Questions AI Can Answer Now

Here are some questions I’ve asked AI recently. For anyone who lives in San Francisco, this should be nothing new; others might be surprised what AI can now do.

I lightly edited the questions and answers to make them into clearer examples. The original answers were a bit verbose, and I pulled out the most helpful parts.

Q: I’m trying to think of an idiom, it’s something like getting your feet out in front of your head?

A:

• “Getting ahead of yourself”
• “Putting the cart before the horse”
• “Jumping the gun”
• “Biting off more than you can chew”
• “Getting too far out over your skis”

Q: What kind of doctor should I see about my bruised hip?

A: Sports Medicine Doctor: This is probably your best bet. They are very familiar with ordering and interpreting relevant imaging like X-rays or MRIs if needed. Many sports medicine doctors are actually orthopedic surgeons who have done additional fellowship training in sports medicine. If a fracture is found and needs specific management, orthopedic surgeons are the definitive experts.

Q: Is there a way for me to see one calendar with all events across all my Google accounts on calendar.google.com?

A:

1. Open calendar.google.com in the first account
2. Go to Settings (gear icon) > Settings
3. Select the calendar you want to share under “Settings for my calendars”

etc.

Q: Quiz me on my understanding of “Why ‘controlling for a variable’ doesn’t (usually) work.”

A:

1. What does it actually mean to “control for a variable” in statistical analysis?
2. Why can’t we determine causality by simply observing association between variables, even when controlling for other factors?
3. What happens when we control for a variable that is in the causal pathway between our explanatory variable and outcome variable?

etc.

More examples:

• What’s the best way to book a rental car in SF? All the websites I’m looking at suck. (Answer: Costco.)
• Someone told me the “logical positivists” shared my belief that stories are overrated. Is that true? (Answer: kind of, but their reasons were different.)
• If my coworker and I started a knitting club together, would the Taft-Hartley Act prevent our knitting club from fundraising for a presidential candidate? (Answer: it’s complicated, but no.)
• Which songs should I listen to in order to understand the phenomenon of Taylor Swift? Focus on her early work. (Answer: Tim McGraw, Our Song, Love Story, etc.)
• Why is my tax software marking this form in this way? (Answer: it’s probably a bug.)

I also use AI every day to help solve programming problems. That’s probably around 80% of what I use it for; these questions are representative of the other 20%.

One or two years ago, before the latest generation of AI models, I didn’t have any good way of answering these questions. Now I do, and it feels like magic.