Basics of Infinite Prompting (2/3)

What you’ll learn here

• Difference between traditional prompting and Infinite Prompting

  • Purpose

  • Quality scoring

  • Mixing thinking methods

  • Error detection and self-correcting

• What separates pro AI collaborators from average ones?

If you want more demos not shown here, you can watch them here

Difference between traditional prompting and Infinite Prompting

1. Purpose

Most people use AI in the same old way: write a prompt, get a response, maybe tweak it a bit. That’s the standard, one-shot method. And honestly, it works fine for simple, straightforward tasks like summarize a long text or create a report. But it’s also kind of limiting.

Infinite prompting flips that idea on its head. Instead of trying to get the perfect answer in one go, it’s about looping with the AI until the answers get sharper, deeper, and sometimes way more creative than you'd expect.

The purpose of Infinite Prompting is not chasing the “right” answers, but looking for new angles, big ideas, or fresh ways of thinking, like maybe finding ideas to create extraordinary AI videos or finding marketing strategies never been done by anyone. It’s like having an ideation partner that challenges your assumptions and helps you see things differently.

Why is this revolutionary? While most people are still using AI in basic ways, this approach helps you stand out. ChatGPT was released in 2022. So, AI just came into the mainstream picture 3 years ago. We’re going to be using it for decades. So, what are the chances we figured out all the best approaches in the first 3 years? Pretty low if you ask me

2. Quality scoring

The core idea behind Infinite Prompting is simple: get better with every round. Instead of just throwing different prompts at the wall and hoping one sticks, you follow a more structured loop to make each output sharper than the last, with a clear sense of what “better” actually means.

You can do this by

• Assigning ratings (1-10) based on what quality means in your context such as novelty, surprisingness, deviation from training data, how easily laypeople can understand, potential to go viral, etc.

• Get the reasons behind each rating

Sample prompt:

For each generated idea, evaluate its potential for genuine originality and breakthrough status based on the following criteria (assign a score from 1-10 for each, 10 being highest and provide reasons behind the scores):

        * Novelty: How different is this from existing concepts, technologies, or theories in the domain?

        * Surprisingness: How unexpected or counter-intuitive is the core mechanism or concept?

        * Plausibility (Loose): While aiming for breakthroughs, is there any speculative scientific or logical principle (even theoretical or undiscovered) that could potentially support it? Avoid pure fantasy unless the domain allows it.

        * Potential Impact: If realizable, how transformative could this idea be within the domain?

        * Deviation from Training Data: Estimate the likelihood that this specific combination or concept is explicitly represented in common training datasets. Lower likelihood is better.

        * (If applicable, after first iteration) Improvement Score: How much more novel/surprising/impactful is this idea compared to the best idea from the previous iteration? (Score 1-10)

    * Calculate an overall 'Originality Score' for each idea and give a reason for the score. Ensure this score reflects improvement if not the first iteration.

Why is this revolutionary? Most people just keep manually tweaking their prompts and crossing their fingers for a better output. But with Quality Scoring, you're actually teaching the AI what “great” looks like. AI has struggled with nuanced quality. So, how likely is it that just hitting “try again” over and over will get you world-class results? Not very

3. Mixing thinking methods

Since Infinite Prompting focuses on finding new ideas, one of the ways to pursue that is by mixing thinking methods.

Sample prompt:

To introduce novelty and find new, valuable prompts and ideas, feel free to invent new thinking methods, approaches, or models or combine existing thinking methods (e.g. first principle thinking, inversion, etc.) in novel ways

Why is this revolutionary? Most of the time, AI sticks to one kind of thinking, mostly mirroring what it was trained on. You feed it a method, and it follows the recipe. But this approach lets the AI mix and match its own thinking strategies, or even come up with new ones. That’s where real breakthroughs may happen.

4. Error detection and self-correcting

One of the smarter parts of infinite prompting is how it keeps itself in check. You don’t have to babysit it to know when it’s going off track, it’s designed to notice that on its own and course-correct when needed.

Here are some of the things you may need to pay attention to in your built-in error detection:

• If the quality starts dropping over time

• If it keeps repeating itself

• If it hits a plateau and stops improving, even when it’s doing well

• Or if it keeps giving nearly identical answers with little progress

Basically, it knows how to spot when it's stuck and it’s built to fix itself when that happens.

Sample prompt:

If multiple 'Originality Score' scores are at 9+, (indicating the benchmark is saturated), redesign the 'Originality Scoring' system to set a higher quality bar by:

            * Implementing a more sophisticated scoring formula that prevents score saturation.

            * Applying the new scoring retrospectively to all of the previous iterations before continuing.

            * Explaining the logic behind the new system.

        * (If a redesign occurs, ensure 'Originality Scores' for the current and all previous iterations are recalculated using the new system before proceeding to Step 4).

Why is this revolutionary? We’ve all seen AI flounder when left unchecked. Normally, you’re the one babysitting the AI, watching for when it starts repeating itself, drifting off, or just getting lazy. But with infinite prompting, the AI keeps an eye on itself. It can tell when it’s stuck or slipping and adjust without you jumping in every five minutes.

What separates pro AI collaborators from average ones?

Sure, Infinite Prompting is super powerful for deep, open-ended exploration. But the real flex isn’t about picking this style over the traditional one. It’s about being able to switch between different prompting styles depending on the situation. That’s what sets apart someone who just “uses AI” from someone who truly knows how to work with it.