Regulations and Retreats

I foresee a near future where groups of people do with mental labor what the Amish did with physical labor, which is to say, “Here is where we draw the line. We choose to live on the this side of it.”

I’ve been a big fan of the Amish for a couple decades now, mostly as a distant observer, not as someone who actually knows what the hell he’s talking about. What I know of the Amish began in conversation with Kevin Kelly. And more lately through my writing partner Matt Mikalatos. Matt and I were working on an Amish AI TV show for Sony for a few years, and that process taught me a lot.

Here’s how Matt puts it, to paraphrase him terribly: “The Amish aren’t against technology. They’re just mindful of how technology impacts the family and community. They don’t have dishwashers because for them washing dishes as a family is time spent together, talking and bonding. They don’t have cars, because they don’t want families to get so far apart in a day that they can’t have dinner together every night.”

When put like that, the Amish sound like geniuses rather than luddites.

I grew up on a farm, tossing hay bales, repairing broken tractors, welding together shoddy fixes to bent plow blades, driving on roads before I could really reach the pedals. Life was hard, but there was joy in the hardness of it all. One of the things my wife and I love about living on a sailboat is that we are constantly having to maintain systems, repair things with our own hands, do things a harder way than most people. We find pleasure in the small struggles. The Amish have that as well.

Right now, machines are getting better at cognition in much the same way that machines a century ago got better at physical work. The Amish decided they wanted to preserve a large domain of physical work for themselves, because it kept their community more tight-knit, made them happier, etc. The reaction I’m seeing to AI-anything right now makes me think that we will soon have similar enclaves for cognition.

These new anti-AI enclaves might not be as obvious or formal, of course. They might take the shape of cliques. Or wings of political parties. Or social media silos. Or just individuals making their own choices about how AI will and will never touch their lives. But I wouldn’t be surprised if down the road (and not too far down it), people will want to live in a country, county, town, or neighborhood that puts limits on artificial cognition similar to how the Amish put limits on artificial labor.

What might some of those choices look like? Where will lines be drawn in the sand? I’ll tell you about some of my choices (not that anyone asked), and I’ll bring up some interesting examples I’ve heard and thought up recently. For me, this is more complex than “AI all bad” or “Bring me the singularity now.” In fact, a lot of what I’m witnessing makes zero sense to me and is full of hypocrisy and contradiction. But we’ll get to that.

My dad loved working with his hands. One of his many hobbies was building furniture. He would spend months on a single corner cupboard, with complex joinery, expensive wood, floating door panels, you name it. All the fine details. Something I noticed over the years: he loved building the first one and absolutely hated the drudgery of building the next 5 (every kid needed one as well). Repetitive tasks filled him with dread, while new creative tasks fired him up and kept him in the workshop into the wee hours.

There was a chair design he fell in love with at a craft show once, a folding chair made out of twenty or so one by twos with just a couple of long bolts holding it all together. He came home and replicated the chair, with a couple of improvements. Then he realized he might have a product here, so he went on to streamline the process with a bunch of jigs. He ended up making thousands of them (they filled a tractor shed, piled to the roof), and he did nothing but complain and never made a penny of profit, despite years of trying to sell them at every craft show within a thousand miles of our farm (seriously).

My dad was a mix of a craftsman and a capitalist. He loved making things by hand, the hard way, but he would’ve loved to have wiggled his nose to duplicate the work so he never had to do the same thing over again. He “borrowed” (pretty close to stole) someone else’s design and tried to profit off it. He made a lot of furniture simply by following detailed plans, with no real creativity or originality. Just finely honed skills and a very expensive set of tools. My dad would’ve loved the things AI let him automate, and he would’ve never stopped being creative and working with his hands. He probably would’ve thought of ways to allow AI to make him more creative. Instead of building according to a plan in the back of a woodworking magazine, he would’ve gone back and forth with an AI agent to come up with plans for something that has never existed and built that instead (my dad never baked to a recipe; he always had to tweak it to make it his own). AI would’ve amplified my father’s creativity, not replaced it. And this is how a lot of people are working with AI today.

Let me tell you how AI has increased my creativity. I do all my writing these days inside NEO, a program I created with Claude. I first wrote a blog post about NEO here, almost a decade ago. Even before writing the blog post, I’d been looking for a piece of software that fit my writing needs. I dreamt up the idea of a Darlings tab, because I always had a separate Word file for every WIP of deleted stuff I couldn’t let go of but really didn’t need. I wanted post-it-notes in my prose for things to come back to so I could forge ahead and keep writing (I used to just put XXX in a draft and do a search for them after). I thought up the idea for an ENTER ENTER ENTER workflow for creating section breaks and new chapters, never having to grab a mouse to format things center, change the font size, etc.

I tried to get NEO made three times over the years. Once was with a developer who believed in the vision. But I was sailing across the Pacific with limited internet, and the development was slow and messy. I once tried to convince a room of Amazon developers to make it internally, at a meeting expressly about this in Seattle. My dream then was that authors would write inside KDP itself, in my ideal novel-writing environment, and then could publish straight to ebook and print-on-demand. It would make it easier for Amazon to ensure quality, and authors could do simple edits right from KDP. Links to author profiles, About pages, “other works” would all be baked right in. Amazon seemed to like the idea. They never did anything with it.

The third time I tried to make this was earlier this year. One of the things AI agents have gotten very, very good at is coding. So I took that old blog post and gave it to Claude, and a few hours and iterations later, I had a very buggy prototype that was already better for my writing needs than anything on the market. My writing output soared. Everything I had intuited about my writing needs turned out to be accurate (funny how writing 20 novels will give you that kind of insight). It wasn’t perfect, but I kept tweaking it to make improvements. I learned a lot along the way. I put hundreds of hours into NEO, making it my dream application. All the heavy lifting, of course, was done by Claude. But was it…?

When I was in college, I minored in computer engineering as a physics major (this was before I switched to English). I took two semesters of programming, mostly learning Java and HTML. We were introduced to Object Oriented Programming. One of the concepts of OOP was that instead of writing every line of code from scratch, you build modules, classes, and objects that you reused. Libraries of objects could be borrowed between projects. Often times, you didn’t solve a problem that had already been solved. If you needed a random number, you could call on a random number generator, use the output, and get on with your day. You rarely built from scratch what someone had already tackled. A lot of coding in class was copying and pasting things from a google search in the right order and then stitching them together.

The more you learned to code, the more you realized that half the battle was knowing what problems had already been solved and where to get those solutions. It was fun, figuring out how to get a program to work. Like working a crossword puzzle. Like building a corner cupboard by hand. I loved it. The same way I love working a Sudoku. But no part of me dreamed of doing it for a living. I didn’t love it enough to want to master it. Not like I loved writing novels.

I shared NEO with some friends a while back, and one of them told me it was his new daily driver. He was obsessed, but he had some thoughts and feedback. I agreed with most of it and made some changes. Before long, more than a few people were using NEO and saying it had supercharged their writing. I mean, it’s not for everyone, but it’s also not for no one. So I made a small announcement, shared a link on my blog, and went back to writing stories and tweaking my writing environment (ya’ll, I was spending less time making NEO exactly what I needed it to be than I used to spend googling and watching YouTube to try and figure out how to get Word to just behave). My time spent fiddling with my writing environment went DOWN while my writing went UP.

If you’ve been around these parts a while, you know how much of my career has been spent trying to motivate other people to get stories out there. Everything I know about writing and publishing I give away for free on my sidebar there, under Writing Insights. I’ve never tried to sell my expertise, or sell a book on writing, or make a penny trying to help others. Not interested in any of that. I’m interested in stories floating through the ether. I’m lost in my own imagination. So I made NEO free, kept it open source, and have allowed collaborators to come in and make it better. It’s now available in a lot of languages. There are features hidden away inside it for power users without getting in the way or even being visible to the rest of us. A YouTube video I put out explaining how to get it and how to use it has over 130,000 views.

Here’s something that might seem unrelated at first, but I’ll bring it all together: If you write on a Windows machine, you’ve probably used Arial or Book Antiqua at some point. Both are pixel-by-pixel ripoffs of fonts that used to be expensive to license (Helvetica and Palatino). Craftsmanship and talent went into creating those fonts, then someone came along, copied them down to the pixel, then used a copyright loophole by changing the code to create the font in order to package them in Microsoft Windows and Office. Apple machines use stolen fonts as well. Linux users have the Liberation family of stolen fonts. These are just a few examples out of HUNDREDS. I can guarantee you without a doubt that every single author screaming about the theft of art has written books entirely in fonts that were stolen from other artists.

Computers themselves are like my father’s jigs for duplicating his labor. REAL BOOKS were written by hand by monks, one line at a time, with real ink. Gutenberg cheapened the whole process. Mass produced books? Print slop! But really, we should be chiseling our thoughts in stone. And no one should be reading anything, because that’s how ideas leak from one brain to another. Humans are so good at rationalizing their hypocrisies that I promise you someone using stolen fonts will learn about this, realize they’ve been doing it, and decide that in this case it isn’t important. While being vehemently anti-AI anything. Humans are strange.

Here’s a point brought up by a commenter on a different post (T.R. Thorson check out his book here). Thorson pointed out that the Star Trek computer could recite any book ever written, and somehow fans of the show (some of them writers) never once jumped out of their sofas to scream at the injustice of a machine having read all books. The anger isn’t the reading. And I promise you it isn’t the not-being-paid for the reading. If every author in the world got list price once for each of their titles that went into the training data (let’s say $100 for a prolific author, but even $1,000 for this thought experiment), you think that a thousand bucks would’ve made them perfectly okay with AI writing fiction? C’mon. That’s just insane. For most authors, there’s no amount of money that would make them okay with AI writing fiction. That’s a machine moving into their line of work! Sure, they’d take a million dollars in a settlement, but then they’d go right back to shouting about the evil of AI doing art. Getting paid isn’t the thing.

It’s not about the money, the theft, or the fact that a machine read their book. You can resolve all these things, and 99% of people are going to stick to their current line in the sand and their current level of anger. The raw feeling of a machine automating what you spent years honing is visceral and it angers the fuck out of people. I faced this when I self-published. Authors who had gone through the hellish path of writing query letters, getting an agent, going out to submission, finally getting a publishing deal, waiting a year for the damn book to finally come out, watched me press SUBMIT on a website, and a paperback book showed up at my house five days later looking like any other traditionally published book. Heck yeah that makes people angry. I mean, they are silly and unjustified in being angry, but they’re still angry.

An aside: There are two types of people: Folks like me who paid off my school loans while making $300 a week shelving books who were THRILLED that other folks were having their school loans forgiven by the government, and people who have suffered who think other people should suffer. Not an AI point. Just a human point.

Back to NEO and using AI to program a tool that makes writing more enjoyable, not just for me but for thousands of others (I’m getting emails and comments from people who are writing their first book with NEO after years of not feeling inspired!). The comments on the YouTube video I posted above are overwhelmingly positive. Now and then you get someone shouting about AI slop. I find these comments fascinating for a lot of reasons. Not least of which is the assumption that coding is the difficult part of programming. It’s not. It’s the monkey part. When a tech company hires people, the coders are the bottom of the pipeline in product development. Product design is the top of the chain. Product managers come next. Coders are far more replaceable.

Product design’s equivalent in writing would be the person who comes up with the idea, outlines the entire plot, writes all the dialogue, has pretty much the entire story in a very detailed draft. This is what James Patterson does, and he’s very good at it. Product managers are the people who take this idea and all the creativity, and they figure out how to bring it to market and where to allocate resources and assign roles to make that happen. This is Patterson’s publisher. The coder would be the freelancer Patterson hires to turn that detailed draft into an actual book, adding a bunch of “he said” and “she saids” and more description to scenes. Fleshing it out. Something almost any decent writer could do.

Patterson (and product designers) make the most money. The publisher (product managers) make the next most. The freelancer (coder) makes the least.

Of course, there are exceptions to the above, but they are the exceptions rather than the rule. There are coders out there who are more like product designers in their vision and creativity, and they can get huge salaries and bonuses. But that’s because they are doing something hardly anyone else is doing. That’s the creative part they bring. In the writing world, there are script doctors who get paid 7 figures just to edit a screenplay to fix all the problems in it. Sometimes a freelancer can do a thing hardly anyone else can do. They might make more money editing a script than the original screenwriter was paid to write it!

But the point stands. Here’s the truth about developing NEO, and the reason no software program like this already existed, and the reason thousand of people are now using it and raving about it: The hard part was never the coding. The hard part was the 20 or so novels I’ve written, the dozens of short stories, all of it with applications that frustrated the heck out of me, while I was mentally making notes about what the dream writing application would look like. Take any amazing coder out there and have them create a tool for writing novels, and you’ll get another Microsoft Word or Scrivener. Bloat. Feature creep. Distractions. Bug-free, maybe? But useless for my needs.

NEO is an example of a different kind of line in the sand, one I find more interesting, and similar to perhaps how my father would’ve used it in his woodworking. An Amish kind of line. The Amish are not against technology. Wheels are technology. A butter churn is technology. Barns and clothes are technology. They are just primarily focused on COMMUNITY and FAMILY. What things strengthen them and what things weaken them? That’s a noble framework.

NEO is without a doubt contributing to creativity in the world. I know this because I’m being more creative. The bonus is that I’m hearing from others who are being more creative. My laptop also makes me more creative. Google has made me more creative. Stolen fonts I’ve used over the years have made me more creative. Print on demand and e-books have added to the store of art in the world. So has fan fiction. When you have the right framework, a lot of complicated decisions suddenly become clearer. If you are AI all bad in life, how are you going to navigate the world ahead? Are you going to use Google Maps? Are you even going to allow yourself to Google? Are you going to learn which fonts are based on stolen labor and ban those from your life? How weird and complicated are you going to make things for yourself?

If you like to write, keep writing! If you like to paint, keep painting! If a big tech company stole from you, join the class action lawsuits! After you get that money, keep being angry at them for stealing from you! If you don’t want your books read by machines, put a clause in each one with a promise to sue if it continues reading past that point. If you want to create a community that shuns all AI, I think there are good models for this already and I’d be very interested in visiting you, perhaps staying for a year or two at a time.

Maybe the most important thing to learn from the Amish is this: They don’t go around yelling at people for living some other way. In fact, they encourage every member of their community to go out and try those things for a year. They have a whole routine about it. They also try new technology all the time to see if it’ll make their community stronger. Just like authors try new things all the time (computers, ebooks, audiobooks) to see if it’ll make their creativity richer. When I see someone pointing and yelling, I usually just block them or delete the comment these days. It doesn’t get a rise out of me, but I typically pick up trash when I see it on the sidewalk and toss it in a bin. I don’t leave it around to get smelly and upset others.

Okay, now that I’ve upset the anti-AI crowd, let me really piss off the libertarians and any MAGA out there who might have stumbled into my blog on accident by touching on the Regulations from the blog’s title, because we are going to need a LOT of AI regulations from government. Not just: who owns the data? How can it be used? But more like: How do we restructure the tax code so the concentration of power and wealth doesn’t end civilization? How do we provide housing, food, and a sense of purpose to people who can no longer find gainful employment or afford to feed their families or purchase a home? I’m talking New Deal levels of government restructuring. The EXACT OPPOSITE of what we’ve been doing.

Somehow, half our population has convinced themselves that dark/gay/trans people are making their lives poorer, when it’s actually corrupt politicians, corporations, and the wealthy doing that. This is fact. But fact doesn’t dent feelings. And if we don’t fix this, nothing else will ever be patched up. We need strong unions, labor protections, an aggressive tax code, a wealth tax, a massive inheritance tax, and elected officials who want to improve our lives and have clever ideas on how to do so. (Newsom signed a bill days ago making it easier for pop-up businesses to have access to abandoned storefronts. A common-sense approach to allocating resources and empowering people who are trying to empower themselves). Not all regulations are anti-business. Some are very pro-business. The best are pro-people and pro-business.

Other things we need to change immediately or we are toast: Abolish the Senate. You can’t prioritize land over people. “Middle America” right now is someone who lives in the suburbs. The “Average American” is a city dweller. Protecting minorities is our duty. Governing for a minority is a terrible idea. Right now, we are getting both of these wrong.

Term limits for all offices, especially the Supreme Court. A massive limitation on the power of the Executive branch (we have limitations in the books that are currently being ignored, like the ability to start wars, the ability to profit off the Presidency, the ability to commit treason). Without this one, we will remain a banana republic. We 100% already are. Rules don’t apply to people in power, and that has to end yesterday. We have to get money out of politics (reverse Citizens United, punish bribes and pay-for-play). We have to ban gerrymandering. DC and PR need to be states. Those are US citizens! Get rid of PACs completely.

Every human on Earth should be guaranteed a roof over their heads, three meals a day, and healthcare. Every single one. That should be a law. We aren’t allowed to spend on other things unless we have those sorted. No more datacenters, rockets, wars, or yachts until there’s a home for each person who needs one, a place for them to get fed, and a place for them to be healed. We start there and then get to the extras. It’s insane that I have to even type these words. Or that the followers of Jesus didn’t hear a goddamn word he said.

What does this have to do with AI, you might ask? Everything. People aren’t angry at AI for reasons that make any sense to me. But people are angry in general for reasons that make perfect sense. They just don’t know who to be angry at or what to do to stop being angry. In most cases, I don’t think they even know why they’re angry. A real work of priceless art was shown to people and told it was AI. Those people spat in disgust at how horrific the work was. They critiqued everything about it. It was a Monet. The art had nothing to do with the anger. People on the right are angry at immigrants, who are a net positive to society (they contribute more than they take; they commit fewer crimes than native-born people; they do jobs natives don’t want to do; they enrich our culture; they arrive at an already-productive age without having been invested in during their primary years). People on the left are angry at AI, because the same companies that have destroyed our societal bonds are the ones bankrolling it. Or any of the many misconceptions about AI and creativity listed in the top half of this blog post.

Folks are just angry. I get it. But I don’t feel it. I’m more confused than anything. Confused that we can’t all come together and agree on common sense solutions to actual problems. Confused at how easily cults arise and wreck society. Confused that the biggest block of voters don’t even vote. Confused that strengthening our communities, our families, and increasing the amount of creativity in the world isn’t an obvious line to draw in the sand, rather than pointing at people and screaming a bunch of human slop. We are in danger of becoming the stochastic parrots. If our humanity is truly at stake, we first need to prove that we ever had any.

The post appeared first on Hugh Howey.