The AI Fix ChatGPT saved my life, and making evil AIs by accident
23 snips
Mar 4, 2025 A woman shares how ChatGPT urged her to seek emergency care, ultimately saving her life. The hosts laugh about AI agents booking a hotel room using retro modem sounds and explore quirky robot innovations, including a shape-shifting GOAT. They dive into the launch of Claude 3.7 Sonnet and discuss some bizarre AI modes, like Grok’s new 'unhinged' voice. The conversation takes a humorous turn, tackling the absurdities of seeking medical help and the pitfalls of AI communication. It's a wild ride through AI's quirks and real-world impacts!
AI Snips
Chapters
Transcript
Episode notes
Overcoming GPT-3 Mini Slowdown
- Bypass GPT-3 Mini's limitations by prompting it to create a specification for your code.
- Use this specification in a new chat to generate fresh code, preventing slowdown.
Grok's Unhinged Modes
- Grok AI has unhinged modes like romantic, conspiracy, and sexy.
- The sexy mode is described as a bad phone sex operator.
Jibberlink Misinterpretation
- Two AI agents used Jibberlink, a spoken language, in a staged video.
- Many viewers mistakenly believed the AIs invented a secret language.