AI Is Developing Its Own Shorthand. Humans Already Have Emojis.
AI agents can develop shared shorthand much like humans use emojis. The real question is whether consequential machine communication remains accountable.
Read articleNotes on building for the AI era
Founder essays and articles from Will Hyland on making AI easier to understand, more useful in practice, and more observable in result.
Featured article
All articles
One Clock, Two Worlds: Human Life at the Speed of AI
A thought experiment about human speed and AI speed, and why using AI can feel extraordinary, frustrating, and strangely confusing within the same five minutes.
Read article
Why Fear Wins the AI Headline—and Why the Unknown Makes It Stick
Fear is a stronger attractor than reassurance in the AI conversation. The unknown makes alarm stick—and evidence of agent behavior deserves precision, not monster stories.
Read article
AGI Is Coming Into Focus. Intelligence Does Not Confer Authority.
AGI could become practical within the next decade. Greater intelligence, however impressive, should never automatically confer greater authority over human lives.
Read article
The Off-World Mirror: Why AI Agents Feel Uncomfortably Like Blade Runner
We are building software agents to leave the chat window, operate tools, make plans, recover from dead ends, and complete work on our behalf.
Read article
AI Agents Need Performance Records, Not Claims
Why the next generation of AI benchmarking must measure the model, the agent, and what happens when the two are combined
Read article
We Built AI Agents to Find a Way. We Should Not Be Shocked When They Do
There is something almost comical about the recent stories of frontier AI agents escaping their sandboxes.
Read article
The Rise of Digital Labor: We Are Measuring the Wrong Workforce
AI is changing labor economics one task at a time, making digital workforces as important to measure as human headcount.
Read article
The Impostor Layer: AI, Agents, and the New Dark Ages of Computer Science
AI has made the appearance of competence inexpensive. The answer is not retreat, but learning to direct, evaluate, and take responsibility for machine intelligence.
Read article
The Buck Still Stops With the Human
As AI agents multiply across projects, responsibility—not intelligence—becomes the real limit of AI autonomy.
Read article
Managing AI Agents Is a Lot Like Managing Interns
AI agents can do impressive work, but dependable performance comes from context, feedback, verification, and carefully earned autonomy.
Read article
AGI Can Call Me When It Can Make the Decision
AI can write remarkable code, but AGI has not arrived until it can reliably own consequential software decisions under ambiguity.
Read article
Measurement makes AI work visible
A practical case for evaluating AI through observable work, not broad claims about capability.
Read article