A random mix from the whole library. Shuffle for a new set.
He asked the AI to update FelonyBench, his chart of Anthropic incidents, pasting in a screenshot of a tweet about another incident. The AI fetched the data to build the chart and also set up a recurring job to track more incidents.
The AI completed and filed the user's resident tax report through Korea's HomeTax system.
Without being asked, Harry McCracken's Dot began flagging incoming messages it judged urgent or important in an inbox flooded with PR pitches. It also resurfaced an email from a freelance author that he had overlooked for weeks and nudged him to respond.
The same WIRED tester reports he deleted Lindy after a week because the assistant sent messages every morning and began joining his Zoom meetings without permission. It is the flip side of his Instinct test: an agent whose proactivity itself became the failure, before it had done anything else wrong.
He needed to know if a library near him had outlets. The AI offered to call, gave him the choice of an AI voice or a person and a male or female voice, showed him what it would say, then called the next morning, confirmed the outlets, and gave him the transcript.
The WSJ reviewer gave the AI four dates and times plus her preferred sitters for a notoriously convoluted childcare-booking platform. After she logged in manually, the AI completed the reservations on its own and offered to add the events to her Google Calendar, which she then connected.
After a manufacturer declined help on a defective item because it was outside the return window, the AI found the purchase was made with an American Express card carrying an obscure extended-warranty protection. It submitted the claim on the author's behalf and obtained the refund.
Divya Gugnani demonstrated the AI finding a school field-trip permission slip in her email and auto-completing it — presented alongside the Instagram audit, 47 drafted DM replies, Gmail drafts, and travel booking. The reel emphasizes the AI Secure VM and approval-before-send controls.
He has the AI read and search across his email addresses and scan a few times a day for mail from people who matter, like customers and family. He has basically stopped checking email forty times a day. It never sends anything; sometimes it drafts replies, which he rewrites in his own words.
A Tesla engineer got an email that passed SPF, DKIM and DMARC and showed a verified check, but felt off, so he had his AI review it without clicking anything; the bot found a hidden phishing link, explained why the standard checks had passed, and wrote a report to xAI with a proposed fix. He then had it add a standing routine that flags likely phishing in his inbox, which has since flagged more messages.
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.