Book Brew 179: The Good AI-Heroin Era Is Ending—Here's Your New Addiction

We've been living in the Wild West of AI for the last few years. No sheriffs, no rules, no gatekeepers. You want to call upon a model that can write your marketing copy, summarize your documents, generate code, or help you with things that probably shouldn't be that easy? Free rein. No questions asked. No liability. No guardrails.

The Wild West didn’t last and neither will this. Someone builds a town, makes rules, and the party ends.

Recently, the US government blocked Fable 5 and GPT 5.6 until someone figures out how to make them play nicely with safety regulations that don’t quite exist yet. The “heroin” is getting cut off and you are being moved to “OTC Tylenol”. And many are having a withdrawal tantrum.

The New AI Classification System

We're entering a regulatory era for AI, and it looks a lot like how the FDA handles pharmaceuticals. Some models are going to be Schedule 1 (heavily restricted, prescription-only, monitored as hell). Some will be Schedule 5 (over-the-counter, anyone can use them, minimal controls).

The resistance has been born out of people who are addicted to the over-access they got used to.

"Loss aversion refers to the relative strength of two motives: we are driven more strongly to avoid losses than to achieve gains." - Daniel Kahneman

Just a few short years ago, you couldn't imagine having a model in your pocket that could generate a thousand words on demand. Now it is fully expected. The novelty wore off and became a need (not how I would personally describe it, but that is for a different soapbox). Now that it's being gated, it feels like a loss. Enter the psychology bias of loss deprivation. You feel the pain of losing something way more acutely than you'd appreciate the safety of it being taken away.

The pharmacy industry has changed a lot once the Controlled Substances Act went into effect in 1970 and I think that is about where we stand with AI - on the cusp of a Controlled AI Act of sorts.

I truly don’t see regulation as a problem. I think access should have been restricted from the beginning. But that becomes a double-edged sword - without all the open access we have all been so happily using since 2022, the AI tools wouldn’t have had the mass usage to be able to train on and improve so quickly.

3 Reasons to Stop Renting OpenAI's Brain

"Organisms that treat threats as more urgent than opportunities have a better chance to survive and reproduce." - Daniel Kahneman

I think what we are going to start seeing is a redirect. Instead of relying on access to the biggest, baddest, unrestricted models, you're going to have to build your own internal infrastructure: fine-tuned LLMs and SLMs trained specifically on your business data.

Three reasons this matters:

1. Control → You're not reliant on what the government decides to allow or restrict. You're running your own model, fine-tuned on your own data, behind your own walls. Regulation doesn't touch you because you're not using their models, you're using yours.

2. Competitive Advantage → A model trained on your business data, your playbooks, your voice, your processes can't be easily replicated by your competitor downloading ChatGPT. That's your AI, running your operations.

3. Accountability & Liability → The businesses panicking about regulation are the ones who've been cutting corners (using public models to handle sensitive data, relying on external infrastructure for mission-critical work). Once regulation hits (and it will), they're exposed.

This is the move. Not "how do I get access to GPT 5.6 or Claude Mythos?" but "how do I build infrastructure that makes me independent from whatever the government decides?"

Shifting from Capability to Judgment

"Failure is by and large due to not accepting and seeing reality as it actually is. An accurate understanding of reality enables you to precisely diagnose, then treat problems." - Ray Dalio

For the last two years, the conversation has been: "Can we use this model for X?" The answer was almost always yes. So people did. They threw models at everything. Marketing copy, customer support, code generation, strategic planning—all of it, no discrimination, no real thought about whether a model was actually the right tool.

Regulation forces a different question: "Should we?"

"To have the power that only strategy can bring, you must be able to elevate yourself above the battlefield, to focus on your long-term objectives, to craft an entire campaign, to get out of the reactive mode that so many battles in life lock you into." - Robert Greene

That's where Next-Level AI Workshops, Fractional COO/CTO work, and the whole "Quadrupling down on human skills" framework live.

You’re gonna start seeing more needs around:

• Clear decision-making about which processes actually need AI vs. which ones just want it
• Risk assessment about data sensitivity, compliance, and liability
• Systems thinking about how a model integrates into your actual business operations
• Human leadership to decide what matters and what's just shiny

The businesses that thrive won't have the most model access. They'll have the clearest thinking about which models to use for which problems, and the internal infrastructure to support those choices without depending on whatever Google, Microsoft, OpenAI, or Anthropic releases next or whenever they crash next.

"The danger inside is controllable and it should be the goal of leadership to set a culture free of danger from each other. And the way to do that is by giving people a sense of belonging... By creating a Circle of Safety." - Simon Sinek

PONDER THIS

1. What tools have you gotten "free" access to that, if you really thought about it, you probably shouldn't have unfettered access to? Why did you want them in the first place?

2. In your business right now, which decisions actually need the powerful models and which ones could live with restricted or internal-only access?

3. What's actually in your control right now: fighting regulation, or building practices that work within whatever constraints come?

Books/Newsletters

• Thinking Fast and Slow - Daniel Kahneman
• Principles - Ray Dalio
• Leaders Eat Last - Simon Sinek
• The 33 Strategies of War - Robert Greene

The Wild West is closing.

And growing up in this space means knowing which tools actually serve your business versus which ones you were just using because they were free and easy.

We help you move from "Can we?" to "Should we, and why?" and then "How do we build something that works even when the rules change?"

Because regulation is coming. The question is: Do you adapt, or do you panic?

Let's figure it out together. Apply →