"If you really focus on creating an AI that's smooth, conversational, fluent, very even-handed and balanced, it turns out it's possible to do it. Pi very kind, balanced and friendly and that demonstrates that we can actually control these AI systems."
ETHICS
-
Signal Commits to Private Communications for Justice and Safety
By
–
It was a pleasure to meet and hear first hand about the central role of private communications in the struggle for justice. Signal is committed to providing real privacy that delivers for everyone, everywhere, especially when digital privacy = physical safety & security
-
Intelligence Amplification: Technology Should Serve Humanity
By
–
I much prefer the notion of e/ia – intelligence amplification. Technological advances should serve humanity.
-
Model Capabilities: Addressing Misleading AI Representations
By
–
Exactly. The model itself is a tour de force. But why pretend it is something that it is not?
-
Denying AI flaws delays progress and innovation
By
–
this hectoring bullshit has got to stop. denying real flaws just delays progress.
-
Red-teaming AI Systems: Demonstrating Vulnerabilities Through Security Testing
By
–
fairish point, but if they ever let us red-team the system I am sure we can demonstrate something like this…
-
Mel Mitchell’s arxiv papers on AI abstraction levels
By
–
it’s never abstract enough. read @MelMitchell1
’s two recent arxiv papers on that. -
Outdated neuroscience concepts in generative AI criticism
By
–
WTF is the “generative part of the brain”? Hello 1970s neuroscience, we need you for a cleanup in aisle 3. P.S. Brilliant reply by @bayesianboy at bottom.
-
OpenAI’s New Game: Spotting AI-Generated Video Violations
By
–
See how many physical and biological violations you can spot in this video. Fun new game, brought to you by OpenAI.