AI - from Article > Podcast > Video?

Imagine turning your written words into a dynamic video with just a few steps. This isn't science fiction—it's the power of AI transforming how we create and consume content. From in-depth articles to engaging podcasts and captivating videos, discover the seamless journey through AI tools that makes it possible. Ready to revolutionize your content strategy? Let's dive into how articles evolve into podcasts and videos, enhancing your reach and impact with every step.

1. Over the past 2 weeks, I created a series of articles about Large Language Models (LLMs) targeted to Creative Professionals. In this case - photographers and architects (though everything I shared is practical and applicable to any Creative Pro!)
2. I opened each article, one-at-a-time, and chose File > Print (print as PDF). This gave me each article as a standalone document I could upload to other LLMs (for evaluation if needed).
3. I created a new Notebook in Google’s NotebookLM.

This is a great product to play with and has lots of potential. If you would like a behind-the-scenes on how this product came to be, I recommend you listen to the podcast shared on Lenny’s Newsletter:

There are multiple file formats that can be added to this tool. I uploaded all of the PDFs created in step 2.

I started in the NotebookLM, by asking it (via the Chat at the bottom) to provide a summary of all the documents included. This is what it gave me back:

Podcast Episode Overview: Getting the MOST from LLMs

This podcast episode will provide an overview of the shared articles focusing on Large Language Models (LLMs) and their applications for creative professionals, especially photographers and architects. The episode will cover:

• Introduction: Begin by introducing yourself, Brian Sykes, and your work helping creative professionals integrate GenAI into their workflows while retaining the human element [your previous message]. Highlight your businesses, theBrianSykes and A.I. Lab [your previous message]. Mention your Substack, the AI Lab, where you share insights and resources on AI for creatives [your previous message].
• What are LLMs? Define LLMs in simple terms, emphasizing their ability to understand and generate human-like text [1, 2]. Explain how they are trained on massive datasets, enabling them to perform tasks like writing, translation, and coding [2].
• Beyond Text: AI for Various Media: Explain that while LLMs focus on text, AI extends to other media like images, video, and music [3]. Mention tools like Midjourney and DALL-E for image generation, Runway and PikaLabs for video, and Udio & Suno for music [3, 4]. Briefly discuss multimodal models that combine different data types, such as text-to-image and text-to-speech [5].
• Vector Databases and AI: Explain the concept of vector databases and how they store information as vectors, which are lists of numbers representing data features [6, 7]. Use the analogy of the hexadecimal color model to illustrate how colors can be represented as vectors [8, 9]. Emphasize the role of vector databases in enabling efficient search and retrieval based on semantic similarity [10, 11].
• The Power of Priming: Stress the importance of providing context to LLMs to get specific and relevant results [12, 13]. Illustrate this with the example of asking for cake ingredients – a generic request yields generic answers, but specifying “Flourless Chocolate Torte” narrows the focus and provides targeted information [13, 14]. Explain how priming enables semantic search, allowing the AI to understand the meaning behind the user’s words [15, 16]. Showcase the practical application of priming using the Flourless Chocolate Torte example in conversation with ChatGPT [16-23].
• LLMs for Photographers:Session and Camera Settings: Demonstrate how photographers can use ChatGPT to get camera setting recommendations tailored to specific scenarios, like high school sports photography [24, 25]. Encourage listeners to input their camera models, lenses, and shooting situations for personalized guidance [26].
• Observing and Recreating Images: Show how Claude.ai can help photographers recreate specific shots by analyzing reference images and suggesting camera settings, composition, and lighting [27, 28]. Discuss Claude's ability to provide textual descriptions of diagrams to guide the setup process [29, 30].
• Exploring the Unknowns: Highlight the importance of using LLMs to fill knowledge gaps [31]. Encourage photographers to engage in genuine conversations with AI, expressing what they don't know and seeking step-by-step guidance [32, 33].
• Meta.AI for Clarity and Social Media: Introduce Meta.AI and its potential for photographers [34, 35]. Discuss its role in conceptualization, research, pre-production, post-production, business and marketing, education and inspiration, and workflow automation [36-39]. Guide listeners on how to use Meta.AI to optimize their social media presence, including profile enhancement, audience insights, content strategy, and analytics [40-48].
• LLMs for Architects:Generating Renders from Sketches: Showcase how architects can use AI tools like Adobe Firefly to transform sketches into realistic renders [49, 50]. Explain the process of creating style references and structure references for achieving desired visual outcomes [50]. Discuss the use of generative fill for adding details and refining renders [51, 52].
• Real-World Applications and Insights: Share insights from architects already using AI in their workflows, as gathered from your community [53, 54]. Include examples and quotes from professionals like Abdul Azeem, Ar. June Chow, Mohamed Fendi, Sameer Z, Itziar Etayo, and Mohamad AbdAllah, focusing on how they use AI for concept generation, texture enhancement, research, and streamlining client communication [55-60].
• Exploring Other LLMs:IBM Granite and Mistral AI: Introduce IBM Granite and Mistral AI as LLMs with strengths in code and creative tasks, respectively [61-66]. Discuss their unique features, such as Mistral's European focus, multilingual capabilities, and emphasis on privacy [67-70]. Demonstrate their practical application by prompting Mistral to act as a studio design consultant for a photographer, providing layout and décor suggestions [71-75].
• Call to Action: Encourage listeners to explore and experiment with LLMs in their creative workflows. Direct them to your Substack for more in-depth articles and resources [from your previous message and source 112, 123].

Please note that while I have tried to provide a comprehensive overview, there are specific details and examples within the sources that could be further explored in the podcast episode. Additionally, you may consider adding personal anecdotes and experiences to make the episode more engaging.

Audio Overview : LLMs

The conversational Audio Podcast that NotebookLM produced - I can't embed in this post. You can listen to it on my Substack. The characters in the podcast talk for 12 minutes about the collection of articles and key takeaways.

5. I took the audio file - llms.wav - to Descript. It discovered there were 2 characters. I went to ChatGPT and had it come up with names for these characters.

It gave me a list of 7 pairs of names, of which I chose Ella & James: These names are classic and professional, suitable for a setting that aims to educate and inform.

I fed those names to Descript as Speaker 1 and Speaker 2. It automatically divides up the conversation with who said what.

6. I asked ChatGPT to use my 12 Essential Prompt Elements to create 2 descriptions ready for Midjourney, to produce photo-quality visuals of these characters.

It provided me with a solid base, that I expanded to produce my 2 characters.

Studio Photograph of Ella, an attractive, professional educational podcast host, engaging in a thoughtful conversation, looking directly at the camera with a slight, welcoming smile. Ella has brown long curly hair in a bun, wearing pro-level headphones, positioned in front of a SURE mic on a mic stand sitting on a desk. Ella has freckles and thick-rimmed red-frame glasses. Ella is wearing a button up shirt and she is a blue jeans kind of woman of 33. Setting: a modern, well-lit studio with soft background colors in shades of blue and grey. Soft, diffused lighting highlights her face gently, creating an inviting atmosphere. Medium close-up, shoulders up. Realistic, photographic style with sharp focus, slightly from above angle. Contemporary, intelligent, trustworthy feel. Portrait orientation. No background distractions, no vibrant colors. --ar 3:4 --style raw --v 6.1

Studio Photograph of James, an average looking male educational podcast host of 40, explaining a concept with animated hand gestures. James is wearing a flannel shirt of yellow and blue. James has thick, wavy brown hair with streaks of grey forming at his temples. James is wearing pro-level headphones, positioned in front of a SURE mic on a mic stand sitting on a desk. Setting: a modern, well-lit studio with soft background colors in shades of blue and grey. Soft, diffused lighting highlights her face gently, creating an inviting atmosphere. Medium close-up, shoulders up. Realistic, photographic style with sharp focus, slightly from above angle. Contemporary, intelligent, trustworthy feel. Portrait orientation. No background distractions, no vibrant colors. Smooth clothing textures, rough wooden furniture textures. Aspect ratio --ar 3:4 --style raw --v 6.1

So what can I do with these? I will show you NEXT time…

Substack Re-Stack

Until next time… Enjoy!

Brian Sykes