Explaining AI stages with visuals from @itsemilyhiggins

Video format

Talking Head Edit

Video outline

1. State a common misconception
2. Build a progressive framework
3. Escalate stakes at each level
4. Pose an existential final question

Hook overview

The Knowledge Gap Hook: State that there are 'X' levels to a concept, but most people only know the first one, creating an immediate desire for the missing information.

Title hook

There are 5

Verbal hook

There are five stages to artificial intelligence, and most people only know of level one

Visual hook

A medium shot of the speaker holding up five fingers to visually represent the number of stages she is about to explain.

Hook strategies

1. secrets-shortcuts
2. curiosity-open-loops

Payoff

The Unresolved Question: After revealing the final, most powerful stage, the payoff is not an answer but a profound, binary question that forces the audience to consider the ultimate implications (e.g., utopian vs. dystopian), maximizing engagement and discussion.

Narrative framework

The Escalating Pyramid

Narrative framework logic

Deconstructing a complex topic into a simple, visual hierarchy that escalates from a commonly understood baseline to a speculative, high-stakes conclusion, culminating in an unresolved existential question.

Narrative framework breakdown

• frameworkName: The Escalating Pyramid
• frameworkType: List-Style Narrative
• coreLogic: Deconstructing a complex topic into a simple, visual hierarchy that escalates from a commonly understood baseline to a speculative, high-stakes conclusion, culminating in an unresolved existential question.
• confidence: 95
• pendingStatus: created

Transcript excerpt

There are five stages to artificial intelligence, and most people only know of level one, the large language model. This is Chachi, BT, Gemini, Claude, you ask, it answers. Most people think this is AI, but in reality, this is just the foundation. The next stage is where things get interesting. This is where you can give AI a job. Not answer this question, but go do this thing for me. This is agentic AI, and this is where we are right now. But now imagine that it's not one agent, but 20. One researches, one writes, another publishes. They delegate to each better, and there's no human involved at all. Infrastructure for this technology is being laid right now. Then there's the big one, AGI, or artificial general intelligence. This is essentially what everyone's racing to achieve. But the fascinating part is that experts can't even agree on whether or not we have already reached it. This is when AI can match or exceed human cognition. It can reason like a scientist, create like an artist, strategize like a better not because it was trained to do so, because it learned it on its own. And the final stage is Super Intelligence, the AI that surpasses human intelligence. It doesn't need o