The Positive Loop

Late-stage capitalism solved a problem nobody asked it to solve: how do you keep billions of people engaged when wages stagnate, housing is unaffordable, and traditional paths to success narrow every year?

The answer was dopamine. Not drugs. Feeds.

Social media, short-form video, infinite scroll, algorithmic recommendations. The entire consumer internet was optimized for one metric: time on screen. The more time you spend, the more ads you see, the more data you generate, the more precisely the next hit can be delivered. It is the documented business model of every major platform built since 2010.

The result is what we described in The Gameable Society: a population fluent in the language of self-improvement but structurally prevented from acting on it. People who can explain periodization, compound interest, and dopamine regulation. Who know the meta. Who have never played the game.

This pattern maps directly to the behavioral dynamics we documented in Honest Gambling: when traditional paths feel closed, people gravitate toward whatever offers the most legible dopamine. Doom scrolling was the cheapest, most accessible version of that. Zero cost, zero effort, immediate reward, zero output.

We turned ourselves into dopamine-driven animals. Not because we are weak. Because the infrastructure was optimized to make us that way. Every pixel, every notification, every autoplay was designed by teams of engineers whose KPI was your attention.

And then something changed.

In the span of roughly 18 months, a pipeline emerged that nobody fully anticipated, not in its speed, not in its accessibility, and certainly not in its implications for everyday people.

The pipeline is simple:

This pipeline did not exist 18 months ago. Not for regular people. Coding required years of learning. Design required specialized tools. Deployment required DevOps knowledge. Research required academic access or expensive databases. Automation required hiring engineers.

Now a person with zero technical background can:

• Run deep research studies, the kind that used to require a team of analysts, Bloomberg terminals, and weeks of work. Fetch real-time data, cross-reference sources, generate charts, publish findings. From a phone.
• Build functional products, websites, tools, dashboards, automations. Not mockups. Working software that serves real users.
• Automate repetitive work, data entry, report generation, monitoring, alerts. Tasks that used to require hiring someone or doing it manually every day.
• Translate creative visions into real output. "I want a portfolio site that looks like this" becomes a real site in an afternoon. "I need a script that does X every morning" becomes a working cron job.

The speed at which this became possible is the part nobody has processed. It is not that AI is improving gradually. It is that the gap between "idea" and "deployed product" collapsed from months to hours, and then from hours to minutes, and it happened so fast that the majority of people still think "coding" is something you need a CS degree to do.

Here is a thought experiment.

Imagine a video game with the TAM of every single human being on earth. Not a niche. Not a demographic. Everyone.

But unlike every other game, playing it actually does something for your life. It makes you money. Makes you more successful. Helps you translate creative visions into real output. Teaches you things you did not know. Solves problems you thought were too complex.

And there is a DLC dropping every day. New capabilities. New models. New integrations. Every morning you wake up and the game is slightly more powerful than it was yesterday.

And 99.99% of people are not even playing yet.

The gaming metaphor matters because it explains the adoption curve. Games spread through word of mouth, through watching other people play, through the feeling of "I could do that." The reason AI adoption will be faster than any technology in history is that the barrier to entry is not learning a new skill. It is typing a sentence.

This connects to the coordination dynamics we study in Belief-Driven Markets and Coordination-Dominant Assets: the most powerful adoption curves happen when the cost of participation drops to near zero and the perceived upside is visible. Meme coins proved this for speculative coordination. AI is proving it for productive coordination.

Technology has always been confusing. That was the moat. The reason most people never built anything was not a lack of ideas. It was a lack of translation infrastructure between "what I want" and "how computers work."

AI collapsed the translation layer.

You do not need to know what a server is. You do not need to know what React is. You do not need to know what an API endpoint does. You need to be able to describe what you want. That is it. The AI handles the translation from human intent to technical implementation.

This is the black box that changes everything. You type an idea in. Something useful comes out. You do not need to understand the internals. You need to be good at having ideas and knowing what "good" looks like.

The historical analog is the printing press. Before Gutenberg, creating a book required being a scribe, understanding bookbinding, having access to materials. After Gutenberg, you needed to have something to say. The press handled the rest. AI is the Gutenberg press for software, research, automation, and design. The skill shifts from implementation to taste.

Here is where this connects back to the dopamine problem.

For the first time, we have an interface that is as addictive as doom scrolling but produces real output. The dopamine loop is not: scroll → content → brief pleasure → nothing. The dopamine loop is:

This has never happened before. Every previous internet dopamine loop was extractive: it took your time and gave you nothing durable in return. Likes, views, streaks. They vanish. You cannot compound a TikTok scroll session into anything.

AI loops are compounding. The website you built yesterday is still live today. The automation you set up last week is still running. The research you published is still being read. Each session deposits something real into the world that persists after the dopamine fades.

The critical insight: the productive loop is just as addictive as the extractive one. The feedback is immediate. The reward is visible. The progress is tangible. The only difference is that when you come up for air, you have something to show for it.

This is the first time the addictive qualities of the internet are aligned with productive output. The phone, the device that turned us into dopamine addicts, is now the device that can turn us into builders.

The speed of change is obscuring the magnitude of change. Here is what people miss:

Most people have not used AI for anything beyond a chatbot conversation. They have asked it trivia questions. Maybe written an email. They have not used it to build a product, run research, automate a workflow, or ship something real. The gap between "I have used ChatGPT" and "I use AI to build things" is the gap between having a phone and building an app.

The analogy is early internet. In 1995, most people thought the internet was for email and maybe some research. The idea that it would restructure commerce, media, communication, and culture was not on the radar for 99% of the population. We are in the equivalent of 1995 right now, except the adoption curve is compressed from decades to years.

The under-allocation we documented in Observed Multipliers (extreme outcomes clustering in compressed windows) is about to happen in productivity, not just financial markets. The same structural conditions that produced 100x price moves (low barriers, network effects, compressed time) are now producing 100x productivity moves. But because productivity gains are less visible than price gains, the market has not priced this in.

Abstract arguments are less convincing than concrete examples. Here is what the pipeline actually looks like in practice:

Each of these examples has the same structure: human intent → natural language → AI translation → real output. The translation layer is the breakthrough. Everything else follows from it.

Here is the part that matters for human wellbeing, not just productivity:

Doom scrolling persists because it is the path of least resistance to dopamine. You are tired, stressed, bored. You open your phone. Infinite content appears. Your brain gets what it wants. The activation energy is zero. The reward is immediate. The cost is invisible (time, attention, mental health) and deferred.

AI tools have matched that activation energy. Open your phone. Describe what you want. Watch it happen. The reward is equally immediate. But instead of losing 45 minutes to a feed, you spent 45 minutes building something. The dopamine is equivalent. The output is not.

The behavioral shift is already visible in early adopters. People who discover vibe coding or AI-assisted building report the same compulsive usage patterns as social media: checking in constantly, thinking about it when away, feeling pulled back to it. Except the output is products, research, automations, and skills instead of scroll sessions and parasocial relationships.

This connects to the core dynamic we identified in The Gameable Society: the spectator class exists because consuming optimization content was easier than doing optimization. AI eliminated the gap. The effort required to do is now comparable to the effort required to watch. For the first time, playing the game is as easy as watching someone else play it.

If this analysis is correct, we are at the beginning of a structural shift in how humans spend their attention. Not a gradual shift. An abrupt one, compressed by the same dynamics that compressed crypto adoption cycles, information propagation, and cultural change.

The pattern across our research, from speculation pressure to belief-driven markets and gameable society dynamics has consistently documented how technology reshapes behavior faster than institutions can adapt. AI is the most extreme instance of this pattern. It is not making existing processes faster. It is making previously impossible things trivial. And the people who internalize this earliest will compound advantages that look, in hindsight, like the early internet, early crypto, and early social media adoption curves combined.