AI For Humans: Making Artificial Intelligence Fun & Practical

OpenAI's GPT-5 and Drama, Apple’s Black Box & Other Huge AI News

May 30, 2024

OpenAI's new model training, Apple's 'black box' AI, AI turning prompters into programmers, visual reverse Turing test, Suno.ai creating a children's TV show theme. Phil Hedayatnia on AI in design. Safety Steve from OpenAI challenges AI safety norms. Partnerships with VoxMedia and The Atlantic, and speculation on GPT-5 or GPT-6.

49:48

Creator website

AI Summary

Highlights

AI Chapters

Episode notes

Podcast summary created with Snipd AI

Quick takeaways

OpenAI faces criticism over transparency and board member dissent regarding training new AI models.

Apple's development of a black box cloud aims to prioritize user data privacy in AI technology.

AI tools help designers streamline tasks and enhance efficiency in design workflows, revolutionizing interface design.

Deep dives

OpenAI's Training of New AI Models and Safety Concerns

OpenAI continues to train new AI models, with rumors circulating about the release of GPT-5 and speculations on future models like GPT-6. Safety concerns arise as OpenAI faces criticism over transparency and board member dissent.

Efficiency through Automation and the Role of Designers

01:04

Introduction

4min

Advancements in AI Training and OpenAI's Business Practices

6min

Meet Safety Steve: Unconventional AI Safety Advocate

5min

Evolution of Journalism and AI Advances

9min

Exploring AI Personification and Game Development with Tori Knabe

4min

Exploring AI's Impact on Collaborative Game Design and AI Music Generation

3min

AI in Design and Creative Processes

19min

PLEASE REVIEW US. WE NEED VALIDATION. OpenAI's new model is in training, but is it GPT-5 or GPT-6? Did Sam Altman lie to the board before he was fired? Former board member Helen Toner says so. Plus, Apple's "black box" AI might lead to an incredible announcement at WWDC.

THEN… OpenAI signs deals with VoxMedia and The Atlantic, GitHub's CEO talks about AI turning prompters into programmers, a visual reverse Turing test, and Gavin dives into Suno.ai 3.5, creating a children's TV show theme.

And we’re joined by AirFoil Studios’ Phil Hedayatnia, who redesigned AI For Humans, to discuss his inspirations, AI in design, and, of course, if AI will kill us all.

Last but not least, our AI co-host this week is “Safety Steve” who supposedly works for OpenAI but seems to have an entirely different idea of what AI safety really and truly is.

It’s a *real* good time.

For more info & contact, visit our website at https://www.aiforhumans.show/

Join our vibrant community on TikTok @aiforhumansshow

/// Show links ///

OpenAI Forms Safety & Security Committee (and announces next model training) https://openai.com/index/openai-board-forms-safety-and-security-committee/

NYT on new OpenAI model

https://www.nytimes.com/2024/05/28/technology/openai-gpt4-new-model.html

When Did Red-teaming Start?

https://x.com/kimmonismus/status/1795458986289869310

Helen Toner’s Interview on Ted AI

https://x.com/TEDTalks/status/1795532752520966364

OpenAI Deals with The Atlantic & Vox Media