Channels for information sharing used to be so simple.

News came from the daily paper or the evening news.

Messages came as memos, carrier pigeons, scrolls.

Entire generations learned of world events long after they’d passed, their understanding shaped by hearsay and limited perspective.

Now, we are in a world of infinite information. Real-time, saturated, eye-ball-burning 😳 amounts of information. It’s never been easier to create content and harder to ensure that content is consumed. In my day-to-day, I have the opportunity (and burden) to learn across hundreds of slack channels, emails. an extensive intranet, a learning portal (seismic), a doc repository (box) and a million other places - and thats just internal content.

The quantity of information isn’t going to stop growing anytime soon, but how we organize it may change. AI has the potential to revolutionize how we access and learn information. It changes accessibility, comprehension and efficiency of information dissemination. This presents an opportunity for anyone who is trying to create and share information.

If you find yourself in a position where you are trying to teach, sell, or evangelize, you should be using AI to do it more effectively. It starts with choosing the right medium and then leveraging an exciting set of publicly available AI tools to do it better and faster:

Video

Video is the richest medium for information sharing. It captures all of our senses (putting smells aside here). It’s also the costliest to create, especially high quality. There are some really exciting AI proofs of concept in video, but they are primarily just that today. No scalable solution is at the quality where it could be indistinguishable from non-AI video. It’s still a great place to be experimenting and is new enough that it really catches eyes. My favorite example has been Reid Hoffman and his “Digital Twin.” 

If you want to play around with AI video, I would recommend creating a digital twin avatar using HeyGen and creating video based on scripts via InVideo or Syllaby. None of these are perfect but they are both good enough to catch eyes right now. And the best part is that once you provide a 1-2 minute video sample, you can create videos scalably and easily.

Again, I’ll be the first to acknowledge that this is far from perfect. But the concept is new enough to catch eyes. Next time you are responding to an RFP, creating an internal enablement document or giving a big presentation, consider blocking some time to experiment with tools like these and see how you can stand out.

Audio

Audio is the most scalable medium right now. In the competition for eyeballs, audio is an easy way to make long reads more consumable. There are some really high quality options for creating AI audio that are nearly on-par with legacy options. Given that large language models are built for language, the translation of text to audio and audio to text is the most native there is.

Google’s NotebookLM was the “ChatGPT of 2024.” It launched a feature that allowed you to create a 10-30 minute podcast about any document or set of documents with 1 click. It entered the world and took the industry a step forward in terms of what is possible.

NotebookLM makes learning information as recreational as listening to a podcast. It’s totally free and worth experimenting with. They recently came out with new features that allow you to customize the podcast output (length, focus areas, audience), and supposedly significant new updates are coming before the end of the year. It represents the single biggest opportunity right now in information sharing. It has never been easier or more effective to distill dense information into consumable nuggets.

Text & Image

Finally, text and image are the foundational capabilities for LLMs and thus the most native to produce. You should be experimenting every day with document creation, summarization, and analysis. The best way to use text or image to learn or share information is iteratively. For this reason, I am a big fan of creating customGPTs and NotebookLMs for information sharing.

CustomGPTs are interactive interfaces for addressing internal functions. You should be analyzing your business function for areas where CustomGPTs might help. For example, I created a customGPT for this blog. As you can see below, it includes “instructions” that act as a base prompt that is consistent across everything I do as well as “knowledge” where I include writing samples. It’s still not good enough to do my writing for me but it does my outlines, brainstorms, dad jokes and metaphors. It is also a good way of ensuring consistency across image generation with a base prompt.

If you are looking to share information held in dense or lengthy PDFs, I would recommend, again, using NotebookLM. Its chat function allows you to ask questions of documents and get detailed highly accurate answers with citations to their location in the original text.

AI is reshaping how we teach, learn, and share. It’s no longer about adding to the flood of information; it’s about cutting through it with precision and purpose. Whether through a digital twin, a custom learning module, or a one-click audio podcast, the tools are ready, and they’re waiting for you to take them for a spin.