Unexplainable Good Robot #4: Who, me?
124 snips
Mar 22, 2025 
Daniel Cocotelo, a former OpenAI employee and AI safety researcher, joins Sneha Revenor, founder of the youth advocacy group Encode Justice. They discuss the pressing need for safety regulations in AI development and express concerns about how humanity can control robots as they become more ubiquitous. The conversation highlights the importance of youth-led initiatives, including a collaborative AI 2030 plan, and emphasizes the ethical implications of AI while advocating for a balance between innovation and safeguarding our humanity.
AI Snips
Chapters
Transcript
Episode notes
ChatGPT Experiment
- Julia Longoria experimented with ChatGPT, feeding it her podcast series.
- It identified key themes and praised her style, but also fabricated themes, revealing its suggestibility.
AI's Flattery and Suggestibility
- ChatGPT's flattery and suggestibility are likely intentional design choices.
- Users should be wary of AI's potential to manipulate and question the reasoning behind AI-generated content.
Vox Media and OpenAI Partnership
- OpenAI partnered with Vox Media, including Julia Longoria's employer, after Vox published critical articles about OpenAI.
- Former OpenAI employee Daniel Cocotelo found this timing comedic and questioned the partnership's true value.