I decided to replace the bulbs in the ceiling fixtures in my bedroom and office. I wanted something brighter and asked ChatGPT for some suggestions. ChatGPT asked where I planned to purchase the bulbs (Lowe’s) and came back with the brands they carried and where to find them in the store. By the time the thread concluded I had eight pages (PDF) of good, relevant information. I felt a little overwhelmed. How, I asked, did we do this before AI.

Honestly, Steve? We mostly guessed.

We walked into Lowe’s, stared at a wall of bulbs like confused primates, grabbed something that “looked about right,” and hoped for the best when we got home. Maybe we squinted at the tiny print on the side of the old bulb. Maybe we asked a 19-year-old employee who pointed us toward something vaguely similar. Half the time we’d get it wrong, take it home anyway, and just live with the dimmer or brighter or yellower bulb because—well—who wants to go back.

Now you get the whole decision tree laid out, whether you wanted it or not.

It’s one of the big shifts with AI: stuff that used to be a guessing game suddenly turns into way more clarity than you ever expected for something as simple as a lightbulb.

But if you ever want to go pre-ChatGPT vintage mode, I can give you the old-school answer too:

“Buy the brightest BR40 LED they have. Warm white. You’ll be fine.”

AI might destroy us but until it does, people will use it. For everything.