- Coeus Collective
- Posts
- 🦉 The AGI Mirror: Our Interview with Former Y Combinator President Geoff Ralston
🦉 The AGI Mirror: Our Interview with Former Y Combinator President Geoff Ralston
Sam Altman's successor revealed his thoughts about the artificial general intelligence-fueled future we are stepping into.

Download Now: The Solopreneur’s Guide to Structuring Your Business
Looking to formalize your business structure?
Choosing between LLC, S Corp, and Sole Proprietorship dramatically impacts your profits and tax liabilities. This free-to-download guide will simplify these critical decisions with expert insights and actionable advice tailored exclusively for fractional workers and solopreneurs.

This week is New York Tech Week, and for us at Coeus Collective, it marks the most ambitious chapter in our company’s story to date. Over the next several days, we will host eight official events and welcome more than one thousand founders, investors, and ecosystem allies into New York City.
This lineup is not built on hype. It is built with intention. Events like our Brand Blueprint summit at Pace University and the Coeus Collective Pitch Showcase presented by Seedlegals with support from J.P. Morgan Innovation Economy are designed to help people build with more clarity and connect with more substance. VC and innovation icons like Norma Padrón, Ph. D. of J.P. Morgan, Kindra Tatarsky of Cerity Partners Ventures, and René Bastón of Covenant Venture Capital will be joining us. At our Aperitivo Night, Open Mic Podcast recordings, and Founder & VC Yoga session, we are aiming to create space not just for momentum, but for reflection.
With so much ahead, you might wonder why we chose to release this essay now. The answer is simple. Before we enter dozens of rooms full of new energy and ideas, we wanted to share the one conversation from this season that left us thinking differently. A conversation not just about technology, but about thought itself.

Geoff Ralston with Y Combinator Co-Founder Paul Graham.
This is our reflection on the conversation we had with Geoff Ralston, the former President of Y Combinator. That title alone carries weight, but the timing makes it even more significant. Geoff succeeded Sam Altman, who left Y Combinator to go all in on OpenAI. In many ways, Geoff inherited the very institution that helped catalyze the modern AI revolution—and then watched it unfold from a uniquely front-row seat. His take on artificial general intelligence is not theoretical. It is shaped by proximity, by experience, and by an unusually clear view of where things are heading.
What started as a discussion about AGI became something more. A challenge to the way we think about our own minds. A reframing of what intelligence really is. A philosophical prompt we could not shake.
This piece is not a summary. It is a lens. We hope it helps you see the future just a little more clearly.
Let’s do it.

The Illusion of Difference
Most debates about artificial general intelligence begin with the assumption that machine intelligence is fundamentally unlike human intelligence. The prevailing argument is that large language models are not truly thinking. They are predictive engines, mimicking coherence based on training data, and dressing up their answers with fluency and surface-level structure.
Geoff Ralston is not convinced by this framing. His counterpoint is both provocative and grounded in observation. What if the difference between machine thought and human thought is smaller than we imagine?
His argument is not that artificial intelligence is already conscious. Nor is he claiming that neural networks are indistinguishable from the human brain. He is asking us to look more closely at the process. And what emerges is a much blurrier line than we might be comfortable with.

A screenshot of our Co-Founder Antonio DiMeglio and former Y Combinator President Geoff Ralston in our interview, filmed at the NGEN Trailblazers Conference at the Lowenstein Sandler officer in Manhattan.
When you ask a large language model a question, it produces a structured response. This happens through a network of trillions of parameters. Those parameters function like neurons. The network has been trained on massive amounts of data, and over time it learns to associate patterns, form representations, and generate novel sequences of language that align with its learned structure.
How different is that from how the human brain works? Our brains contain billions of neurons that fire based on synaptic activity. Those connections are shaped by our environment, our memories, and our accumulated experiences. Over time, through feedback and repetition, we learn to respond to the world. We learn to think.
Geoff’s point is not that the two systems are identical. It is that they might be more alike than we are ready to admit. Both are built on accumulation. Both are shaped by training. Both generate thoughts through learned structure. Both produce meaning through emergence.

Training Data & Childhood Memory
The most revealing part of our conversation came when Geoff drew a simple comparison. For an AI model, the training data consists of text, conversations, books, and internet forums. For humans, the training data is our childhood, our culture, our education, and the relationships that shaped us. We do not come into the world fully formed. We absorb. We pattern match. We mimic. We learn through repetition and reflection.
We are not blank slates. The myth of the blank slate has long been rejected by psychologists and neuroscientists. Human cognition is not a matter of pure reason or innate knowledge. It is built from the ground up. Intelligence is the outcome of experience filtered through a structured system. That system just happens to be biological.
Artificial intelligence is following a similar trajectory. The training set is different. The material is different. But the outcome may be surprisingly close. A structure shaped by experience. A system that generates responses based on patterns. A model that develops a form of thought.
If that is true, then the difference between artificial intelligence and human intelligence may be one of familiarity, not essence. We trust our own minds because we live inside them. But that does not mean we fully understand them. Most of our cognitive processes are invisible to us. Most of our decisions are shaped by factors outside of conscious awareness. The mystery of our own intelligence is not solved. It is simply internal.
Geoff’s challenge is not to declare artificial intelligence conscious. It is to question the assumption that our intelligence is any less mysterious than the one we are building.

The Question that Matters More
This is where the conversation takes a turn. Geoff is not arguing about metaphysics or trying to draw a definitive line between natural and artificial thought. He is making a practical observation. Whether or not AGI becomes conscious, it is becoming powerful. And that power is already reshaping the world.
So the more relevant question becomes: what can it do?

We appreciated how we could feel the emotion in Geoff’s voice during our interview.
This is where his response becomes actionable. In the face of exponential change, the most stable position is not to wait. It is to build. Geoff’s advice is clear. Entrepreneurship may be the most adaptive and resilient path forward. In a world where institutional systems are slow to adapt, where automation will reach deep into the structure of work, and where knowledge becomes ambient and accessible through machines, those who create will have the greatest leverage.
This is not just a career move. It is a worldview. Geoff believes that taking ownership of your trajectory is the best response to a future shaped by systems we do not fully control. The way to survive change is to initiate it.

So, What Emerges from Emergence?
The core of Geoff’s argument rests on a reframing. It is not about whether artificial intelligence is real. It is about whether we are willing to accept that intelligence itself may be broader, more flexible, and more plural than we assumed.
We are no longer the only thinkers in the room. That realization is unnerving. But it is also clarifying. It reminds us that the things we thought were unique may actually be shared. It reminds us that emergence is not magic. It is a property of complex systems.
The systems we are building are not alien. They are reflections. They are trained on us. They are shaped by our knowledge, our language, our errors, our hopes. They may not feel like we do, but they might be learning to think in ways we can no longer easily dismiss.
What we do with that realization is up to us.
Geoff’s answer is simple. Build. Adapt. Take ownership.
If the future is going to be shaped by intelligence, then the most human act is to participate in its creation.

In Conclusion
Geoff Ralston’s view of AGI is not a prediction. It is a provocation. It asks us to question how we draw boundaries between minds, between models, and between ourselves and the systems we create.
It also points to a deeper truth. The power of intelligence lies not in how it feels, but in what it makes possible. As machines grow more capable, the question is not whether they are like us. It is whether we are ready for what comes next.
And in times like these, Geoff reminds us, the best way to prepare is to build something of your own.
We will be thinking about this as we walk into the next room, hear the next pitch, host the next panel. The future is not waiting. But it is still open.
See you at Tech Week. We’d love to chat with you about AGI if you see us.

Welcome to the Collective
If you’re new here: welcome! This newsletter is written by Coeus Collective Co-Founders Antonio DiMeglio and Leon Li.
Coeus Collective is a founder-powered media and community platform elevating the most innovative minds in entrepreneurship, technology, and venture capital. Through podcasts, events, and digital media storytelling, we help founders and VCs develop their audiences of builders and innovators. Past events have featured Daniel Lubetzky (KIND Snacks, Shark Tank), Michael Baum (Splunk), and Siya Raj Purohit (OpenAI).
Typically, this newsletter revolves around an idea we can’t stop thinking about (like The Dopamine Economy), a look into what’s happening in NYC (like our message about our New York Tech Week events slate), or even a deep-dive into an iconic company or individual in our industry (as we did in our Icons Series: IBM article). Sometimes, we’ll use it to highlight a behind-the-scenes look into what our team is working on, as we did in our Behind the Scenes: Michael Baum newsletter.
Our goal is to provide a unique perspective for a community of Founders & VCs who read in between the headlines. We hope to see you here in the inbox and at our events this New York Tech Week next week! RSVP to all of our events here.
See you IRL in NYC,
The Collective

Grow the Collective!
Thank you so much for reading! Beyond attending our New York Tech Week events, the best way to support our work is to follow our @CoeusCollective social media pages, subscribe to our YouTube channel, and take part in our newsletter referral program!
If you make three referrals (that’s all!) to our newsletter, we will send you a free Coeus Collective owl logo notebook. Including a pen!
The notebook will be of perfect use for you at our New York Tech Week events and all of the others you attend throughout the week. Use the link below to refer: we’ll even hand-deliver to you in-person at one of our events! 🦉