Creative process mapping, knowledge graphs for scientific discovery, How Dwarkesh Patel uses AI to amplify intelligence, and more

📖In this issue

• Framework: Creative process mapping for AI augmentation

• NIH on risks and benefits of AI in diagnosis

• Instructional designers augmented by AI

• A big news week in AI, including frontier-performance open source models from Meta, top level inexpensive models from OpenAI, SearchGPT from OpenAI, and more

• Markus Buehler on knowledge graphs for scientific discovery, isomorphic mappings, hypothesis generation, and graph reasoning

• How star podcaster Dwarkesh Patel uses AI tools to amplify his intelligence

🤸 Framework

For superior Humans + AI creative outcomes we need to focus on process.

LLMs are most effectively applied at well-defined points within broader processes, using precise prompts and context for highly specific tasks such as content generation, ideation, structure mapping, analysis, adaptation, synthesis, evaluation.

The first step is to pull back to the broader process and identify where AI can add value to complex cognitive and creative processes, as shown in this chart. Click through for the full size image and applications of the framework.

👩🤖Humans + AI update

NIH on risks and benefits of AI in diagnosis

The U.S. National Institutes of Health (NIH) “found that an artificial intelligence (AI) model solved medical quiz questions—designed to test health professionals’ ability to diagnose patients based on clinical images and a brief text summary—with high accuracy. However, physician-graders found the AI model made mistakes when describing images and explaining how its decision-making led to the correct answer.” These findings help us design better Humans + AI systems.

NIH findings shed light on risks and benefits of integrating AI into medical decision-making

Instructional designers augmented by AI

A study shows that Instructional Designers (ID) aided by AI, and even novices aided by AI outperform IDs who do not use AI.

How Close is AI to Replacing Instructional Designers?

Humans + AI dentists

“Pearl, one of only three startups that has an FDA clearance to use AI to assist dentists make accurate diagnoses and provide patients with a “second opinion,” has raised $58 million. It is not a substitute for dentists, but partly helps patients be more comfortable with recommendations.

Pearl raises $58M to help dentists make better diagnoses using AI 

🔥Hot news in AI

• Meta launches open source Llama 3.1 models with performance similar to top commercial models - The Verge

• OpenAI launches inexpensive and very high performing GPT-4o mini with two months free fine-tuning - ZDNet

• OpenAI announces SearchGPT that will complete with Google and Perplexity to 10,000 beta users - Semafor

• X enables a setting on by default to train its Grok AI with your tweets - Ars Technica 

• There continues to be a proliferation of text-to-video generators. Here is a short zombie movie from Chinese company Kling, which is getting some attention.

🎙️This week’s podcast episode

What you will learn

• Accelerating scientific discovery with generative knowledge extraction

• Understanding ontological knowledge graphs and their creation

• Transforming information into knowledge through AI systems

• The significance of ontological representations in various domains

• Visualizing knowledge graphs for human interpretation

• Utilizing isomorphic mapping to connect disparate concepts

• Enhancing human-AI collaboration for faster scientific breakthroughs

💡Resources and insights

How star podcaster Dwarkesh Patel uses AI tools to amplify his intelligence

Dwarkesh Patel has swiftly risen to the top of podcasting world, with recent interviews including Mark Zuckerberg, Tony Blair, and Demis Hassabis.

This article summarizes how he uses AI to help his thinking and work, including:

• AI Integration: Dwarkesh uses LLMs to process and synthesize information, significantly boosting his learning efficiency.

• Learning System: He employs AI to create and manage spaced repetition prompts, ensuring long-term retention of new concepts.

• Podcast Workflow: Dwarkesh uploads context about his podcast guests into AI models to generate deeper insights and formulate better interview questions.

• Book Learning: He inputs text from dense books into AI models to generate summaries and explanations, aiding his understanding of complex topics.

• Worldview Building: Dwarkesh writes reflections and uses AI to connect and analyze ideas, helping him build a coherent and comprehensive worldview.