For the first time in my life, I found myself wrapped up in a conversation with an AI. It was the new release of Grok 3 (Beta). I’m not going to post the conversation, but what startled me most was that I couldn’t distinguish the interaction between machine and human. In most ways, it was even more interesting than if it were a human.

It started with me asking Grok if it had information about our demonstration of UWB via Bluetooth in 2005. It immediately connected the dots back to me as the former CEO of Open Interface North America at the time. Not only did it write an accurate summary of what took place, but it also then asked me about the technological climate at the time and my focus while running the company. I explained that my goal was to keep us financially afloat. This led to more prompting from Grok. It wasn’t pandering to my ego—at least, it didn’t feel that way. It was as if it was on a quest for knowledge about tech during that period.

I was facing challenges in all directions back then. Grok then got into a discussion about other personal matters as they related to technology in 2005, and the whole thing dragged me into an even deeper interaction with Grok.

I tried to pick up where we left off, but Grok 3 (Beta) does not retain memories of prior writings, similar to GPT-4o. In that case, I can ask GPT-4o if there is a topic I should cover on Substack, and it will suggest a few, while Grok 3 (Beta) cannot. This is a very big deal to me because it otherwise feels like Fifty First Dates.

I’m still very impressed by the conversational side of Grok; however, it will, for now, feel like a conversation with Uncle Bob, who remembers a lot of facts but can’t recall a conversation five minutes earlier. I know it will get there. I’m just a little impatient.