Thinking Out Loud
Are you still talking to yourself if a robot is responding?
This morning I set out for a quick fifteen minute run, but it became an easy forty-five because I got wrapped up in a conversation with ChatGPT. And if you’re only familiar with OpenAI’s Voice interaction from their TV ads you should know they are selling the capability short.
Each ad involves a couple of old fogies being fun and a little silly with this impressive technology. The ads are clearly meant to show you how easy it is to use, but the disconnect between what they show and how I use it is striking. They make voice interaction look more like a party trick than a valuable tool.
I spend a lot of time using Gemini, ChatGPT/Codex, and Claude and there is no question that OpenAI is ahead of the game on voice interaction. I use it all the time. I will pause a YouTube video or podcast and open up voice mode to ask questions about the topic. I on the stationary bike watching Brian Cox talking about black holes, and then I am asking ChatGPT about event horizons 1, Riemann sums, Zeno’s paradox, and exploring how they relate to each other. 2 I learn really well through dialogue, and voice mode gives me an outlet to interact with the material.
Back to this morning’s run: lately I have been revisiting graph databases as I think more about knowledge graphs and building a scaled pointer system for agents. So I as I went out for my run I threw on a podcast and within five minutes I had spun up ChatGPT Voice and started asking questions. And the questions quickly brought me to my… Read more