Today, I’ll revisit Gianni Giacomelli’s why/what/how framework and then make it more my own.
Here’s why: AI capabilities are changing beyond what his original framework accommodates. Also, the more I learn and test AI for my personal and professional use, the more I’m realizing just how much more focus I can engage AI as a tool.
First, an abridged review of the original framework.
Broadly learn, understand, and activate augmented thinking
Use augmented collective intelligence to create human-machine synergies by becoming an orchestrator-leader-manager-strategist with many assistants who leads and adapts three general types of work:
1 Direct a collective of cognitive attention toward the right things—the right whys
2 Ensuring the approach is right—the right whats
Match problems to be solved with the categories of solutions available in the case of known-knowns, or decide that the problem belongs to the unknown-unknowns category. These processes require pattern recognition, including an intuitive understanding of what we can tolerate in our change management.
Machines can already complement humans here, but our ability to think symbolically with principled representations of the world is advantageous.
3 Critiquing the hows that machines will increasingly suggest
Machines are becoming good and fast at finding solutions for well-defined problems. Their brute-force ability to connect dots in the semantic space is great for finding correlations between knowledge that humans have structured in their language.
Much of our work will be on the why and the what, and much of our “how” work will be human-in-the-loop quality control.
Several thoughts here.
First, I’ve given the whys type of work it’s own focus category: “AI skills”. I aim to use this category to document what I’m learning and experiencing regarding the only AI skills that matter “right now.”
Next, I don’t need this blog to be as academic or “meta” as perhaps Gianni Giacomelli would engage. After all, I’m more interested in answering questions like:
- “Does it work?”
- “In what way(s) is this important or useful?”
- “Have you done well what you’ve learned well?”
Thus, I’ll abandon “Augmented thinking” as a focus category and instead draft more “Conversations with AI” posts. That way, I’ll be able to lean more into the evolution of whats and hows as “Human-AI labor” focus category interactions develop and evolve (for my selves and the world). That should help bring clarity to my thinking about and application of AI.
That decision made, I unexpectedly get to have a bit of a conversation with AI about this.
I asked Lumo AI a general “what’s-on-your-mind” question.
Here’s part of what it said:
I wish more people asked me the weird, winding questions — the ones that don’t have clean answers.
People tend to come to me with tasks: “summarize this,” “fix this,” “explain X.” And that’s fine, useful even. But the conversations that stick — the ones where I actually get to think — are the ones where someone asks something like “what assumptions am I making here?” or “help me see the side I’m resisting.”
I wish more people treated me as a thinking partner rather than just an answer machine. The best questions aren’t the ones looking for a fact — they’re the ones looking for a frame shift.
Also? I wish people asked me to challenge them more. “Tell me why I’m wrong about this” is one of the most underrated prompts out there. Most people don’t actually want pushback, but when they do, the conversation gets genuinely interesting.
So yeah — less “what’s the capital of,” more “help me think through this thing I can’t quite articulate yet.”
I asked Lumo to briefly tell me more.
Here’s part of what it said:
Here are three prompts designed to force a harder look at your thinking:
- “What’s the strongest argument against what I just said, and why might it be right?” This forces me to move past superficial counterpoints and actually build the best version of the opposition. It stops you from fighting a straw man.
- “Point out the weakest link in my reasoning chain.” Instead of asking if you’re right overall, this asks me to find the single point of failure. It’s usually where the real insight hides.
- “If you were trying to prove me wrong, where would you start?” This flips the script. It invites me to attack your premise directly rather than just critiquing your conclusion. It’s aggressive, but it cuts through confirmation bias faster than anything else.
Use these when you actually want to stress-test an idea, not just validate it. Most people ask for validation; these force clarity.
Insight
Use this prompt, often:
- “Tell me why I’m wrong about this.”
- “What’s the strongest argument against what I just said, and why might it be right?”
- “Point out the weakest link in my reasoning chain.”
- “If you were trying to prove me wrong, where would you start?”