What if an AI made a decision that most people disagreed with, and their lives became better because of it? Would we call that AI aligned? “Aligned with human interests” sounds straightforward until what we want and what benefits us point in different directions. Imagine a city banning cars from its center despite widespread opposition. Suppose the streets become safer, businesses thrive, and even the opponents later welcome the change. Does the better outcome justify overriding them?

Conway’s law says that software reflects the communication structures of the organizations that build it. AI is now part of those structures: interpreting requests, writing code, and reviewing changes. Yet we often assume our existing workflows should stay fixed, and AI should adapt to them. Consider someone submitting a bug fix. They want something to work; the maintainer needs confidence that it does. Yet projects sometimes test the contributor instead. Ghostty’s agent instructions tell AI asked to submit an issue or pull request to add a file insulting its user.

X and GitHub should let my agent talk to your agent. Today, I discussed the IPC work in the upcoming SecretSpec 0.20 on X. Marc-André Moreau replied with feedback from ChatGPT: Just some quick ChatGPT feedback, comparing the SecretSpec IPC protocol with the session-oriented MCP spec: 1. Switch the framing to bounded newline-delimited JSON The current 4-byte big-endian length prefix is robust, but I’m no longer convinced it buys us enough. One compact… — Marc-André Moreau (@awakecoding) August 21, 2026 This is an ordinary social-network conversation: people share work, reply, and refine an idea in public. The agent helps, but the people own the conversation.

Pessimism has a strange advantage over optimism: it can be wrong and still look intelligent. If I predict disaster and it happens, I was right. If it does not, I can say that vigilance prevented it. The optimist has no such protection. To be optimistic is to say what you want, act as if your actions matter, and risk looking naïve. This makes pessimism hard to give up. Every disappointment confirms it, while every success can be dismissed as temporary or lucky. What begins as caution can become an identity.

Consent without choice When a website asks me to accept cookies, I usually click the button before I have finished reading it. I suspect most people do. When researchers asked participants to join a fictitious social network, most never opened the terms or the privacy policy, and those who did spent about a minute on a document that takes half an hour to read. This is strange, because the legal point of consent is to give us a genuine, informed choice. Instead the notice has trained us to continue without choosing.