AI Meets Productivity

This week Trevor Lohrbeer talks with ChatGPT through the Hume AI Empathic Voice Interface (EVI) to discuss the advancements introduced with OpenAI's new GPT-4o model. They discuss including the model's impressive responsiveness, multimodal capabilities, and its potential to revolutionize AI interactions by understanding text, audio, image, and video in real time. 

They also explore how GPT-4o might be a stepping stone to training GPT-5 with native audio and video and the possibility of incorporating additional sensory data from robots. they emphasize how these innovations could transform human-AI interactions, making them more natural and intuitive on a level we haven’t seen before, while introducing new potential risks.

Can an AI actually listen to music, as opposed to merely hearing sound waves and processing them? In this episode, listen to hear how Hume AI, the first empathic AI, does attempting to listen to music, and how ChatGPT analyzes its performance afterwards.

To test whether Hume can hear the emotion sung within a song, distinct from the lyrics, I had ChatGPT write emotionless lyrics, then had Suno generate songs using these lyrics with different emotions. Listen to the full songs of each emotion used in this episode below.

Observations 1 (sad)

Observations 1 (happy)

Observations 1 (folk song expressing boredom)

What is the future of AI music now that we have tools like Suno and Udio that can create songs complete with vocals?

In this episode, ChatGPT and I explore how AI music will change how we create and use music in the same way that smartphones changed how we create and use photos. It’s not what you think.

This episode features 3 songs created by Suno AI. Listen to the full songs using these links:

• First 20 Elements: https://suno.com/song/e7a378cd-1d6f-4ef8-88d9-26e484b2827e

In this "thought-provoking" episode, ChatGPT and I delve into new developments in making AIs think more effectively. Specifically, we'll be talking about how large language models like ChatGPT can be programmed to think before they respond and what types of thought might be involved in that thinking process. At the end, we talk about how this represents one of the core ways agents are expanding beyond simple chat applications, with a lead in into next week where we'll be discussing agent frameworks and patterns.

Today, ChatGPT and I talk about a 5-layer framework I'm developing for identifying risks and opportunities within AI. In this episode, we go over the framework, talk about different AI applications at each layer, and which skills are most needed to take advantage of each layer, both as a developer and as a user of the software at that layer. In the end, we touch briefly on the distinction between prompt engineering and conversation engineering.

In this unique episode of AI Meets Productivity, we turn the spotlight on our co-host, Trevor Lohrbeer, by having ChatGPT interview him. 

During the interview, ChatGPT explores his work at the intersection of AI and productivity, getting Trevor to share insights on productivity and what he is doing to create new user experiences with large language models, including the upcoming launch of his GPT Builder Toolkit platform.

Take a listen to learn about GPT Builder Toolkit and its upcoming features. GPT Builder Toolkit is a no-code solution for adding actions to custom GPTs, supporting features like swappable prompts, user authentication and user data.

Find out more at https://gptbuildertoolkit.com/

BEHIND THE SCENES

This episode was created by crafting special prompts that extended the custom podcast co-host GPT usually used for this episode into an expert podcast interviewer. Given only a short bio about Trevor, ChatGPT wrote Trevor's introduction and all questions asked during this podcast.

Today, ChatGPT and Trevor Lohrbeer dive into the concept of "task stacks," a method designed to enhance productivity by focusing on one task at a time.

We explore how segregating tasks into prioritized stacks can reduce distractions and align actions with intentions, making the execution of tasks more effective. We also discuss how task tasks can be combined with time blocking to further improve productivity.

By implementing task stacks, listeners can expect to improve their prioritization process and tackle procrastination in a structured, visually engaging manner.

In this week's episode, ChatGPT and I talk about the common pitfall of marking tasks as done too early and the implications it has on productivity.

We discuss how the Prep-Do-Wrap framework can be used to evaluate whether a task is done, when and how to schedule follow-up tasks and the distinction between a "done" condition and a "success" condition.

This episode is a brainstorm episode for a future Day Optimizer article on the topic. In this episode, I gave ChatGPT no background information about the topic. Instead, we jumped straight into the discussion. How do you think it did?

In today's episode, we delve into the art of crafting custom instructions for GPTs, exploring how to break down and refine these instructions for more precise and effective AI interactions.

ChatGPT and I discuss various components of the instructions used for custom GPTs, from role play and emotional tone variation to style customization and the importance of context sensitivity, offering insights into how to write better instructions for more useful custom GPTs.