Linguistic prediction time! For the near future — and perhaps much longer — slang and informal jargon are going to get _very_ strange, and change even faster to the point where the change is so rapid that their utility effectively disappears. I’m talking about not just the words, but specific patterns of speech, construction of argument, and even punctuation. It’s also why I changed the settings for the blog to allow AI agents “use to answer questions” access, though _not_ training access. I realized this because I saw yet another social media post complaining about the repetitiveness in AI generated scripts on YouTube. Which I totally understand; the shallow impressiveness is indeed, _shallow_. > If I hear “it’s not this, it’s this!” In a YouTube script one more time, I will lose it. > > I’m so tired of AI generated scripts. And I thought about my own writing, which very, _very_ definitely has its own stylistic tells, many of which overlap with supposed “AI writing” tells, but predates the technology by decades. And the meme where _Frankenstein_ is judged to be “AI written” simply because it was part of the training corpus. And how we know that training is still going on (and even incorporating poisoned data). Slang — and to a lesser degree, jargon — have always been about distinguishing in/out groups verbally; the word choice is the signifier of what group you belong to. Hot, rad, bad, sick, ill, tubular, cool, awesome, on fleek; the list literally goes on for pages. But in these cases, if you detune the actual _words_ , the grammar and syntax is essentially the same. If there is a high enough ratio of context clues to slang, it’s basically a simple substitution cipher. See also Review: Viewing Yellowjackets (Showtime) From A Certain Point Of View Similarly, there are trends in language structure, particularly evident now thanks to the internet and short-form video. Be sure to like and subscribe, ring that bell to be notified, all begin originally as very effective calls-to-action that are copied so much that they quickly suffer the same fate as “on fleek” when I try using it in public. And now we’re seeing the same pattern happen not just with calls-to-action, but _to the actual structure of the writing itself._ You know, the sort of thing that correlation machines are perfect at imitating. I’ve noticed AI influencing my own writing; not in a form of echolalia, but in subtler ways. I copyedit less, for example, and when given the choice of two sources for something factual, like why Americans use utensils the way they do (it’s not Benedict Arnold, by the way), I would rather cite a source from 2017 than a source from 2025, even if they say the same thing, because I don’t have to worry if the 2017 source is AI-written or not. I’m relying on my history of always sounding like this, plus my natural error rate, and what I choose to cite to help cement the reality that I’m not using an LLM to write this. (There’s also an additional information layer that I’m awkwardly using and not elaborating on intentionally.) And you _know_ that content creators are paying attention to these trends as well. And if you’re not an old fart like me, you don’t have the ability to point to your past work to prove that you’re “real”. See also When AI Mirrors Your Grudge -- Or Your Self-Blame You just have your language and context. The closest thing we have to this in real life is the creation of rhyming slang, though after its creation it seems to be, once again, essentially a substitution cipher like “cool”. Any kind of linguistic arms race with an LLM puts us in a stark John Henry situation. I don’t know what the eventual outcome will be; there’s no way to force it on the chainwax (NSFW link). It may be that skibidi toilet and 6-7 are the first organic reactions in their surrealism; evocation rather than communication. Because just like some people mistake my use of big words and a large vocabulary as implying that I’m looking down on them when I’m not, it’s going to get more and more common that people are accused of using AI to write articles and scripts and so on, just for writing using the same styles and conventions that people were already using in the training data. Maybe we’ll find additional ways to write between the lines, conveying information through _interpretation of context_ rather than context itself. Maybe it’ll actually be helpful in the long run, allowing more people to identify hacks and propaganda (which is, by necessity, are repetitive and formulaic) along with LLM-written scripts and articles. Regardless of exactly _what_ it looks like, though, it’s likely to look unusual, and strange, and probably confuse and annoy a lot of fascists the same way modern art does (NFSW, 100% worth it). And I am 100% here for _that_. Featured Image by Willi Heidelbach from Pixabay See also Our Accuracy Is Even More Important The More Blatant Their Lies Are

We have our zombie apocalypse. Replicants are among us, wreaking havoc. Honey the cat is loose. It’s going on right now. It just isn’t what you expected. ## Call and Response First, I have to tell you about the "Chinese Room" thought experiment. Originally described by John Searle, it goes something like this: > [In this thought experiment], Searle…supposes that he is in a closed room and is receiving questions in Chinese. While he cannot understand Chinese, he has a large collection of Chinese phrasebooks in the room, with questions and matching answers. When he receives a question, he need only to look up the same sequence of characters in one of the books and respond with the indicated answer, even though he does not understand the question nor the answer….Searle argues that, without “understanding” (or “intentionality”), we cannot describe what the machine is doing as “thinking” and, since it does not think, it does not have a “mind” in anything like the normal sense of the word.https://en.wikipedia.org/wiki/Chinese_roomhttps://en.wikipedia.org/wiki/Chinese_room This concept makes the Turing test – sometimes thought of as a test to see if a machine “thinks” – worthless. If the “phrasebooks” are complex enough, a machine could pass the Turing test but be entirely without _thinking_ , without consciousness, without awareness. Searle was originally talking about the difference between humans and computers; his point isn’t limited to them. What he is describing is the distinction between _conciousness (or awareness)_ and _intelligence_. A computer – or any entity – could be highly intelligent but completely without consciousness. It could be intelligent enough – have enough “phrasebooks” – to give appropriate responses and behaviors that mimic consciousness without there being “anyone home”. That’s the idea behind _philosophical zombies_. ## Complicated Replicants It’s also what Deckard was trying to determine with his Voight-Kampff test in _Bladerunner_. There, the replicants (androids) were undetectable except by their visceral emotional reaction (or lack of reaction) to a series of upsetting images and questions. While it tracks that a philosophical zombie wouldn’t be capable of real empathy [1], there’s a flaw in trying to measure that with external observations like shown in those films. The assumption of the Voight-Kampff test. While the Voight-Kampff test works for the film, it falls prey to the same failure as the Turing test [2]. It assumes that emotional responses are complex enough that they cannot be replicated accurately by a machine. To use the “Chinese Room” example, they’re assuming that the person inside won’t be able to deal with the complex incoming messages, dropping some entirely, not having time to look up others, and not be able to write responses fast enough. But that isn’t a problem of the _system_. To solve this problem, you can still use the same Chinese room. You just need more people inside the room, making sure that each has a complex enough “phrasebook”. **Both the Turing test and the (fictional) Voight-Kampff test are testing a _quantitative_ difference, not a _qualitative_ difference.** _They are just testing processing speed_. Which means that those tests _cannot_ determine if a human-shaped individual has “consciousness” or “awareness”, or if it is just a very complicated (meat) machine. If you asked a philosophical zombie if it had a consciousness, it would know to say “Of course I have a consciousness” and appear upset just as _you_ would…merely because its “phrasebook” would tell it to act that way! See also Seeing Ourselves In The "Other" (Guest Post) ## Read the Questions Twice Before Starting the Test So how could you _really_ test for a philosophical zombie? I think it’s pretty simple – though a bit complicated to do in practice. You look for **introspection** and **self-consistency**. See if the individual is resolving conflicting ideas and thoughts. Determine if they’re incorporating new information and ideas into their existing thoughts and world-view. Let’s go back to the “Chinese Room”. Normally, your Chinese room can expand its range of responses with a pretty simple addition method. The hard part is teaching “If you get this format of a message, take the first part to match your input, then take the second part to form your response.” Once you’ve brute-forced that lesson, everything else can follow that example. But having an internally consistent world-view is a _qualitatively_ different problem. In such a scenario, it wouldn’t matter how many people you had working, or how complicated the phrasebooks were. It would be _extremely_ difficult – maybe impossible – to even _appear_ to have this kind of reflection and introspection. Think about it: To achieve that goal, you’d have to give the people inside the room (who do not speak or read a bit of Chinese) instructions written entirely in Chinese, about comparing two different Chinese texts, making a value judgement about the _content_ of the texts, then editing the _phrasebooks(!)_ and generating new output that makes sense. In Chinese. With _no_ translation manuals. Can you think of any human-shaped creatures right now **who don’t seem to “get” empathy** (though they try to say “the right things”) and **who seem utterly unable to have a coherent set of principles**? Indeed, they will endorse completely opposing ideas (and lies) from one minute to the next with no sense of irony? Photo by Charles Deluvio on Unsplash Yeeeeeah. This “zombie” infection has spread pretty damn far, and unlike COVID-19, a mask isn’t going to help. ## Zombie Creation Engines Looking at the world right now like this could seem … well, hopeless. But there’s one big sign of hope. One thing that might turn back this infection. **Sufficient – _and the right kind of_ – conflict.** All of us enter states of intelligent non-consciousness from time to time. Remember the last time you drove home and don’t remember it? Or were working on something and hit “the zone” and don’t really know how you did the task you did? One big theory of consciousness suggests that consciousness plays a role in resolving internal conflict. Our consciousness is middle management. In situations like driving a familiar route, our consciousness doesn’t have to resolve any kind of conflict. All your body has to do is “repeat the pattern of turns that get us to where we’re going”. In those sorts of situations, you could think of it as our awareness “powering down” temporarily. In those situations, we’re still “conscious” in the sense of being _awake_ , but we aren’t really “conscious” in the sense of being _aware_. That kind of supervisory awareness only kicks in when there are competing options and a decision needs to be made, or an ethical dilemma resolved. For most of us, switching consciousness back on is not a problem. Boots right back up, and once our consciousness kicks back in, change can happen. It would only be a problem if our consciousness stayed shut down most – or all – of the time. See also Tennessee GOP Apparently Willing To Sacrifice Millions Of Taxpayer Dollars, Citizen's Health To Spite Planned Parenthood ## Unempathetic Zombies Yes, this really happened. This model totally explains the “empathy gap“. That’s the phenomenon where someone about faces on a position when it affects someone close to them, like a member of their family. When you’re against gay rights but change your stance when your beloved child comes out, or think COVID-19 is a conspiracy until it hits you and yours, that conflict between caring for your family and other beliefs is enough to kick awareness back online. (This goes right along with Mead’s ideas about consciousness as well [1].) It isn’t the fact of conflict _alone_ , though, that is needed to reboot awareness in these people who have fallen prey. In fact, simple, direct conflict can actually **harden** an existing belief system, making it less likely for any kind of change or introspection to happen. Cult-like groups often view people arguing with them as a badge of honor and a sign of belonging. When religious groups send out missionaries, it’s less about conversion and more about making the missionaries feel isolated from everyone who is not a member of the cult. It makes it _easier_ for the cult leader; the rank and file feel isolated and rejected by the outside world, so they tune out all information coming from anywhere other than the cult. We’ve seen this process happen with the political right in the United States. We’ve seen it with politicians who didn’t fall in lockstep be called “RINOs” (akin to excommunication), with the divisive rhetoric of the Tea Party, with the blatant doublespeak of Trump, with his regime’s constant cry of “Fake News”, and most recently with the idea that an anonymous person on the internet (this “Qanon” crap) is the only one telling the “real truth”. Not like there’s a TV show about pretending to be someone you’re not on the internet or anything. This is where the biggest problems arise. Not only do these systems take advantage of people, but they encourage unreflecting knee-jerk reactions. They isolate and harden people against _thinking_. They _actively discourage_ listening to other voices than their own. **They. Create. Zombies.** [4] ## What do you do with a philosophical zombie? Which brings us back to conflict. Obviously, it’s not _just_ conflict. If the conflict is anticipated and the person has a stock response ready to go, conflict doesn’t have much of an effect. (This seems to be the case with people in general, not just philosophical zombies.) It’s possible that being subtle in your arguments or appealing to common values _may_ make your words have more of an impact, but _only_ if people don’t think you’re trying to change their mind. [3] And that’s pretty difficult when you are, after all, trying to change their mind. So direct conflict doesn’t always (or even _frequently_) work with regular people, let alone philosophical zombies. To be able to get through to the non-conscious, the message seems to require three things: * It can’t be interpreted as a talking point other zombies have already answered. (It does _not_ matter if that answer makes absolutely _zero_ sense.) * Sufficiently high stakes so it cannot be ignored. * A personal connection (or themselves) being effected by those high stakes. ## Kill Means Kiss The empathy gap indicates that when those stakes are high enough, and personal enough, that there’s a chance awareness will kick back in, knocking the person out of their zombified state and allow them to re-evaluate decisions. [5] Depending on the severity of the “zombie infection”, the person might be able to resume normal thinking and awareness. (This is what happens at the end of _The Brainwashing of My Dad_ , for example.) See also Ohioans: If You Think Booze Should Be Legal, You Have No Reason To Oppose Marijuana Legalization This sounds a lot like exaggeration until you hear people who have left cults (religious, social, or political) try to describe how they thought while were still in that cult. The way they thought seems alien and mystifying – to the same person who _thought it_. They literally can’t make sense of what they thought and believed before. I’ve had a sinking feeling for a while now that all this I just described is not just a thought experiment. It matches up with what’s going on a little too well. Nobody is coming to save us from this kind of zombie apocalypse. Nobody else _can_. Outside influences don’t have that personal connection. Things like articles and blog posts and television shows and speeches won’t cut it with a philosophical zombie. Instead, it is up to each of us. We must confront those succumbing to these cult-like ideas. It is not enough to remind them of others. It is not enough to just appeal to their empathy. _**We must teach them that “kill” means “kiss”.**_ I’m paraphrasing _Pontypool_ , but the point is actually spot on. It is _not supposed to make sense_. It _cannot_ fit in the expectations of the zombie cult. Right now, the cult-like programming these individuals have undergone (whether through the TV, talk radio, the internet, whatever) actively resists **any** kind of logical or emotional argument that you might have. It is _not_ enough to just try to refute what they’re saying. It is _not_ enough to just point out the flaws. Instead, you – yes **you** – have to completely confused and side-step their pre-programmed responses. As tempting as it might be to use facts, or appeals to empathy alone, you will need to step completely outside the box. Break the pattern they’re expecting. Subvert their paradigm. Keep them off balance. Make sure they _have_ to think to make _any_ sense of what you’re saying and doing. It isn’t easy. But I think it is _necessary_. Tell them that “kill” means “kiss”. And keep fighting back while keeping your soul. Featured photo by Jeremy Keith under a Creative Commons Attribution 2.0 Generic license. [1] Mead’s conception of consciousness implicitly and explicitly requires compassion and empathy for others, or at least the ability to determine the motivations of others. For him, this is exemplified through the saints and (interestingly) capitalism, since to form a better communal society with increased benefits for all, one must be able to empathize with the other. (Capitalists who don’t figure out what other people need and want don’t sell a lot of stuff.) [2] This is actually the point of the _Blade Runner_ franchise. [3] My own masters thesis seems to have measured this effect. [4] Yes, like _Pontypool_. Go watch _The Brainwashing of My Dad_ and tell me that doesn’t sound like a _Pontypool_ situation. (And yes, that’s where “Mrs. French’s Cat” in the title comes from.) [5] Yes, I’m assuming that the original philosophical zombie’s decision was morally wrong. After all, if it made sense morally, why would you need to avoid reflecting on it?

Even if you don’t want to commercially sell your writing (or other creative endeavor), if someone else wants to use it, _you should have a contract_. And while that contract doesn’t need to be super-fancy, there are several parts that it needs to contain. This isn’t just a hypothetical; I’ve got real-world examples here (though my privacy policy is in effect). **Also, I am not a lawyer, I am not _your_ lawyer, and this is officially _not_ legal advice, as I am not a lawyer.** That said, this is the stuff _I_ look for as a layperson and as an author, and what I made sure was included when I was a publisher. ## Even If You’re Not Getting Paid, Only Grant Others Rights With Intention I have a friend who writes well, but has absolutely zero interest in pursuing being published. While they call their work “fan fiction,” it’s fan fiction about their own original character, so there’s no rights complication, which makes this example all the better illustration for us. They post their fiction on one of the fan fiction sites, but it’s behind a login due to it being of an sometimes adult nature (in the sense of problematic relationships as well as naughty naughty smutty bits). They got a request via social media to repost the fiction about that character elsewhere so their friend could read it. Yes, it was that vague: “your writings about [Character Name].” I told my friend, in no uncertain terms, that they should absolutely _not_ just agree, even if they have no intent of selling the work or trying to get it published. My reason was simple: There was a lack of clarity as to where it would be published, how long they’d have the rights to _keep_ reposting (or republishing) the work, which works specifically or the _entirety of their fiction related to that character_ , whether that person could then _charge_ others to read the work, whether minors would have access to the work, and whether it would be put somewhere that AI agents could (or would) train on the data. Regardless of the stated reason (“so my friend can read it”, which is odd since it’s posted to a fanfic site already), without a contract there’s no specified limitations on what someone else can and can’t do with your work, and for how long. If the person’s on the up-and-up, then writing down those limitations is not going to bother anyone. Perhaps it may seem like overkill, but since you’re just putting in writing what they said they were going to do anyway, why should they care? For the conflict-adverse out there, the point of the contract here is to _avoid_ conflict down the line. **By being clear about what is happening and why, there’s less room for misinterpretation,** **with a goal of avoiding conflict later on.** See also Review: Viewing Yellowjackets (Showtime) From A Certain Point Of View Which is, unfortunately, a good segue into the other example. ## Whether Through Incompetence Or Malice, Bad And Missing Contracts Cause MORE Pain And Conflict A different friend, Sue, was approached by Glenn a few months ago. Glenn (2) was developing a product for tabletop role-playing games, and wanted to use Sue’s setting as part of the product. Glenn also approached me, and I declined to participate. After seeing the contracts now, that was absolutely the right choice. _Because there essentially is not one_. There is an NDA, yes. It is extremely poorly drafted, overbroad, and violates requirements that it itself invokes (specifically the Trade Secrets Act; the NDA omits required whistleblower protections required by that law). The only portions that actually pertain to Sue’s IP in the slightest are these two bits: > Confidential Information includes any and all data, assets, or logic contained within the [REDACTED] Google Drive directories, Discord servers, Slack channels, and Canva Business ecosystems. This includes, but is not limited to [list of game mechanics]. …. Working Documentation: All internal files, chat logs, prototypes, and asset pipelines accessible via the secured corporate environments. Simply saying “if you post this in this channel, it becomes our trade secret” does not automatically make it so. The fact that you’re in discussions to _acquire_ that IP, yes, absolutely. Already existing work **itself** does not automatically become **someone else’s** trade secret because you discuss the IP in their communication channels. (1) That’s a patently ludicrous claim. An additional “deferred compensation & transition agreement” further muddies things by calling “participants” volunteers until the company secures external funding, at which time they “immediately transition into a formal Salaried Position [sic] at a rate commensurate with my level expertise and role mapping.” Plus a signing bonus that, when funding is secured, “serves as the immediate compensation for loyalty during the inception era,” whatever that might mean. Your alarm bells should be going off. There is no mention of when, exactly, that bonus will be paid (despite the phrase “immediate compensation” it just says that it “serves as” compensation. There is nothing specifying “within 60 days of angel investment” or the like, and it sure sounds like payment amounts are going to be entirely at someone else’s whim _at that time_. It _is_ possible to say “I’ll pay you later for work now, based on future income.” I’ve done it myself when running Kickstarters. And I clearly specified what portion of royalties (if any) or the flat rate or flat payment for the work would be, contingent upon the Kickstarter being funded. See also Learn To Write Better From Films (feat. Morbius) These laughably bad documents do none of these things. And you’ll notice that there’s still a whole lot of nothing about intellectual property for the _setting_. While there’s some verbiage in an “onboarding document” about game _mechanics_ as trade secrets (1), the whole relevant section is this: > III. INTELLECTUAL PROPERTY & TRADE SECRET NOTICE All verbiage, game mechanics, mathematical formulas, and linguistic structures presented during your tenure are the exclusive property of [REDACTED]. > > Non-Disclosure: You are privy to a 50-year trade secret discovery. Unauthorized disclosure of any logic core, is a direct violation of international trade secret law. Glenn’s “onboarding document’s” overbroad “linguistic structures presented [sic] during your tenure are the exclusive property” is just ludicrous. It is never defined what “linguistic structures” covers or “presented” means; is he literally trying to claim that he owns parts of speech if you just say it around him? “Logic core?” I’m pretty sure that logic is not a trade secret. Please. At best, it showcases Glenn’s ineptitude and inexperience, at worst, it showcases an attempt at a predatory rights grab. It is worth noting at this point that Glenn allegedly **PUBLICLY** posted a non-trivial amount of Sue’s not-previously-published work that was produced prior to her working with Glenn, without express permission, and without attribution to Sue as the author. Prior to posting it, Glenn had refused Sue’s attempts to get clarification as to rights and permissions. While Glenn’s incompetence (or malice, or both) is largely to blame, Sue also should have known better. ## The Parts That Were Needed 1. Who the contract is between, _exactly_. 2. What work(s) the contract covers, _specifically._ 3. What rights, explicitly, are being granted to the work by the creator, and that **all other rights are excluded**. In general, _the entity being granted the rights should be able to use them_. 4. How long exclusivity and non-exclusivity last. 5. The context in which the work is going to be used. 6. A reversion clause and/or kill fee. You can see all of these portions in my old contracts, which I have kept up for exactly this reason. I never asked for film or audio rights as a publisher, because I was not going to use them and did not have the setup to do so. I asked for exclusivity for a period of time, and non-exclusive rights… but _only_ as part of the anthology that a story appeared in. That meant I couldn’t just post the story elsewhere, or release it separately, or train AI with it (even though that didn’t exist when I made these contracts). There’s an explicit reversion clause, so I couldn’t just sit on a work forever. There’s an audit clause. And so on. An important addendum: Some publishers _will_ try to get unspecified rights, or rights to methods of distribution that have not yet developed. Whether or not that publisher is compensating you sufficiently for those future rights is your decision. See also Value Flows Toward The Author. Always. Because Sue did not hear the alarm bells early on, she’s facing a _lot_ more difficulty, grief, and heartache now trying to fix the damage caused by Glenn’s abuse of her creative content. ## Ask For Changes, Clarifications, And Specifications Contracts, in general, should be regarded the same way you’d regard a “Wish” spell, or an agreement with the Fae. The wording of _how_ you agree matters as much as the agreement itself. **Specificity is your friend.** If you find yourself in a position like Sue’s, please feel free to use my old contracts as a starting point for negotiation. **If you’re dealing with substantial amounts of money,_consult a lawyer versed in contract law_ prior to signing anything, even if a contract meets all the marks I point out here.** As a final note, it _is_ possible to ask for changes to a contract yourself. The contract for my first “pro” fiction sale came with an exclusive “in perpetuity” clause. If I had just signed that contract, _I would not have the rights to republish my own story, even though it came out 17 years ago_. Instead, I asked for it to be modified so they had five years of exclusivity and non-exclusivity afterwards to use it as part of the anthology only. **As far as I know, I’m the only author in that anthology who has the rights to my own story again.** Because I asked. So do not hesitate to consult with a lawyer. Do not hesitate to ask for provisions to be changed. **And if they’re suddenly really nervous about writing down how they are going to use your work, and refuse to clarify,_walk away_.** Featured Image by Andreas Breitling from Pixabay * * * (1) I am not discussing their “trade secrets” or naming them because, well, they’re not exactly trade secrets, no matter how many capitalized letters you put in there, or how much you claim they are. It’s stuff that I’ve seen in TTRPG circles (though perhaps less formally) for over a decade. “Glenn” has formalized what I would call a bunch of house rules familiar to many who play TTRPGs and added a bunch of hype and capital letters. Yes, I refer to this same footnote twice. (2) In case “Glenn” sees this and recognizes himself: Hey dude, you want to out yourself as the person I’m criticizing here, go for it. Or you can take this as a chance to get your house in order quietly, so that people don’t connect you with what I’m relaying here. Your decision.