What AI Said I Said About AI
I ran my Substack posts through Claude. The results were eye opening.
Welcome to the Cloud Database Report. I’m John Foley, a long-time tech journalist, including 18 years at InformationWeek, who then worked in strategic comms at Oracle, IBM, and MongoDB. I invite you to subscribe, share, comment, and connect with me on LinkedIn.
I was putting notes together this week for a conversation with colleagues on trends in the news media and wanted to be well prepared with examples of my work. In the process, I came up with a new use case (for me anyway) for GenAI.
Working from a hotel room in San Francisco, I input three Cloud Database Report posts into Claude, Anthropic’s chatbot, to identify common themes. These are the topics I explore and write about week in, week out.
The resulting summarization and analysis by Claude was pretty good. And it got me to thinking: What if I were to feed all of my published content over the past 30 years in an AI model? That would give me the history of business & tech from my vantage point: from the Bell companies to OpenAI, from OS/2 to iOS, from x86 to Blackwell, from data warehouses to data clouds, from terabytes to exabytes.
That’s a project for another day, but something any career writer could potentially do to document and learn from with their life’s work.
‘Evolution, transformation, new frontier’
Back to the exercise at hand — Claude’s analysis of my analysis. I cut and pasted excepts from three of my most recent blog posts (linked below).
Here is my query and Claude’s response, unedited and in its entirety.
Prompt: “What are some common themes in the following articles?”
Output: “These articles share several interconnected themes about the evolving relationship between data and AI:
Data architecture evolution - All three articles discuss how traditional data architectures are being challenged and transformed by AI requirements. The first article specifically questions whether the 50-year separation between processing and storage will continue in AI data centers, while the others explore new data models emerging for AI applications.
Multimodal data importance - There’s significant focus on how AI systems, particularly modern LLMs, are working with diverse data types (text, audio, video) in new ways. The second article specifically discusses “advanced multimodal” approaches and their importance to AI development.
Emergence of agentic data - The second and third articles highlight “agentic data” as a new frontier - both the data that powers AI agents and the data they generate through interactions. This represents a shift from static data repositories to dynamic, learning systems.
Cloud infrastructure’s critical role - The articles discuss how cloud partnerships and infrastructure decisions are becoming strategic factors in AI deployment. The third article specifically mentions Salesforce’s partnerships with Google Cloud and AWS as critical to their agent strategy.
Industry transformation - There’s recognition that businesses are in early stages of AI agent adoption, with competitive advantage potentially coming to early adopters. The articles suggest we’re at an inflection point where the relationship between data and AI is fundamentally changing business operations.
Historical context and evolution - All three articles place current developments within historical context, noting how data management has evolved over decades from early computing through relational databases to today's AI-centric approaches.
The overarching narrative is that AI is not just consuming data but transforming how we think about, store, process, and generate data across business and technology sectors.”
>end of Claude output<
The full, original dataset
No a bad recap, IMO! If the above AI-generated synopsis piques your interest, feel free to go to the original sources and read the full posts.
Next: 50 years of Microsoft
Microsoft was founded 50 years ago, in April 1975, and a good part of my career as a tech journalist has been spent covering the company.
In my next post, I will share highlights, and a few lowlights, from my conversations with Bill Gates, Steve Ballmer, and Satya Nadella, and other takeaways from years on the Microsoft beat.
Great use of GenAI, John.
As we know, LLMs know what they know by mining databases. So as you’ve demonstrated, let the models go to work on our own data.
It’s what organizations should be doing. Why not individuals?
The new motto: Here, Claude — mine mine.