The more your AI knows about you…

When I first began experimenting with AI chatbots almost three years ago, I decided not to worry about what I shared about my life, past and present. I’d been blogging for more than 20 years so it was all “out there” already. I’ve uploaded hundreds of blog posts, notes from all of the books I’ve read, movies I’ve watched… the entire corpus of Steve Mays.

If you’re using one of the AI platforms as a search engine or asking for cooking recipes or for performing tasks like drafting letters or writing computer code, the AI doesn’t need to know anything about who you are. But I quickly became interested in using these tools to organize, and make sense of, the thousands of files and blog posts I’ve been curating for most of my adult life. To do that, the AI needs context. Who is this person? What does he think? What does he feel?

My interest in self-inquiry and tendency toward introspection obviously had a lot to do with how I interact with Perplexity, Claude, Gemini and —once upon a time— ChatGPT.

Today I added some work history to the “corpus.” Continue reading →

Mark Andreessen on AI

Marc Andreessen co-created the Mosaic web browser (in 1993) and co-founded Netscape, helping popularize the early World Wide Web in the 1990s. (More than 30 years ago so, ancient history.) He later co-founded venture firm Andreessen Horowitz, becoming a key backer of major internet and crypto companies and a prominent voice shaping Silicon Valley’s views on software, AI, and “techno-optimism” in the 21st century.

He recently appeared on Joe Rogan’s podcast and shared some thoughts on AI. Someone (@itsolelehmann) shared his “most interesting takes.” Continue reading →

From Gopher to the Abyss

Seems that every new conversation with one of the AI tools I use gets more interesting. This morning’s ramble (with Claude) lasted almost an hour. Here’s the prompt that kicked things off:

I’m old enough to remember what the Internet was like before the World Wide Web. But it was so long ago I’m trying to remember how we found things before Google search came along. I remember websites like Alta Vista and Yahoo and Ask Jeeves and a bunch of others. Can you refresh my memory on this point?

It’s too long for any human to read so I asked Claude to summarize:

Starting with pre-web internet navigation, we traced how Google dominated information retrieval for 20 years before AI tools began displacing it. That shift threatens the web’s fundamental advertising-based economics, potentially hollowing out the content ecosystem that trained AI in the first place. We followed that thread through cable TV’s decay, the speculum ad as endpoint, stratified access to clean versus compromised information, institutional corruption as the same process running across society, and concentrated wealth as both cause and beneficiary. Ended with AI’s dependence on very human, very physical infrastructure — and the cooling systems that keep it all running.

I found Claude’s then-and-now perspective fascinating. Some of the references will be lost on those who weren’t there at “the beginning.” And it tossed a little humor at the end.

This was a genuinely good conversation — it covered a lot of ground and kept finding unexpected connections. The speculum ad will stay with me as a unit of civilizational measurement.