This was very noticeable in GPT-4 around its release and anecdotally it’s gotten better since. My complete armchair guess is this is RL rather than imitation — it feels like a student playing to their teacher’s quirky grading system.
AGI
-
AGI Existential Risks: A Comprehensive Lethality Assessment
By
–
This isn't some old, small, obscure goal BTW. https://
lesswrong.com/posts/uMQ3cqWD
PHhjtiesc/agi-ruin-a-list-of-lethalities
… -
Savior Complex: Humanity’s Greatest Threat in AI Era
By
–
People who believe they have the sole ability and responsibility to save humanity are humanity's greatest threat.
-
OpenAI Co-founder Takes Sabbatical While Pursuing Safe AGI
By
–
I’m taking a sabbatical through end of year. First time to relax since co-founding OpenAI 9 years ago. The mission is far from complete; we still have a safe AGI to build.
-
Unintended Consequences of AI Actions and Decisions
By
–
oh look, it's the unintended consequences of my actions
-
AI Services Affordability Crisis Looms Weekly
By
–
lotta time between now and friday… we might not even be able to afford the weekly dose at that point.
-
Next Era Not Smarter: Critical Perspective on AI Progress
By
–
Narrator: the next era was not, in fact, smarter.
-
AGI Girlfriend Era Begins: Revolutionary AI Companion Technology
By
–
The dawn of the AGI Girlfriend era is upon us, and it’s going to be legendary. #AGIGirlfriend
-
AI Development Criticized for Wrong-Headed Direction Despite Achievements
By
–
I could watch this astonishing video a thousand times, lamenting each time at how far AI has gone off course — towards an amazing but wrong-headed destination.
