Antony, this captures something we don’t talk about enough—the feeling of always being behind in AI, even for those shaping it. The insight on recursion is key, the key that unlocks—AI’s ability to improve itself is accelerating, but so is our own learning curve when we integrate AI into workflows.
The point about organized data unlocking latent value is a huge one. We’re seeing this dynamic unfold not just in business, but in governance. Take the DOGE email compliance exercise—it wasn’t just a bureaucratic request, but an AI-driven restructuring of what work was recognized and rewarded. The same principle applies across industries: AI doesn’t just analyze data, it reshapes incentives, workflows, and even what is seen as valuable. Just the surface of my thoughts.
Been exploring this idea in my own writing—how AI oversight mechanisms are evolving from tracking to structuring. Would love to hear your take on where you see this heading. Here's part 1 of my 10 part series and would love your feedback at each turn. Good stuff coming!
Totally agree with you re: NotebookLM - it’s incredibly useful. I keep finding new and varied use cases for it.
And you are a Readwise fan too - I've got around 25,000 highlights from 276 books I've read on Kindle since 2010. I had no idea about about the new AI chat feature - there goes Sunday afternoon :)
Absolutely. Readwise is shockingly good. All that reading and note-taking has become many times more valuable.
Antony, this captures something we don’t talk about enough—the feeling of always being behind in AI, even for those shaping it. The insight on recursion is key, the key that unlocks—AI’s ability to improve itself is accelerating, but so is our own learning curve when we integrate AI into workflows.
The point about organized data unlocking latent value is a huge one. We’re seeing this dynamic unfold not just in business, but in governance. Take the DOGE email compliance exercise—it wasn’t just a bureaucratic request, but an AI-driven restructuring of what work was recognized and rewarded. The same principle applies across industries: AI doesn’t just analyze data, it reshapes incentives, workflows, and even what is seen as valuable. Just the surface of my thoughts.
Been exploring this idea in my own writing—how AI oversight mechanisms are evolving from tracking to structuring. Would love to hear your take on where you see this heading. Here's part 1 of my 10 part series and would love your feedback at each turn. Good stuff coming!
https://ailoops.substack.com/p/what-i-learned-from-a-simulated-ai
Totally agree with you re: NotebookLM - it’s incredibly useful. I keep finding new and varied use cases for it.
And you are a Readwise fan too - I've got around 25,000 highlights from 276 books I've read on Kindle since 2010. I had no idea about about the new AI chat feature - there goes Sunday afternoon :)