Notebook
Quarterly Letter: The Character Layer
Why the Next Iconic Consumer Companies Will Compete on Philosophy
By Cameron McLain
few weeks ago, I watched my 6-year-old son ask an AI assistant for help with something that was upsetting him. The exchange was patient, thoughtful, and at times genuinely wise. It struck me that this app, and the ones that come after it, will shape how he thinks far more than any technology I grew up with.
Reid Hoffman, LinkedIn's co-founder, shared with us on Giant Ideas that he thinks AI is best understood not as a tool but as a new kind of "creature" - neither human nor merely technological, but a third category we are only beginning to make sense of. My son will grow up alongside these creatures, and they will influence his psychological development and his worldview more than any medium before them. As Marshall McLuhan put it, "the medium is the message", and the medium has now changed.
For most of computing's history, technology could only encode logic. Machines followed instructions. They processed information but did not interpret it. AI, which is probabilistic and intelligent, has changed that. Setting aside the AGI debate: today's systems already pursue goals and exercise judgment - and that judgment is being embedded everywhere: in our work decisions, our relationship dynamics, our children's lives. It will shape us, just as we shape it.
From Logic to Values
When I spoke recently with Robinhood co-founder Baiju Bhatt, he said one of the reasons he admired Steve Jobs so much was because you could see Jobs' personality in the software he built. Software has always reflected the priorities of its creators. But previous generations of software expressed those priorities at the layer of speed, design, efficiency, and engagement.
AI is different. It will reflect its creators, but it is sophisticated enough to offer a worldview and sustain a relationship. Conversation operates at the layer of human experience: memory, prioritisation, emotional texture, judgment under ambiguity. AI is GPUs in a data centre diligently repackaging the world's knowledge, but to the person on the other side, it is experienced as something to think, feel, and grow with. Whether or not anything like consciousness ever emerges from these systems, the experience of relationship is already real for the user, and that is what shapes them.
Lots of Intelligence, Not Enough Character
Most consumer products today optimise for productivity, engagement, and dopamine. They shape how people spend their days and, ultimately, how they experience their lives. AI will amplify this dramatically.
Some argue that AI should be "maximally truth-seeking." But why stop there? If an intelligence is going to be omnipresent in our lives, and in our children's, why not also encode it with wisdom, kindness, and good judgment? Truth-seeking is necessary, but it is not sufficient.
It is imperative that we are deliberate about the values and philosophies we train these systems on. For most of history, a society's values were debated, shaped, and transmitted by religion, education, literature, and institutions. Increasingly, AI's implicit values will do that work too. This does not mean machines are moral agents. It means they are imbued with the values they are trained on, and they propagate those values throughout the world every time they are used.
Fortunately, across cultures, there is a surprising consensus on what a good life looks like: meaningful relationships, purposeful work, vitality, and connection to something larger than oneself. Aristotle called it eudaimonia. In Japan, ikigai captures a related intuition, meaning found at the intersection of craft, contribution, and joy. For most of history, these ideas lived in philosophy and religion. Increasingly, they will be product decisions, beginning at the intelligence layer, in how we train the models.
The Next Iconic Companies
The most important competition in AI, then, may not be between companies or between countries. It may be between worldviews. What kind of intelligence do we want around our children? What conception of a human being gets reinforced a million times a day? And how will that shape the people on the other side of the screen?
These are no longer abstract questions. They are design questions. The founders who answer them well will build the next iconic consumer companies. That won’t be because their models are marginally smarter, but because they carry a coherent view of what humans truly value. Apple differentiated through design taste. Once the intelligence infrastructure layer is commoditised, the next generation will differentiate through philosophy. We have written before about the infrastructure boom financing this moment (The Bubble and the Engine) and where value accrues at the application layer (How Giant Will Invest Behind the AI Wave); the character layer sits above both, and it is where we believe enduring consumer franchises will be won.
You can already see early signs. Khan Academy's Khanmigo refuses to simply hand students the answer, guiding them to it through Socratic questioning, because its makers believe learning, not task completion, is the goal. Hume AI builds its voice models to optimise for the user's wellbeing rather than engagement, and spun out a nonprofit that explicitly prohibits manipulative design. Neither is winning on raw intelligence; both compete on a view of the human they serve.
How We Invest Behind This
At Giant, this is increasingly how we evaluate founders working at the human layer of AI. We are looking for a clear and considered view of the person on the other side of the product, and the conviction to design for them. The founders who take that responsibility seriously are the ones we want to back. Stay tuned, we're working on a few interesting ideas in this space.
