I used Google’s Notebook LM in September 2024 and I was impressed with the podcast it created, sharing a summary of my blog. A month later I had it do the same for a video I created with Joe Truss. This is a novel theory, not a general knowledge concept and yet the AI grasped the majority of the concepts and did a very good summary.
Today I went back to Notebook LM because I heard it can now do video summaries. Again, I was impressed. While the accompanying visuals were not ideal, (we discuss complex geometry), the audio summary was excellent and it was valuable to see what takeaways were summarized and how the ideas were structured.
I then explained some of the geometric issues and the AI produced a pdf with the correct geometry. Joe and I then tried creating a slide deck, another new feature. The resulting text was excellent again, but some of the images were not quite accurate, yet we could see the possibilities in correcting the details and providing other sources to make it give us impressive results.
Reflecting on these improvements it occurred to me just how good this tool is now, and yet this is the worst version it will ever be. Artificial Intelligence and robotics are both advancing exponentially in capabilities. It’s exciting to think that what we are capable of using these tools for now will be considered simplistic if not archaic in just a few years.
Today I saw a video about a Chinese company that is selling a three and a half foot tall humanoid robot for the price of a new iPhone. It is not a simplistic toy, it is extremely agile and comes with a fully programmable operating system, meaning it is completely trainable for skills it doesn’t come with. That same company is expecting to reach 1,000 units produced a month by the end of this year.
We are in an era where advances happen daily, and what we marvel at today will be commonplace tomorrow. Every day the advances get a little better and so we are perpetually living with the worst technology we’ll ever know.

