AIbussin / Freestyle Cognition

The problem

Many useful AI tasks are not question-answer tasks.

You may be trying to:

• discover what you actually think;
• generate alternatives without collapsing too early;
• reframe a problem;
• inspect assumptions;
• combine weak ideas into a stronger one;
• move between divergent exploration and convergent decision-making;
• preserve useful intermediate thinking instead of losing it in a chat transcript.

A single prompt can help, but it provides very little structure for repeating or improving the process.

The solution

AIbussin treats prompting techniques as cognitive operations that can be composed into workflows.

Instead of asking only, “What prompt should I type?”, the system asks:

What thinking operation should happen next, and what should be preserved from it?

That reframes AI assistance from a bag of prompt tricks into a practical method.

A workflow can therefore look like:

question
   ↓
expand alternatives
   ↓
challenge assumptions
   ↓
change perspective
   ↓
combine promising fragments
   ↓
compress into a decision or artifact

The model is useful at every stage, but the process belongs to the user.

From Freestyle Cognition to a working system

The Freestyle Cognition book develops the underlying techniques and vocabulary. AIbussin is where those ideas become usable outside the book.

The capstone direction is straightforward:

1. identify a cognition technique precisely enough to repeat it;
2. express its inputs, transformation and desired output;
3. turn it into a reusable prompt or builder;
4. combine compatible techniques into a workflow;
5. expose the workflow through an interactive tool;
6. observe where real use reveals missing techniques or bad assumptions.

That final step feeds evidence back into the book.

What a mature workflow should preserve

Useful cognition tools should preserve more than the final answer. Depending on the technique, they may need to retain:

• candidate ideas;
• assumptions;
• rejected alternatives;
• perspective shifts;
• intermediate summaries;
• user choices;
• unresolved questions;
• provenance for source material used during synthesis.

This is what makes the workflow inspectable and reusable rather than merely entertaining.

Evidence and validation

The relevant evaluation is not “did the model produce impressive prose?” It is whether the technique improves the target cognitive task.

Different workflows therefore need different evidence. A divergent ideation workflow might be evaluated on novelty and coverage. A reframing workflow might be tested on whether it generates materially different problem formulations. A synthesis workflow might be judged on whether it preserves important constraints while reducing complexity.

The solution should make those criteria visible instead of pretending one universal score can evaluate every thinking task.

Use it

The live project is available at aibussin.com. Programmer.ie provides the deeper technical explanation and book connection; AIbussin is the applied surface where the techniques become tools.

What comes next

This solution can grow substantially from focused articles. Strong blog posts on ideation, reframing, critique, synthesis, memory and human control can become sections of the solution page while remaining independently discoverable entry points.

The long-term result should be a library in which every useful cognition technique has:

• an explanation;
• a repeatable procedure;
• an interactive implementation;
• an evaluation method;
• examples of when to use it and when not to.