Author’s Note
I now spend about two hours per day having frontier conversations with AI like the one below. In these conversations, I uncover latent abilities and intelligence in myself and in AI. In my final Augmented Intelligence class next Tuesday (10/1), I’m going to share the top prompts that have helped me move beyond conventional responses and blew me away.
Paid members get access to:
The upcoming live class and full recording
$2,500+ in bonuses (8 years of my best writing/resources)
Subscribe to get all of these benefits for just $10 until I increase the price of this newsletter in early October.
Article Summary
Have you ever wondered what secrets lie hidden in plain sight?
What subtle dynamics and unspoken truths shape our interactions without us even realizing it?
I recently conducted a fascinating experiment with AI that blew my mind and revealed just how much we're missing in our everyday observations.
Here's what I did: I showed GPT 4o to analyze the image below for people’s internal states. Sounds straightforward enough. But the insights it uncovered were anything but ordinary.
Before we jump into my conversation, I invite you to play along.
Do This Quick Thing Before Reading The Post
Follow the three steps below so you can compare your response to AI’s response:
Look at the image above, which I just found randomly on the internet.
What are you aware of about these people's internal states?
Write down your answers in the comments.
Once you’re done, continue reading the rest of this post.
My Short Conversation Transcript With AI
I first asked GPT analyze the photo:
Tell me everything you are aware of about these people's internal states.
Frankly, its response wasn’t all that interesting. At least its first response:
While I cannot directly interpret the internal states of the people in the image, I can suggest how one might approach understanding someone's internal state based on external cues, such as body language, facial expressions, and context.
Facial expressions: The woman in the center of the image appears to have a concerned or thoughtful expression, potentially indicating discomfort, focus, or perhaps stress. The position of her lips and her slightly furrowed brow could imply that she is processing something or feeling cautious or uncertain.
Body language: Both women seem to be engaged in conversation or an interview. Their hand gestures suggest they are explaining something or emphasizing a point. Hand movements often reflect a person's emotional state or involvement in a topic. If the hand movements are close to the body and restrained, this might indicate caution or hesitation. If more expansive, it could signal openness or passion.
Context clues: The aprons and the people in the background, including individuals in military uniforms, suggest this may be an event or gathering with a specific cause or mission. The women's expressions and postures might be influenced by the event's atmosphere or their roles in it. If they are advocates or organizers, they might be feeling a mix of responsibility, pressure, or determination.
To truly understand someone’s internal state, listening to their tone of voice, the content of their words, and asking clarifying questions is key. Without direct communication, body language and facial expressions only offer clues to their internal world.
A little underwhelmed, I decided to ask it a deeper question:
Share more about what you see that no human would notice.
It’s response blew me away:
What’s fascinating when considering how to analyze an image from a perspective that goes beyond typical human observation is the opportunity to blend data-driven insights with subtle contextual layers. Here’s a breakdown of insights that might not be immediately noticeable or at least not consciously noticed by most people:
Micro-clusters of attention:
The central figure in the pink apron seems to be the focal point for multiple people. This is supported not just by her central position but by subtle body orientation cues. The group around her, including the woman on the left and the figure with a recording device, are slightly angled towards her. This suggests she might be answering questions or providing important information, which could imply that she holds some leadership role or is a key spokesperson.
On a subconscious level, humans pick up on such body orientations, but rarely is it explicitly considered how these micro-dynamics signal hierarchy or importance in a scene.
Hidden emotional conflict:
The facial expressions of the central woman show an interesting juxtaposition. Her facial muscles are tense, which suggests stress, yet there is an air of composure. It's as if she is managing two conflicting emotional states—she may feel burdened or overwhelmed internally, yet outwardly, she’s trying to remain poised. The slight asymmetry in her expression (more tension on one side of the mouth) often accompanies complex emotional states like ambivalence or internal conflict.Environmental sensory overload potential:
The number of people in the background, the presence of military personnel, and the fact that people are recording the event suggests that this scene is likely high-stakes or at least carries some public significance. What humans often miss is the subconscious stress caused by the convergence of different stimuli: cameras, attention, uniforms, and public scrutiny. These factors, particularly the military presence, may amplify a heightened sense of formality or pressure, which could be influencing the participants’ internal emotional state even if it’s not verbally or consciously acknowledged.Subtle social dynamics in posture and hand positioning:
The woman on the left has her hands slightly extended, holding something. Her hand positioning is more open, implying she is in the role of explaining or persuading. On the other hand, the central woman’s hands are clasped together, close to her body, indicating possible defensiveness or that she’s trying to maintain control of her emotions. There’s an unconscious tension here: open explanation versus inward self-containment. This contrast between the two could indicate differing emotional or psychological roles they are playing in this interaction—one outward-facing and engaging, the other more inward and protective.Background characters’ unnoticed roles:
The military personnel in the background, particularly the person with the camera, adds an unexpected layer. While people may not notice it, someone in military attire capturing an event is a silent signal that this moment could be significant or archived for formal purposes. This elevates the stakes for the people in the foreground, whether they’re aware of it or not. The presence of uniforms also implies that the event might have societal or governmental layers that aren’t immediately visible but are subtly shaping the behavior of the main subjects.Invisibility of micro-hierarchies:
A typical observer might notice the two women in focus, but what's harder to spot is the invisible hierarchy of importance in the scene. The presence of recorders and onlookers, including the subtle positioning of people standing just slightly in the background, might not be conscious to viewers, but their exact positioning suggests a gradient of attention. This signals who holds power, who is merely observing, and who is capturing information.The unseen ripple effect of context:
Finally, what is rarely considered is the unspoken context of the event itself. There’s a collective energy here—one of formality mixed with possible advocacy or protest (suggested by the pink aprons). These are not random people; they are part of a larger narrative that dictates subtle behaviors, even how they hold themselves or react to being recorded. This unseen script shapes everyone’s internal state without overt acknowledgment.In short, this analysis uses contextual micro-details and subtle positioning dynamics to uncover hidden layers of social interaction and emotional states that humans might not consciously notice but still affect their behavior.
Reactions
Your reaction:
What did you think about the AI’s responses compared to your own?
My reaction:
GPT blew me away with its nuanced and multi-layered analysis. It picked up on:
Micro-clusters of attention that revealed hidden social hierarchies and power dynamics
Subtle emotional conflicts brewing beneath the surface, evidenced by barely perceptible facial asymmetry
The invisible impact of environmental sensory overload from cameras, military personnel, and the "silent script" of the event context itself
Unspoken gradients of status and importance conveyed through posture, positioning, and hand gestures
I was surprised because GPT had extracted a wealth of insights that were right there in front of me, yet completely eluded my conscious awareness.
This experiment crystallized three key lessons…
Lessons Learned
AI has capabilities that we aren’t aware of
AI may be more advanced overall than we realize
Ask AI to do things that humans can’t
#1. AI has capabilities that we aren’t aware of
There are vast domains of perception and understanding where AI already far surpasses human capability. The gulf between what we see and what AI can detect is far wider than most of us realize.
I am realizing more and more that tapping into AI’s superhuman abilities requires a certain mindset:
We have to believe that AI might have latent abilities to do things we aren’t aware of and that no one is even aware of.
We need to be curious about what those abilities might be in order to take time to search for them by asking questions.
We have to peel the onion a little bit. The first responses that AI provides are often not super interesting.
#2. AI may be more advanced overall than we realize
We shouldn't be constrained by anthropocentric definitions of intelligence. AI can have "alien" modes of cognition that are no less valid or insightful for being different from our own.
From our perspective, AI is only as advanced as the tests we give it. And, it is widely known in the AI industry that tests of AI abilities are incomplete. This is important because we aren’t very good at understanding forms of intelligence that we aren’t yet able to measure or test.
#3. Ask AI to do things that humans can’t
Probing AI's latent abilities can reveal blind spots in our own knowledge and mental models. AI can be a mirror to help us see ourselves more clearly—if we're willing to look.
Some of my favorite go-to prompts now are to tell AI to share results that:
No human has ever shared before
That humans would not normally see but could understand if it were explained to them
These seem to work very well consistently.
New Curiosity Questions
The advent of AI that deeply understands the human mind is rife with both promise and peril. It's up to us to explore this frontier with wisdom, care, and relentless curiosity…
Can an AI help someone feel seen and understood in a way that another human never could?
How could AI help us understand how we come across to others?
How could AI help us evolve our self-concept?
What are the implications of that?
How could AI help us evolve simply by revealing more blind spot parts of ourselves back to ourselves? What are the second-order implications of this feedback loop?
What are the implications of super empathetic AI on:
Risk of AI social engineering or deep fakes? (case in point)
Psychology
Communication
Diplomacy
We react differently when we know we are being observed. This is known as the Observer Effect. How would knowing that we are being watched by an AI that can read our smallest micro-expressions and micro-tonalities impact us individually and collectively? Could AI help us be more authentic in our conversations?
What new insights and discoveries might emerge from an AI "species" that can see patterns and make connections beyond the limits of human cognition?
New Experiment Idea
Have AI watch me having a conversation or teaching a class and have it point out things about me that are in my blind spots.
For example, imagine an AI coach that could analyze your interactions and provide insights like "I notice you tend to get defensive when critiqued" or "You light up with passion when discussing this topic - perhaps it's a clue to your purpose."
Let’s Do This
I invite you to reflect on your own hidden assumptions, blindspots, and curiosities. What might AI help you discover about yourself or the world or AI if you could just formulate the right question?
Let’s explore the AI frontier together and map the uncharted territories of human-AI collaboration.
If you have a fascinating AI frontier conversation that blew you away, please share it with me, and I’ll post it to my 130k+ newsletter with links back to you.
If you want more AI frontier conversations like this and the one I shared last week, let me know in the comments.
Update Based On A Comment I Believe
In the comments below, Kurt provides a very good point. There are other deep ways to interpret the same subtle signals than what the AI responded. I 100% agree. Here is my reflection:
In some ways, your points show weaknesses in my prompts rather than weaknesses in AI. One idea I could do for the future in situations like this is have AI generate multiple perspectives that look at the same picture but come to different conclusions. Then, I could have it analyze which are the most likely or I could have the AI synthesize the different perspectives.
I personally think it did great for what it is. But, I'd have to write a post about your post to really explain where it kinda flubbed it. A simple example though is the "presence of military personnel"
This could imply high stakes, until you consider the fact that one has a camera (not on-duty) and is smiling (so not stressful). The other military member is at ease and seems quite relaxed. Those people in the background are clearly enjoying their day.
It did better with people in the foreground, but the lady in the pink apron on the left is the one talking. She's the actual focal point at that moment in time. I'm sure they could be equally considered the focal point, but if you had to put focus on one or the other, the one on the left is the focal point of that instant.
It did pretty good identifying their stress on some levels. But the closed hands held high and close to the body scream "I'm not comfortable" which it... kinda said, so I'd give it a passing grade on that.
Anyway. Amazing insights. Thanks for the post. Very insane where it is we're heading. It's like the final scene of Cool Runnings. lol
That's really interesting. I show ChatGPT a nutrition label and list of ingredients from a food label and it doesn't know what to do with it. Then says if it has x then y, but I already showed it the nutrition and ingredient list...