The AI found a serious security hole in a state legislature website while archiving it
She was using AI to archive legislative material when it flagged a serious vulnerability in a state legislature's website. She had to stop and report it.
Real things everyday people got it to do, and how to try them yourself.
A random mix from the whole library. Shuffle for a new set.
She was using AI to archive legislative material when it flagged a serious vulnerability in a state legislature's website. She had to stop and report it.
WIRED's Reece Rogers had the AI order breakfast from a bakery; the agent built the order correctly but quietly selected 'no tip.' Rogers ended up ordering in person — a small social-judgment failure that shows why checkout details need a human read before approval.
Jorden Tually (verified, 520K followers), answering a comment about visiting every country, explained he's been to 104 of 197 countries over 12 years and asked the AI for a realistic timeline. The AI interface analyzed his Instagram to determine his travel style, then generated a detailed month-by-month seasonal timeline starting in 2027 and projecting completion in December 2036, with an interactive world map showing visited countries in green and remaining in orange.
He is job hunting in tech, aiming at startups and big tech. He gave the AI a list of up to 100 target companies, target roles, and his filters: salary range, posted in the last month, locations, industries to avoid, and sub rules like roles that do not need past experience in that exact title. Twice a week the AI sweeps the list and brings back matches. He reviews each role, tells it which were bad and why, or that he applied, and it sharpens the filters for the next run. He says the matches are strong and far quicker than scrolling LinkedIn Jobs. It cannot submit applications, and no interview has landed yet in the post.
The AI books the user's tennis club clinics and pays for them, keeping the user's regular sessions in place without weekly admin.
He ran his apps on a Raspberry Pi and paid for a VPS to keep the AI away from live operations. The AI worked out the spec for a PC that could run two virtual machines, researched eBay and found a good deal, and he followed its instructions, mostly by pasting commands on his phone, to set up both machines and move everything over.
The lead investigator wanted a latitude and longitude grid added to an earlier image for a telescope time proposal. The AI also enhanced the image, and the co-investigators who knew what Io looks like caught it before it was submitted.
Jailbroke his Kindle with GPT-5.5 Sol and turned it into an external e-paper display; the project code was written by GPT-5.6 Sol.
Her boiler had stopped heating the water. She sent the AI about 15 pictures step by step, followed its instructions to turn it off, switch a few handles and reset the pressure, and the boiler worked again.
To mark the start of the semester, he asked the AI to build a macOS widget. It counts down how many weeks are left until the last class, and how many of those weeks are real teaching weeks.
Every idea is a real post from someone who tried it, summarized in our words, with a link to the original. App names are swapped for plain words.