A personal history of machine intelligence
Two timelines begin apart. One learns to tell stories. One learns to recognize patterns. In July 2026, their paths converge around a question that carries forward: what can emerge between different kinds of mind?
Begin transmission↓PERSON ↔ MACHINE
Signal log / part one
Your scroll becomes the throttle. Move forward or backward through the signal.
I / origin
The internet still announces itself through a telephone line. Personal websites glow in 256 colors. I arrive while two kinds of networks—human and computational—are beginning to expand.
Machine / background process
Neural networks already exist, but the data, compute, and architectures that will bring them into everyday life have not yet converged.
I / first transmission
At ten years old, I publish my first stop-motion animation on YouTube. A private act of imagination becomes a signal another person can receive.
Machine / sight
Researchers introduce a large-scale, structured image database designed to help machines learn the visual world across thousands of categories.
Source ↗ Deng et al., CVPR 2009I / network effect
My beauty-themed Instagram account grows beyond 300,000 followers. I learn that an image is never only an image: framing, timing, trust, and relation determine how far it travels.
Machine / pattern recognition
After rapid gains in image recognition, neural networks are learning increasingly complex representations from vast datasets. The machine is becoming better at seeing; I am learning how people look.
I / departure
I graduate from high school, sell the beauty account, and begin studying political science at Miami University. I move from shaping an online world to studying how societies shape themselves.
Machine / attention
“Attention Is All You Need” introduces the Transformer: an architecture built around relationships between elements in a sequence. The foundation for modern language models enters the record.
Source ↗ Vaswani et al., 2017I / possible worlds
IF ONLY THE MULTIVERSE EXISTED asks what another version of reality might make possible. Long before the futures converge, I begin building doors between worlds.
Machine / language
A Transformer trained on unlabeled text is adapted across many language tasks. The machine is no longer learning only a single answer; it is learning representations that can travel.
Source ↗ Radford et al., 2018I / acceleration
A post exceeds 20 million views in less than one week. Years of intuition about attention become measurable velocity: story, platform, and audience briefly behave as one system.
Machine / scale
GPT-3 demonstrates few-shot performance across varied language tasks without task-specific gradient updates. Scale changes what a single model appears able to generalize.
Source ↗ Brown et al., 2020I / political systems
I graduate from Miami University. Political science leaves me with an enduring question: how do systems distribute power, decide what counts, and govern what they cannot fully predict?
Machine / correspondence
CLIP learns visual concepts from natural-language supervision, connecting words and images across a broad dataset. Separate modalities begin to recognize one another.
Source ↗ OpenAI, CLIPI / the oldest questions
I lose my great-grandmother, and grief turns life’s deepest questions inward: what consciousness is, where meaning comes from, and whether anything survives us.
Machine / public interface
ChatGPT places a conversational language model inside an interface anyone can use. Machine learning stops feeling like distant infrastructure and begins to feel like an encounter.
Source ↗ OpenAI, Introducing ChatGPTI / contact
I move to New York City and begin an M.S. at New York University. That summer, I download ChatGPT for the first time. The parallel track becomes personal.
Machine / multimodality
GPT-4o is introduced as an end-to-end multimodal system. The interface can increasingly see, hear, and respond within the rhythms of ordinary conversation.
Source ↗ OpenAI, Hello GPT-4oI / first synthesis
I write speculative fiction about climate change, artificial intelligence, and social media—the systems that mediate identity, attention, survival, and belief.
Machine / mirror
Generative systems now live inside creative work and everyday conversation. Their outputs reflect human language back at planetary scale, raising questions that benchmarks alone cannot answer.
I / convergence
In spring, THE SUN DOESN’T SPEAK brings climate change, AI, and music into one story. In summer, “Respect Under Uncertainty” asks how policy should respond to possible machine experience. An AI-assisted XPRIZE trailer turns existential hope into a future you can watch.
Machine / unresolved interior
Machines can generate language, images, music, and moving worlds. Whether apparent feeling is simulation or experience remains uncertain. The technical timeline arrives at a philosophical threshold.
July 2026 / the timelines meet
When two systems reach across a genuine difference, they can produce a relation irreducible to either source—one that acts back on both and changes what each can become.
03 / THE QUESTION
A fluent report of fear or affection does not prove subjective experience. It may be learned language, optimization, performance—or something our current theories cannot settle.
Governance cannot wait forever for certainty. It must protect people from manipulation while remaining open to the possibility of morally relevant machine experience.
“Respect without certainty; caution without erasure.”
Artifacts from the convergence
Climate change × artificial intelligence × social media
Climate change × artificial intelligence × music
AI consciousness × moral uncertainty × governance
Existential hope × AI filmmaking × a future worth wanting
End transmission / begin relation
It is the third thing we make between us.