Machines, Intelligence and Subjectivity
A language model can solve a problem and describe an inner life. What would establish that it actually experiences anything?
John Hardy’s engineering notebook
Current bench Atom · Z80 systems · AI and authorship
A language model can solve a problem and describe an inner life. What would establish that it actually experiences anything?
This article argues that current AI systems can model thought with extraordinary power without becoming conscious subjects. It explains the difference between simulation and experience, syntax and meaning, and output and inner life for a technical audience.
This article argues that code generated without sufficient human authorship may fall into the public domain by default. It frames that legal risk as a neglected issue inside the current rush toward AI-assisted software development.
This article examines the legal parallel between the monkey selfie case and software generated by large language models. It considers what happens to ownership, licensing, and open source when code lacks traditional human authorship.
On the January 2025 Copyright Office advice that vibe-coded software is public domain, why chat-window prompting fails to transfer authorial intent, and what a clay-like programming interface would need to do differently.
Explains why I build this in public and how writing reshapes the technical decisions. It also sets out how I use deliberate review to keep pace without losing my voice.