Governance doesn’t have inherent moral valence; it’s a neutral framework for delivering against intent.
AI governance, however, does: LLM accuracy, as a primary measure of how AI fulfills human intent, makes accuracy a moral quality of governance.
To be right is to be good…
Using your company’s ChatGPT business plan is essentially a consensual forfeiture of your uncensored Google search history to your employer, qualified with a series of prompt exchanges that reveal how dumb you actually are to have needed to search what you searched.
I have a running text conversation with myself. There’s about a 2-second delay between sending the text to me and receiving it from me.
It gets me every time still — the delayed inbound text, rendering me a dopamine-fiending amnesiac, who wonders, “Who Might That Be!”
Months ago, I went to the bike helmet store and picked up a bright orange Smith helmet. “This is cool,” I thought in my head, which was immediately met with a “Orange? Bold choice.” by the bike helmet store worker. I bought black instead.
Traumatic brain injury/rizz trade-off.
Everyone on my feed talking Huberman. Must repent for whatever sins I have unwittingly committed to get myself this
adjacent to sincere utilitarian bro popsci culture.
I joined a “play music every day in March” Telegram group. I was expecting what you’d expect: learning some new chords on the guitar, amateur GarageBand mixing. So far we’ve got: shakuhachi flute, Uruguayan bossa nova, didgeridoo circular breathing, glossalia, casual humming…