In partnership with
Good morning.
Today's issue is about the gap between how confident AI sounds and how right it actually is, plus a personal project: I turned myself into a bobblehead this week just to see how good the tools have gotten. Both are good reminders that impressive and correct are not the same thing.
AI INSIGHT
It Sounds Right. That Is Not The Same As Being Right.
AI has gotten better at sounding sure of itself. That confidence is the problem, not the fix.
A model that hedges every sentence loses your trust fast. One that answers cleanly, in full paragraphs, with numbers and a calm tone, earns your trust automatically, whether it deserves it or not. Polished language does that job on its own, no matter what is actually underneath it.
The gap between the two keeps growing. Newer models write with more structure, cite more sources, sound more like an expert. None of that tells you the answer is correct. It tells you the model got better at the delivery, a separate skill from getting the facts right.
Here is the number: 35% of American households now use AI to find information. Only 14% of people say they trust what it tells them, according to Census Bureau data reported by USAFacts.
That gap is not irrational. It is people noticing something true. Fluency and accuracy are not the same thing, and most of us were never taught to tell them apart, because a confident wrong answer looks exactly like a confident right one.
Anthropic's own explainer on trusting AI walks through why that confidence shows up even when the model itself is uncertain underneath.
The fix is not distrust everything. It is treat a persuasive answer as a starting point, not a verdict. Ask where a number came from. Ask what would change the answer. A model that can back up its own claim earns more trust than one that just states it well.
FROM OUR PARTNER
2 Free AI Courses. No Credit Card Needed.
5,000+ professionals use Skill Leap to get ahead with AI. Right now, two of their best courses are completely free - Claude 101 and the 14-Day AI Boot Camp.
Claude 101 covers prompting frameworks, Artifacts, file analysis, and real-world workflows in 19 lessons.
The Boot Camp covers ChatGPT, Gemini, Midjourney, and prompt engineering in 16 lessons. Downloadable workbooks. LinkedIn certificate.
Zero cost, no credit card, no catch.
AI TUTORIAL
Turn A Photo Of Yourself Into A Bobblehead
Image generated using Gemini
I turned myself into a bobblehead this week. Grey ballcap, black polo, my nickname on the base, the whole look. One photo in, one image out, about ten minutes total.
1. Pick one clear, front-facing photo. Good lighting and a plain background give the model the most to work with.
2. Upload it to ChatGPT, Gemini or Claude and describe the figure style you want: bobblehead, boxed action figure, or caricature. Name the specific details that matter, like an outfit, a hat, or text on a base or box.
3. Ask for two or three versions before you settle on one. The first pass rarely gets the proportions right.
4. Once you like a version, ask for one change at a time, like fixing the face or resizing the head, rather than a full redo.
Try This With AI
What it does: Turns any clear photo into a personalized bobblehead, action figure, or caricature you can save, print, or share.
"Act as a character figure designer. Using the photo I'm uploading, turn me into a [bobblehead / boxed action figure / caricature]. Give it a [grey ballcap and black polo, or describe your outfit], a base or box that says '[name or nickname]', and a plain [gray] background. Make the proportions look like a real collectible, not a cartoon."
Customize it: Swap in your own outfit, name, and figure style, or use a photo of a grandkid, a pet, or the whole family for a matching set.
If it does not work: If the face comes out too cartoonish or does not look like you, the source photo is usually the problem. Upload a sharper, more front-facing shot and add "photorealistic figurine style" to the prompt instead of "cartoon" or "toy."
FROM OUR PARTNER
Learn AI in 5 minutes a day
You don't have to scroll every AI thread, track every new tool, or watch every demo.
The Rundown AI breaks it all down for you — the latest AI news, tools, and tutorials in one free 5-minute email every morning.
Trusted by 2M+ professionals at Apple, Google, and NASA.
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.
📱 ChatGPT Now Switches Itself Into Teen Mode. Accounts it reads as under 18 default to study nudges, break reminders, and no romantic language. Parents can link in and set quiet hours.
🚗 Ford's AI Reads The Owner's Manual So You Don't. The Ford and Lincoln apps now answer questions about your specific vehicle, pulling live data on tire pressure, towing limits and maintenance. Rolling out to 8 million owners.
🎰 ChatGPT Cannot Win You Powerball, But It Can Help. AI cannot beat random draws. It can steer you off birthdays and neat patterns, so if you do hit, you split with fewer strangers.
🎮 Gamers Are Prompting Their Own Call Of Duty Clones. People are building playable shooters and dungeon crawlers by chatting with AI models. They are rough and expensive to make, but the demos keep improving weekly.
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.
💰 Mike H., Hollywood, SC. His daughter-in-law's family is outgrowing their house and wanted a budget to see where the money was going, but spreadsheets were not her thing. So Mike had Claude build her a browser based budget tracker instead, in short sessions over about two hours. He kept refining it prompt by prompt, income tracking, editable categories, spend charts, and Claude caught and fixed problems on its own. She now has a budgeting tool she will actually use.
READER POLL
How much do you trust AI's answers by default?
TOOLS FOR YOU
🛡️ Tool Spotlight: Incogni
Today's Insight was about knowing when to second-guess something that sounds authoritative. The same instinct is worth aiming at your own information online. Incogni scans data broker sites and sends removal requests on your behalf, the kind of listings that quietly resurface your address, phone number, and relatives' names for anyone willing to pay a few dollars. It runs in the background and reports back what it found and removed. Start here.
PASS IT ON
Know someone who would like this newsletter? Send them here.
Missed an issue? Read the archive.
Advertising Disclosure: We evaluate all recommendations of products and services independently. Clicking on links provided on this page may result in AI for Daily Living earning compensation, which supports independent publishers like us.
