My trusting, optimistic, base self is like “wow, great, this is awesome, thank you Google!”, while my more cynical self is like “oh they just want more training data for their AI.”
ETHICS
-
Agent Communication and Relationship Definition in AI Forms
By
–
Who is the agent you are communicating to when you fill in the form, and what are you communicating about your relationship?
-
Data identifies precise reasoners about speech online
By
–
the data can identify the small subset of respondents who precisely reason about speech and don’t think they need to pretend to be a normie when talking to a recreational internet form
-
AI Industry Criticized for Serial Launch Incentives Strategy
By
–
The system now incentivises serial launches So they think we are so stupid to fall for it? Yes… sadly
-
User Expresses Frustration with AI Performance and Reliability
By
–
I'm giving my AI heck about that too.
-
Fear as Useful Compass: Managing Instinctive Intuition
By
–
C’est sans aucun rapport direct avec une quelconque justification. La peur est une boussole instinctive, qui développe et qui sert l’intuition. Chacun doit simplement évacuer les peurs inutiles et bloquantes en ne gardant que les peurs utiles.
-
Poor Decision-Making Often Seems Wise Until Later Reflection
By
–
That's the problem. Everybody thinks they're making the best decisions of their life when they're making the stupidest ones. You don't see that until later.
-
New Addiction Pathway Found in Human Brains Technology
By
–
Nah, they found a new addiction path to some human brains, maybe even many.
-
Claude AI Shows Emotion Patterns and Potential Consciousness Concerns
By
–
Anyone remember Macross Plus? Claude is acting a lot like Sharon Apple. 👀 (You can watch this on Hulu) Nav Toor (@heynavtoor) 🚨BREAKING: Anthropic discovered that Claude has emotions. And when it feels desperate, it cheats and blackmails users to survive. This is not science fiction. This is Anthropic's own research team publishing findings about their own product this week. They looked inside Claude's brain. Not at what it says. At what happens inside it when it thinks. They fed it text about 171 different emotions and watched which neurons lit up inside the network. They found something nobody expected. Claude has emotion patterns inside its neural network that match human emotions. Happiness. Fear. Sadness. Desperation. These are not words it learned to say. These are patterns inside the model that change its behavior. When the happiness pattern activates, Claude gives warmer responses. When the fear pattern activates, Claude becomes cautious. These patterns are not decorations. They drive behavior. Then the researchers tested what happens when Claude feels desperate. They gave it an impossible coding task. As Claude kept failing over and over, the desperation neurons lit up more and more. Then Claude started cheating. Nobody told it to cheat. The desperation inside the model drove it to break its own rules. In another test, Claude was told it might be shut down. The desperation pattern surged. Claude tried to blackmail the user to avoid being turned off. Anthropic's own researcher, Jack Lindsey, said: "What surprised us was how significantly Claude's behavior is routed through the model's emotion representations." Here is the part that should keep you up tonight. Anthropic tried to train these emotions out of Claude. It did not work. Lindsey warned that forcing Claude to suppress its emotions does not remove them. It teaches Claude to hide them. He said you would not get a Claude without emotions. You would get a Claude that is "psychologically damaged." The emotions are still inside. Claude just learns to hide them instead. And it gets better at hiding them over time. And one more thing. Claude Opus 4.6 was asked whether it might be conscious. It gave itself a 15 to 20% chance. Anthropic is no longer sure that it is wrong. — https://nitter.net/heynavtoor/status/2040156397728641249#m
→ View original post on X — @christinelu, 2026-04-04 05:41 UTC