🔑 The One Thing AI Still Can't Do for You

Good morning. Someone close to me bought a used EV this month with AI doing the research and him doing the challenging of it. Today covers that, plus the family calendar project my wife and I are only just starting ourselves: one shared calendar people will actually check instead of quietly ignoring.

Someone close to me just bought a used EV, and AI did most of the heavy lifting on research: comparing trims, checking recall histories, running the numbers on total cost of ownership. He used ChatGPT alongside a dedicated automotive research tool built for exactly this. Genuinely useful.

It also needed constant correction. Every time the AI got confident about something, he had to stop and push back with hard questions: how the range estimate actually held up on a highway drive in summer heat, what charging near his actual daily route really looked like, whether the used battery pack still had a warranty that transferred. AI does not naturally think that hard about your specific life unless you make it.

That is the real lesson, not that AI is untrustworthy, but that it is a research partner, not a decision maker. It found options faster than he could have alone and asked questions he had not thought of. It also happily agreed with whatever direction he steered it, until he steered it somewhere better.

The catch: none of this replaces a mechanic's inspection or a test drive in the conditions you will actually use the car in. AI can tell you what to ask. It cannot answer for the machine itself.

If you are using AI for a decision that matters, EV or not, the move is the same. Let it do the legwork, then challenge it before you trust anything it tells you.

Go further: See how a fake brand fooled ChatGPT into recommending it anyway

FROM OUR PARTNER

Leadership is asking: are we getting value from AI? Which tools are worth the spend? Where are we exposed? Right now, most teams have no idea.

Harmonic Security Usage Explorer changes that.

You get a complete picture of how your organization uses AI, automatically categorized into custom tasks and use cases.

You’ll see the projects being worked on, who’s using what tools, where AI investments are driving value, and where employees are engaging in risky behavior.

CIOs can rationalize spending and cut wasted licenses. CISOs can pinpoint where risk exists and neutralize it. AI committees can show exactly how their efforts are paying off.

See it in action

My wife and I are both still figuring this out. She got a step ahead of me a few weeks ago by connecting ChatGPT to her Google Calendar, and it has already caught a couple of scheduling conflicts we're not entirely sure how it spotted. I have not done that part yet.

But the bigger problem is not one person's calendar, it's that everyone already has one. A couple probably each run separate work calendars. Their kids have calendars on their own phones. Nobody wants to abandon what already works just to retype everything into one master calendar someone else built. The real task is combining what already exists, not starting over.

Step 1. List every calendar already in use in the house: each parent's work calendar, each parent's personal calendar if it's separate, and each kid's calendar, a teenager's phone calendar counts.

Step 2. Ask AI for the exact steps to make each of those calendars visible, at least as free or busy blocks, inside one shared view on whatever app your family already leans on, since the steps differ between Google, Apple, and Outlook, and differ again if one calendar is tied to a work account.

Step 3. Create exactly one new calendar for things that need full detail visible to everyone: school events, family plans, shared appointments. Nobody has to publish their whole personal schedule to make this work.

Step 4. Decide per person how much detail they share versus keep private. A teenager's calendar might stay busy and free only, while the new shared calendar shows full event details.

Step 5. Spread ownership of the shared calendar by category instead of by person. Whoever books the dentist adds the dentist. Whoever signs a kid up for practice adds practice. It never becomes one parent's job by default.

Step 6. Run it two weeks, then ask everyone what they actually checked and whether they're still missing something from someone else's calendar.

If it does not work: AI may assume your whole family uses Google Calendar, or give you generic sharing steps that do not match your actual mix of apps. Correct it immediately with your specific platforms, since the sharing steps genuinely differ between them.

What this doesn't fix: It cannot make a teenager keep their calendar updated, or settle a disagreement between two parents about what counts as a family event. That negotiation is still yours.

Go further: See how ChatGPT's connectors work, the feature behind my wife's experiment ◆ See Google's own guide to sharing a calendar with someone

FROM OUR PARTNER

Most AI resources are scattered, untested, or buried behind a paywall.

The Shift puts it all in one free newsletter, covering what’s new, and what practically works under 5 minutes a day.

A vault of 3,000+ vetted tools so you’re never guessing. A 1000+ prompt library for every use case you’ll actually hit and free AI courses that make sure nothing stays theoretical.

Plus, right now, 3 subscribers win a free 1-year Claude Pro subscription just for signing up.

Subscribe to The Shift

NEWS YOU CAN USE
This Matters at Home
Short, practical summaries to help you stay updated on recent AI news—and what it means for everyday life.

🔥 AI Is Learning to Decide Which Wildfires Get Fought First. Beyond early detection, researchers are building AI that predicts how multiple wildfires will spread two weeks out, helping crews decide which fires to prioritize first.

⚡ Turns Out AI Agents Guzzle a Lot More Power Than a Chatbot. A climate scientist tracked eight weeks of his own AI agent use and found it burns roughly 600 times more energy per prompt than a simple chatbot question.

💬 1 in 4 Americans Now Turn to AI for Emotional Support. A new survey found 27 percent of Americans now use AI chatbots for personal or emotional support, and nearly a third of regular users call theirs a friend.

🧳 Europeans Aren't Just Asking AI for Travel Tips, They're Booking With It. New research found up to 92 percent of AI-using travelers say it shaped their trip decisions, not just their research, moving AI from browsing tool to booking agent.

REAL WORLD
How We Used AI This Week
Real people, real uses—here’s how AI is quietly making life better, one small win at a time.

📚 Erin K., Alpine, AZ. She walked into her local library, photographed the shelves, and told Gemini what she had already read and loved. It scanned the photos and matched specific titles sitting right there to her taste, historical fiction with a touch of magic, slow burn romance, strong female leads, including Canticle for Leibowitz, which she ended up loving. She walked out with a list of new authors she never would have found just browsing the spines herself.

👉 What's the last thing you asked AI at home? Tell us here. One question, 30 seconds.

READER POLL

HomeZada turns the pile of manuals, warranties, and photos every homeowner accumulates into one organized system. Point your phone at a room and its AI photo recognition catalogs what is there, worth having if a fire or theft ever forces an insurance claim. It also predicts when big-ticket items like your roof or water heater are due for replacement, with a rough cost estimate attached. Start here.

PASS IT ON
Know someone who would like this newsletter? Send them here.

Missed an issue? Read the archive.