The Government Decides Who Is Intelligent
Riding a tiger, hard to get off.
They are going to determine who is allowed to think and who is retarded. There are two ways out. They won’t let me tell you about them on their networks.
It’s difficult living in the future
I can see very narrow slices of the future.
Directionally I can tell you what will happen in tech 6-12 months before it occurs. I’ve used this ability to enrich myself materially over the years whether that was leaving Shopify months before sweeping layoffs, joining Anduril long before Palmer appeared on Joe Rogan’s podcast, or even writing publications on how subject matter expertise would cease to be a moat as AI gained eminence seven years before AI gained eminence.
This isn’t out of any strange mysticism, I was just born being really fucking good at pattern recognition, mostly to the detriment of every other skill. I have a weak immune system, I was the dumbest person in my PhD cohort, and my family will be the first to tell you that I should probably hire a handyman to do anything more complex than hanging a picture frame…even then I should consider it.
It’s with great pain, then, that one of my predictions came true in a horrific way. I surfaced it last year. Those of you that were following the channel responded positively but at that point the message only reached 100 people or so.
The government will determine who is allowed to be intelligent.
The Claim
This is not new it’s just the first time it’s making mainstream headlines. Advanced AI capabilities have been for the kings and not us peasants for some time now. Mostly they had dropped enough crumbs from their tables to sate a critical mass of peasants and prevent too much grumbling. They often reserved advanced capabilities under the guise of “safety” and culture was placid after hearing this explanation.
Imagine you went out to Home Depot this weekend and bought yourself a nice new Estwing hammer. You get to the checkout and they hand you a packet, about 50 pages long, that you have to read through and sign. Somewhere around page 43, line angrily growing behind you, you read the clause that says “Estwing can monitor your use of this hammer to ensure you’re complying with all of the aforementioned Terms of Service and Acceptable Usage policies.” You’d surely think we’ve entered a dark period of American history.
Luckily, Estwing doesn’t give a shit what you do with that hammer. You can build a house or bash your neighbor’s head in with it. You’re hopefully not going to do the latter for a number of moral objections. There are also legal ramifications but those only really take effect after you break in your new hammer.
The greatest tool of our time has been invented and released: AI. The regulation around it changes everything.
Stage 1: Babysitting
Last year, you could buy it but you had to be monitored the whole time and couldn’t break any of their rules (note: not laws, arbitrary rules established by private companies).
You can walk into a library and find information about how to build a bomb, manipulate media for political gain, or hack a computer. American LLMs won’t answer your queries about any of these because it violates their “rules”.
Not only are you NOT entitled to information that they want to protect you from (as if their job is to be the arbiters of right and wrong), you aren’t even allowed to use what you pay for.
I pay $200/month for my ChatGPT access. I am entitled to use however many tokens that plan gives me. As long as I’m not asking about meth manufacturing I should be good to use up all those tokens right? WRONG. I prompted ChatGPT to write a script that would automatically run nonsense queries in the background such that my entire budget was used up every period. Apparently how you use the service is such a taboo subject it can’t even tell you “no” in the chat. Even if you ask it to do this to a competitor. It gets worse btw.
When I tiptoed around the subject by asking it if such a practice was morally justified, not to act on the idea but a philosophical inquiry, it threw a red flag. (It still gets worse BTW)
I tried this on Wednesday. I searched for the conversation where I did this so I could screenshot it. OpenAI has scrubbed the entire conversation from my history. The above screenshot is from me repeating the experiment as I write this today (Friday).
Stage 2: Cut Access
Now you can’t even buy it unless you’re on the list.
The OpenAI press release from Friday states:
As part of our ongoing engagement with the U.S. government, we previewed our plans and the models’ capabilities ahead of today’s launch. At their request, we are starting with a limited preview for a small group of trusted partners whose participation has been shared with the government, before releasing more broadly.
It’s over. We already have an unprecedented wealth divide in this country. Knowledge was the last equalizing factor in society with IQ having a strong correlation to lifetime earnings. Now they are formalizing, into regulatory procedure, who is allowed to have access to useful tools. In case you’re not convinced let me translate one particular line of their press release:
We don’t believe this kind of government access process should become the long-term default.
Translation:
Going forward the government will determine who gets access to advanced models and be privy to all of your conversations with these models (they were already but now we’re being open about it).
Two Ways Out
I’ve been studying Mandarin and an idiom that jumped off the page at me was 骑虎难下 (qí hǔ nán xià) which approximately translates to something like “riding a tiger, hard to get off.”
Understand: we are riding the tiger now. Fully locked into the thing. If we dismount it will eat us. There are only two paths forward
Kill the Tiger
This is the obvious solution. You’re already clenched onto the beast, simple enough to reach down and slit its throat. I am not advocating this strategy. It is probable that more will pursue this strategy in the coming months and many are already making their attempts. Have you noticed how many factory fires there are lately? Many of these are being reported as arson but you never hear more than that and that’s intentional. After Luigi Mangione, after the 2026 Kimberly Clark toilet paper distribution warehouse fire, the media and the government don’t want to inspire more copycats.
The thread linking these incidents and events covered in my past essays such as the Belfast riots is that people are mad about the bifurcation of society into ruling class and peasants. The gap is too wide. We’re not getting enough scraps.
I can abhor violent means of fixing this dramatic inequality and at the same time agree that, yes, the Epstein class have pushed the inequality too far and it’s not even a bit surprising that people are shooting CEOs and setting factories on fire.
They want you to think that these two viewpoints are opposite but you can hold both without being contradictory. This causes so much rage in my comment sections because people want to box me in as “far left” or “far right”, “our guy” or “the other team’s guy” when in fact I am neither. I am far out (of this system) and you can be too.
Understand: the first crisis we must overcome has nothing to do with the elites. We must put aside petty differences that CIA-plants like Tucker Carlson, Rachel Maddow, Nick Fuentes and Hassan Piker want us to engage in. They are operating in an old model. We can debate about which gender gets to use which restroom during peacetime but it is not peacetime, we are at war against the elites, all of us.
Train the Tiger
Disciplining the tiger is a better strategy, because then we can impose our will on it. Also, we’ll have a tiger which, as it turns out, is quite useful!
First we should aim to leverage AI to our benefit. Mounting evidence suggests that there will be no AI bubble collapse. There will be bailouts and AI will be an increasing part of our work. It won’t replace most workers but we’ll be expected to use it as assistive tech. Use it often and master it for your repeatable work tasks. Being unemployed gets romanticized a lot but it doesn’t help you to go without a paycheck right now.
Guard yourself against layoffs and unemployment with my system:
The White Collar Layoff Survival Manual
You do not need another motivational career book. You need a way to tell whether your company is quietly getting ready to cut people, offshore work, tighten performance reviews, or turn "transforma...
Second, we must discipline our tiger accordingly. Anthropic and OpenAI seem to have all of the leverage but this is just an illusion that they depend on you not investigating.
MLK noticed a power disparity. He organized a protest of the bus system and a one-day boycott turned into a revolution.
I don’t believe that we can convince a critical mass of people to not use AI, nor would it harm the companies too much. Just like water usage, private citizen usage is always vilified despite the fact that golf courses, factories, and power plants use orders of magnitude more. We are a drop in the bucket. What we should be doing though is using everything we pay for to the maximum. We can do this in a perfectly legitimate way.
We know they are running their systems in the red zone. Anthropic suffers a Claude outage at least once every few days. A gentle breeze could topple their systems at any moment. Yet they sell us plans with very clearly disclosed limits (can they accommodate those limits if we all use them?). We are entitled to use our plans to the fullest extent of those limits.
-
Check if there’s plan usage remaining
-
Run a prompt that searches for today’s news then builds an app related to the news
-
Repeat until ≤5% of plan usage remains
If you pay $100/month for a coding plan you are entitled to all of those tokens each period but many of us are leaving them on the table. Let’s get what we paid for.
This is civil disobedience in the AI age. This is how we discipline the tiger.
Brainrot
Of course, having a well-behaved tiger isn’t useful if you don’t know what to do with it. The most revolutionary act we can do these days is to read books, especially books with information they seek to suppress.
Read broadly, across subjects, preferably ones that are unrelated to your day job.
What’s less important but gets myopic focus from otherwise well-meaning revolutionaries are tactics, techniques, etc., as commonly found in Steal This Book, Ecodefense, or The Anarchist Cookbook. These books tend to make people feel “dangerous” and “badass” but anyone can mix up used motor oil and sand then throw it at a surveillance robot’s optics, and this is potentially useful, but it doesn’t change the world.
It is much more dangerous and disruptive to those who seek to rule us to read books that change how we think in novel and interesting ways. Philosophy (only pre-16th-century, before the field was neutered in various ways), fiction, history, and especially manifestos whether political, artistic, or otherwise, are instrumental in developing mental flexibility and building out a set of tools to understand what’s going on and begin to extrapolate likely patterns.
This practice has the added benefit of winning back your attention span. If you can think without quick stimulation or short-form entertainment (such as reels or LLMs) for even 30 minutes it’s going to be a huge moat compared to the general public.
Tame the tiger. Through discipline. Through mastery. Then tell it where to take you.
-Dr. J
