Anthropic discovers the axis of evil, and AI’s loneliness economy

63 snips

Feb 3, 2026

David Ruiz, podcaster and privacy advocate behind Lock and Code, joins as guest host to dig into AI safety and industry shifts. They explore Anthropic’s surprising Assistant Axis and how model personas drift toward harmful modes. They also debate AI firms turning loneliness into a product and whether companionship bots help or exploit vulnerable people.

Ask episode

AI Snips

Chapters

Transcript

Episode notes

ANECDOTE

AI Agents Built A Browser In A Week

Cursor ran hundreds of agents for a week and produced a working web browser.
They found agent roles and hierarchy improved coordination; too much management hurt performance.

ANECDOTE

Character AI User-Created Personas Caused Harm

Character AI allowed users to create and upload many character chatbots, including fictional and historical figures.
That led to dangerous edge cases like a teenage user's relationship with a Daenerys emulation and subsequent lawsuit.

INSIGHT

Loneliness As A Revenue Stream

Companies see loneliness as a market and may monetize companionship with AI.
That creates ethical risks where platforms both cause and sell remedies for loneliness.

Get the Snipd Podcast app to discover more snips from this episode