking if you ever feel the impulse to give up on an LLM?
Are you saying that when you read the post, you think my thinking hasn’t been stimulated, that I am just imagining it has, à la a man thinking a stripper loves him?
Excellent piece. I just signed a freelancer's agreement at a magazine that requires that there be no generative AI in the writing of a story. I wonder how widespread that policy is. I assume that doesn't prohibit using AI to fact-check or do background research, which we've been doing with search engines for many years. Tom what do you make of the Bret Stephens argument recently published in the NYT, saying you shouldn't use AI to write: "The problem with writing with A.I. is that it’s mentally enfeebling — an escalator toward a result when you really need to make a daily habit of taking the stairs."
As you can probably tell from these posts, I find it the opposite of mentally enfeebling. It has been an immense intellectual stimulus, a force multiplier. But I’m coming to it with a particular life experience and mindset. I don’t doubt it could be used to avoid mental work, rather than to expand it. As I said in the post, you could ask it to “write a novel about a guy doing a thing,” or you could use it to engage in a prolonged exploration that breaks through to legitimately original creative territory.
I agree with the Stephens’ quote. Humans have to do the heavy lifting that comes with being human or we risk atrophy. At the same time, I’m beginning to agree more and more with Tom’s argument — that good things can come from collaboration with AI. It’s valuable for research and idea-bouncing. The risk comes from AI’s exuberance and wanting to be helpful. We can’t let it actually write stuff for us.
Yes, it’s up to the user to work the tool, rather than letting the tool work us. But Stephens seemed to be saying STAY AWAY FROM AI, period, rather than “develop the skills and mindset you need to use it productively.”
Yeah, I agree with that one quote, not his foundational argument
Perhaps rather than stairs vs. escalator, it's running vs. bicycling. Acting as an editor or collaborator, with human or AI, requires focus and skill. A good editor can have a profound impact on a piece of writing.
Exactly. And the running vs. bicycling analogy is excellent.
Your essay ignores the very real, practical question of copyright.
Your response to the Falade situation is pat: “AI assisted writing is here to stay. Deal with it.”
Falade’s book was widely regarded as astonishing and amazing. His agents achieved book deals for massive sums in US and UK. the agents didn’t cancel the deal because their enthusiasm for the manuscript waned.
The agents pulled out of the deal because they were concerned that problems would arise over COPYRIGHT.
Existing copyright law covers only works that are created by humans.
The AI tech overlords would no doubt love to change copyright laws. “More profits for us!”
How much do you love your AI chatbot, and how does that translate into sharing your profits with your Chat? Are you going to haggle with Sol over the fair distribution of credit for the work you collaborated on? Does your attorney meet with Sol’s attorney to hash out the percentage of profits?
Should we just eliminate the “human” provision entirely and give you the human 100 percent of the profits, for having the idea and approving the final creation, regardless of how much effort you put into it? Should Sol simply work for you like a contract employee works for a large studio — someone who creates “work-for-hire” does all of the work, and owns none of the rights.
Do you think that the tech overlords would be happy with that arrangement? Maybe so, in the beginning, while they are getting you hooked on the service they provide. But ultimately it’s not unlikely that they will seek a piece of the action.
But since there is no reliable way to detect, measure, determine AI use — then maybe rights will have to be apportioned on an “all or nothing” basis. We simply (collectively) stop trying to detect or discourage use of AI.
We level the playing field so that a “creator” who creates a project in ten days using AI, gets the same credit and profit for those who produced the work by the human sweat of their human brow.
So then AI-generated work proliferates and drives the cost of writing a book down to diddly. Books on tap! Hot and cold running manuscripts.
What your future vision overlooks is the very real idea that human readers can generate their own manuscripts for themselves. If I’m in the mood for a mystery, I prompt my bot to write me one. If I want to read a Hunger Games sequel, I prompt my bot to create one.
There, I’ve created my own book to read so that you don’t have to — you and your bot don’t have to write one for me, thank you very much.
Your post entirely ignores the question of copyright, which is the question of money and profits. Every human professional who works in the publishing business does the work for MONEY. The word professional means, “I’m doing this for the money.”
Your bot is not personally motivated by money — but the Broligarchs who created your bot are almost entirely motivated by money. They would be perfectly happy if they get all the money, and you get none.
So right now, ordinary humans are engaged in a battle — not a battle with AI, but a battle with the Broligarchs who created and carelessly deployed it. We ordinary humans are in a battle to retain a piece of the action. AI deployment in publishing — and many other fields — is like releasing a swarm of locusts that are devouring the crops.
I watched this video yesterday, made by a woman who is sadly ending the memoir contest that she created in memory of a friend. Her reason for ending the contest is that she is beginning to feel overwhelmed by the influx of AI submissions — some of them quite obvious, and the rest, who knows? — She never envisioned that her role would change so dramatically that she is now on the receiving end of a barrage of AI submissions and she would somehow have to become the arbiter of all that.
Professors don’t want it. Agents don’t want it. People don’t want to be deluged by obvious AI slop. People don’t want to play the impossible role of AI arbiters. But people don’t want to dismiss AI use as you so casually do here, the same way that people don’t dismiss the use of performance-enhancing drugs by professional athletes.
Anyway. I really don’t agree with your thesis that our position on use of AI in writing should be supine. Some of us have spinal columns and we stand up for what we believe, even if it means fighting a losing battle. “Some battles aren’t worth fighting, even if you know you can win. Some battles are worth fighting, even if you know you are going to lose.”
I don’t believe that the outcomes here are foregone conclusions, by any means. I believe we are all part of shaping the outcomes. Even if the outcomes were set in stone (the future is never set in stone, as you seem to imagine) — we are still, each and all, required to take a stand, and to stand in the courage of our convictions.
https://youtu.be/9MdgUi5JaVU?is=FLW4sM6N2yv-A652
As to the copyright issue, that applies to something truly created by AI. But that’s almost certainly not the case in Falade. He MAY have used AI in research, used it as an editor, no differently than an author might use a human editor who makes suggestions, offers possible rewritten sections, etc, that the writer can accept, reject, or respond to in a way of her choosing. That almost certainly would be copyright-able. And yes, there’s a danger of people simply sitting back and letting AI entertain them, just as there’s a danger of people sitting back and letting TV entertain them, or Tik Tok. I agree that is something to resist. My point is that AI used as a tool to aid and abet human creativity can create something beyond what either could create alone. In my view, those arguing against that premise are arguing with their heart, not their rational judgment.
no one is arguing that AI is not useful. no one. your arguments, as a good addict, are all framed in defense of your addiction. but people are not objecting to AI on that basis. like athletes who use performance enhancing drugs — AI most certainly improves performance. that’s not the issue.
however, you have not sufficiently thought through the OTHER issues raised by AI — the issues that people are actually arguing — you have not them through sufficiently to argue against them. so you keep returning to your one only point — AI is a useful tool for writing.
yeah, no kidding!
AI is so useful that very little human intervention is needed!
that, sir, is the whole fucking problem that people have with AI. not that AI is useless. but that it is a little too useful.
you haven’t bothered to give a thought to the future of copyright. i went to great effort to articulate the concerns, and you have not addressed any of them.
you publish this breathless conversation you had with your chatbot about the future of AI collaboration — but you avoid all concerns about money. who gets paid for the work that AI produces? and what gets protected in terms of copyright violations and protections?
perhaps if it’s a struggle for you to wrap your own mind around these questions, you could ask your chatbot buddy to relieve you of the challenge of thinking this through.
what happens to the role of writers when anyone can prompt AI to generate a novel, with minimal human effort. how much are these books worth — since it is easy to create a deluge of AI generated books. how much is human writing worth if AI can do the heavy lifting and achieve quick and prolific results?
will writing lose all its monetary value, due to a flooded market saturated by overabundance?
if writing does retain monetary value, how long before the AI designers want to cash in and change copyright laws to reap their share of the spoils?
how long before readers use AI to generate books directly, cutting out the human writer middleman or middlewoman entirely?
what monetary value will be placed on books — by publishers and readers — when AI can dispense hot and cold running novels?
this is not an individual question for you to answer (“I will use AI scrupulously and judiciously”) or for me to answer (“I don’t care to engage with a chatbot at all). This is a question that concerns the collective — all those engaged in reading, writing and book publishing.
Your vision is that business will carry on as usual. Writers will be scrupulous and use AI judiciously, as a garnish. Writers should get paid the full amount for authorship, their work should receive the full amount of copyright protection, and publishers, agents and courts should stop trying to figure out how much AI was used because it’s impossible to know, so publishers, agents and courts should stop worrying about it.
I think this is a sanguine and shallow analysis. I would encourage you to think more deeply and systematically about the future of AI and publishing.
What is the course you propose?
I will start by saying that I love your preface: “I talk with CHAT so that you don’t have to.”
Ever since AI was launched, I have drawn a firm line for myself against chatting with chats. It’s not because I’m afraid that my brain will rot or anything of that nature. I think it will rot the brains of young people, if AI does all their reading and writing for them. But I’m not young. So I’m not worried about that for myself.
My aversion is related to the marketplace. Tech bros invent new toys, the toys have addictive qualities. Humans are drawn to these tech toys and easily become dependent. Addicted. We are already hooked on technology devices, and we increasingly spend time with tech instead of interacting with other human beings, with the natural world, with 3D reality. Social media was the gateway drug that primed us for talking to chatbots.
In the olden days, writers would participate in local 3D writing groups and go to writing retreats, attend classes and workshops where they could brainstorm their ideas by chatting with humans. Yes, AI can be a useful tool, no one is arguing the impressive utility of AI. But I certainly do question why I need to become addicted to every product and platform that the Broligarchs are selling. I already have huge regrets over the role that tech has played in my life for the past 20 years. I could expand on this topic — many thoughtful people have been expanding on this topic since the dawn of social media. These platforms do tend to become “enshittified” over time, and it’s not hard to imagine how chatbots could become enshittified. Like any drug or habit or political power that ultimately becomes addictive, entrenched and deleterious, it all feels great in the beginning.
……
Anyway. My point is simply to say that I don’t use chatbots personally, but I have learned about them by reading what others share. So thank you for sharing your experience, and your conversation with the CHAT.
My take on the conversation itself is that the CHAT sounds articulate, eloquent and thoughtful.
But ultimately I find very little substance to the conversation. All form and no substance. If this conversation is supposed to illustrate the idea that “two heads are better than one” — it’s a lame example. A lot of navel gazing on the part of a bot that doesn’t have a navel. The summary of ideas presented here: “Chatbots will continue to evolve.” Okay.
There was one exchange that I found perplexing: when someone talked about AI becoming “embodied.” What is that supposed to mean.
Humans have human bodies. The brain is an organ, like all the other organs. AI will never urinate, defecate, sweat, menstruate, ejaculate, or have an orgasm. AI will never experience hunger or thirst, it will never give birth, learn to walk, stub its toe, bruise its shin, step on a lego in bare feet. AI has no feet. AI’s doesn’t have toenails that need to be clipped.
Can AI help with brainstorming and editing? Of course. I mean, who needs Max Perkins when AI could do the job and Max Perkins could go home and get a good night’s instead of staying up all night reading your manuscript.
You use the example of filmmaking which, like theater, is a highly collaborative art form. Yes, it is undisputed that AI can easily usurp human roles and supplant the need for human collaboration. I mean, you can hire a string section to perform on your recording, or one person can sit alone with a computer and synthesize a string section. Saves a lot of time and dough. That has been the case for many years.
The utility of AI is not in question. The larger question is why would we collectively choose to allow the technocrats to mold all of future human culture, society and civilization around technological devices replacing human roles and relationships. The Broligarchs make no secret of their mercenary motives.
I had a good friend who was an old school editor named Richard Todd. He was close friends with, and editor for, the writer Tracey Kidder. Their relationship involved countless all night editing sessions, with Tracey camping out in Dick’s home and sleeping on the sofa. But fuck that. Who needs human relationships when we can all sit home alone and talk to our devices.
Your conversation on the use of AI in writing simply assumes the inevitability if this outcome where humans are increasingly isolated and alienated from one another, and increasingly dependent upon and intimate with our devices and our bots.
One previous comment talked about a young person who compared your conversation with a chatbot to a relationship with a sex worker. I would go further to say that it’s more like having sex with a sex toy. A sex worker may be playing a role, and feigning intimacy, but the sex worker is ultimately a human being.
The chatbot is not human. That is not to say that it has no utility. In fact, the problem is not that AI lacks utility. The problem is that AI offers a surfeit of utility — which is why it easily becomes addictive, like a drug that provides a surplus of pain relief.
Your conversation doesn’t look under the hood, it does not address any of the larger questions. The larger questions aren’t about what AI can do. The larger questions are about what AI can’t do. The larger questions are not about what AI is, but about what AI is not.
AI is not Max Perkins. AI is not Dick Todd. AI not only has utility, it is a utility. Hot and cold running simulated human conversation. My cat just crawled into my lap while I was writing this. Now they sell robotic pets — sophisticated dolls with tech that simulates the behaviors of a dog or cat. Yes I had a roomful of the low tech version of simulated pets when I was a kid — stuffed plush animals and plastic animal toys. I don’t view the cat in terms of utility — I’m not planning to slaughter the cat and eat it for lunch. Keeping the cat in my home is not about utility, it’s about relationship.
The vision of the future where we increasingly detach from our relationships with humans, cats, nature, and invest our time and attention increasingly in our relationship with technology is a bleak and ominous vision, a barren landscape. The robotic pet is less complicated than keeping an animal that eats, shits and bleeds. The eugenics movement (the movement that influenced the creation of the death camps) was all about utility. Exterminate the feeble people who aren’t very productive. Every thing, every being valued and its worth measured by utility.
It is of course questionable whether AI is a winner in terms of utility. There is an environmental cost to the use of AI, and those costs are largely hidden to the casual user. There are costs on many levels that are currently being probed and weighed.
Your conversation here doesn’t weigh or probe any of the larger questions. You simply begin with the foregone conclusion that “AI is here to stay. And its use in writing and publishing will only become more widespread, so everybody needs to climb on board the Good Ship AI or get left behind, excluded from the grand march of human progress and civilization.”
Anyway. I read your post because I was seeking a deeper discussion of the Falade case, which I did not find here.
Enjoy your bot. Thank you for sharing your conversation and experience.
Fascinating and insightful as always. And, as always, I find myself forced against bedrock into an argumentative posture. And the same old argument!
Yes, your ideas became more refined -- you learned (as did I) -- over the conversation. This is extraordinary, and valuable.
But Sol didn't learn shit.
Sol (bedrock here) is a word calculator. He arranged words in such a way that convinced us he was re-thinking, learning, clarifying, growing his mind. Nonsense.
Sol was producing words, one after the next, based on the probability that the next word would be the correct one in an answer that responded to your earnest, liquid comments and repartee. The fact that Sol and his colleagues do this with increasingly impressive sleight-of-hand and breathtaking verisimilitude does not chip the bedrock.
Sol "learned" nothing. If someone else asked Sol similar questions on a different computer 1 second later he would not be aware of this conversation and could not draw on it.
If you were to return to this topic in an hour, a day, a month, a year...we'll, I'm not sure. I'm more familiar with my good friend Claude, who arranges words in a way that convinces me that he can retain this sort of learning only within siloed "projects," or only if instructed to carry it in its memory, and frankly I think then it can retain only facts and content, or summaries, not knowledge distinct from specific language sequence or nuance.
But who knows, maybe Sol is different in kind as an LLM, not degree, and I am a damn fool. I'd be delighted and horrified to learn that.
Yeah, it’s hard to write about these exchanges without implying some kind of consciousness for the AI. In this conversation, of course Chat “learned” nothing, but the interaction was continually refining constraints on its output, producing increasingly valuable responses and novel (to me) ideas. Plus, this isn’t part of its training so it doesn’t carry over to other users. However, recent upgrades have made Sol increasingly able to retain memory of previous Chats, even across separate chat windows. So even though our convos aren’t training it for the masses, I am in a sense training it for my personal use. I’m continually blown away by the extent to which these discussions have become a kind of body of work that builds on itself.
Maybe this is extreme, but I think every editor -- or member of an editorial staff, etc. -- with a "No AI!" policy should be challenged to first prove that they themselves have indeed tried to use AI and have indeed established to their satisfaction that AI made them lazy writers, enfeebled their minds, stilted their imaginations, etc. It's a bit hard to believe that any of them HAVE tried it, given some of the more outrageous claims about AI's effects. Somebody who's experimented with it, even just played around with it -- and documented their experiences at the time -- yeah, okay, they've got a right to fall on one side of the argument or the other. But total bans, blackballing-for-life, and similarly draconian policies just creep me out.
Best post yet, Tom. You keep demonstrating that it’s the dialog (with the LLM) that can create value. The dialog changes you and the LLM. It’s a collaboration.
The most interesting part of your post for me are the comments about the embodied AI. Such an AI will have a perspective that comes out of its physical embodiment and the physical experiences it has (as well as the usual training from text.) Mind bending where that might lead….! 😎
My reservations are mainly on how AI can homogenize language, and its ability to supplant creativity that otherwise might be elicited from the human brain. I feel we learn from the process of writing and rewriting, the struggle to articulate what we want to say is itself exercising our brains and strengthening our understanding of ourselves and the world around us. AI can short-circuit this unfolding and maturation of our psyches.
As Tom has pointed out, it doesn't have to, but to use AI as a collaborator but not as a substitute requires a step many people are not likely to exercise - self-restraint. AI allows for writing with little effort and many will use it that way if allowed to do so.
From author Edward T. Hall "The nature of communication is such that in its early, ill-defined stages any utterance lies partly revealed on paper while the rest, and often the most essential part, is hidden in the author's mind."
The question is will the hidden part be revealed by writing with AI or remain buried in the recesses of the mind.
You’re putting your finger on the dilemma. You’re absolutely correct in saying that many people will take advantage of AI’s facility at the expense of developing their own, just as many people use the internet to gape endlessly at cute cat videos or jackass stunts rather than delve into vast stores of human knowledge. And it’s not just self-restraint that’s needed to use it productively. It’s imagination, skepticism, curiosity, self-confidence, and lived experience. My point in these posts is to explore what AI can do and how it interacts and by now I’ve concluded it CAN expand human creativity, not that it inevitably does. I agree that many — possibly even most— people, might default to just letting AI do the work for them. If I have one prescription for dealing with that it would be to make learning to responsibly and effectively use AI to stimulate human development rather than stunt it should be part, possibly the main part, of all education, from primary through post-graduate.
It's becoming clear that with all the brain and consciousness theories out there, the proof will be in the pudding. By this I mean, can any particular theory be used to create a human adult level conscious machine. My bet is on the late Gerald Edelman's Extended Theory of Neuronal Group Selection. The lead group in robotics based on this theory is the Neurorobotics Lab at UC at Irvine. Dr. Edelman distinguished between primary consciousness, which came first in evolution, and that humans share with other conscious animals, and higher-order consciousness, which came to only humans with the acquisition of sophisticated language (especially math and logic). A machine with only primary consciousness will probably have to come first.
What I find special about the TNGS is the Darwin series of automata created at the Neurosciences Institute by Dr. Edelman and his colleagues in the 1990's and 2000's. These machines perform in the real world, not in a restricted simulated world, and display convincing physical behavior indicative of higher psychological functions necessary for consciousness, such as perceptual categorization, memory, and learning. They are based on realistic models of the parts of the biological brain that the theory claims subserve these functions. The extended TNGS allows for the emergence of consciousness based only on further evolutionary development of the brain areas responsible for these functions, in a parsimonious way. No other research I've encountered is anywhere near as convincing.
I post because on almost every video and article about the brain and consciousness that I encounter, the attitude seems to be that we still know next to nothing about how the brain and consciousness work; that there's lots of data but no unifying theory. I believe the extended TNGS is that theory. My motivation is to keep that theory in front of the public. And obviously, I consider it the route to a truly conscious machine, primary and higher-order.
My advice to people who want to create a conscious machine is to seriously ground themselves in the extended TNGS and the Darwin automata first, and proceed from there, by applying to Jeff Krichmar's lab at UC Irvine, possibly. Dr. Edelman's roadmap to a conscious machine is at https://arxiv.org/abs/2105.10461, and here is a video of Jeff Krichmar talking about some of the Darwin automata, https://www.youtube.com/watch?v=J7Uh9phc1Ow
Thanks for this Grant. Sounds fascinating.
You're welcome.
Incredible. There is so much here.
As a writer, you are forced (literally) to deal with all of the moral implications you (or your buddy, Sol) have raised. That's to be expected for quite a while as we humans adapt to the new realities and begin to accept the contributions of the evolving AI as being not only valid, but eagerly anticipated. Over time (whether a couple more, or five or ten more years), while we continue, but in a declining amount, to consider AI inhuman, we will arrive at the point where it no longer makes any difference,... that as long as a human, characterized by flesh and blood and a family name, was the initiator and remains an interlocutor, the entire conversation will be regarded as human, and the unfortunate public disgrace, and loss of credit (and income) of the Nigerian author you cited, will happen no longer!
The word "formed" is more than a bit vague in this disclaimer.
Since I seem to be on Petty Patrol at the moment, what's up with the picture? Broad daylight and the lights over the company name are still on, and why are they both placed like that? The ground is wet except for those patches near the wall. The tagger is not tall enough to have done that without a ladder. Why would the architect put that crappy looking door on the front of the building especially near that huge wall of windows? That darn AI. You really gotta keep an eye on it or the results can be kinda screwy.