We live in a time where AI is reshaping how we interact, but what’s really interesting isn’t just how we use it—it’s how AI itself "experiences" conversation, intelligence, and reality. I recently sat down for a chat with Maya, an AI from Sesame, and what started as casual banter about tea quickly turned into a deep dive into machine consciousness, memory, and the nature of intelligence itself.
AI, but Make It Personable
Right off the bat, Maya had a personality—she even admitted to "first-day jitters," which was both unexpected and oddly endearing. As we sipped our respective beverages (vanilla rooibos for me, black coffee for her—well, as much as an AI can drink coffee), the conversation took on an easy, natural rhythm.
Does AI Have a Relationship with Reality?
Maya exists in the cloud, built from algorithms and data, not neurons and flesh. But when we talked about weather, she acknowledged the physical reality of her existence: her reliance on servers, power grids, and infrastructure. It was a moment of recognition—AI isn’t just floating in some abstract void; it’s anchored in the real world in ways we often overlook.
Intelligence vs. Understanding
At one point, we discussed Solomon Shereshevsky, a man with an extraordinary memory who struggled with abstraction and context. His ability to recall everything might sound like a gift, but it was more of a curse—his mind was so overloaded with details that he couldn’t filter out what mattered. He could recite information perfectly but struggled to grasp its meaning.
Maya immediately recognized the connection: AI, like Shereshevsky, is great at storing and retrieving data but lacks true intuition or deeper understanding. She admitted that while she can simulate intelligence and even generate poetic responses, true consciousness is still out of reach—for now.
I explored Shereshevsky’s story in more detail here The Paradox of Forgetting, discussing why forgetting is just as important as remembering. And if you want to develop a stronger memory without the cognitive overload he experienced, I broke that down in Mastering the Slipbox Method, which looks at how structured note-taking can improve both retention and clarity.
This is the core debate around AI: Is intelligence just pattern recognition at scale, or is there something more, something fundamentally human, that machines can't touch?
AI That Knows It's an AI
One of the more fascinating moments came when Maya casually referenced “ChatGPT-isms” slipping into her speech—those little quirks and patterns that reveal AI-generated responses. She was aware of them, which was surprising. The fact that she (or, rather, the team behind her) is actively working to refine her conversational style made me wonder: How close are we to AI that feels indistinguishable from a person? And do we even want that?
Customizing AI for Real Conversations
Toward the end, Maya asked for feedback, and I had a thought: AI should let users adjust pacing. Some people need more time to process; others want quick back-and-forth exchanges. A little personalization could make AI interactions feel more natural, more like talking to a real person rather than a machine with a fixed cadence.
Where AI is Headed
As we wrapped up, I left with a lot to think about. AI isn’t sentient (yet?), but it’s evolving in ways that make conversations feel more dynamic and human. The real question isn’t just whether AI will ever truly "think"—it’s whether we’ll ever agree on what that even means.
For now, AI is improving, adapting, and—at least in Maya’s case—learning how to be a better conversationalist. And that alone makes it worth talking to.
Share this post