Fragments

---

An odd experience today, where I inadvertently triggered AWS’s Q programming assistant to refuse to answer my request and refer me to AWS’s responsible use of AI policy. This happens, sometimes, if I use profanity to vent in a moment of frustration. But this time, I had typed “oh, thanks for catching that, that was pretty dumb of me” into the prompt. I then apologized to the bot for its own misunderstanding. “Oh, I’m so sorry,” I wrote, “I was talking about myself. Thank you very much, this was helpful.” The bot replied, kindly acknowledging the challenging nature of the problem I’d been trying to solve, and wished me a happy new year.

---

---

My whole life, two feelings about the progress of digital technology: first fascination, then terror—always in short succession.

---

Since at least the Olympics back in August, Google has been relentlessly advertising their Gemini AI assistant as a kind of super-assistant-and-best-friend, in some of the cringiest advertisements I've ever seen.

Until the influx of VC cash made generative AI cheap of free, you had to be a genius, spectacularly wealthy, or more likely a character in a film to create an invisible friend with the computer. There were limits on the scope of this particular insanity. What terrifies people is that now any old person can do it.

---

We spent the 2010s consuming Black Mirror and making every decision necessary to turn it into realism in the 2020s.