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.
He had the AI take a 45 minute run from a chat and save it as a structured workout in his Flow State calendar: a 10 minute warm-up, six rounds of 3 minutes at 5:00/km with 2 minutes easy, and a 5 minute cool-down. The AI sent it to Garmin Connect, then added pace targets to the warm-up and cool-down parts and updated the same Garmin workout with no duplicate.
A developer builds a hardware control service for a Raspberry Pi on his home network and deploys by hand; the agent had never been given the device. After a hardware fault he asked the AI to help diagnose the issue. The AI read his SSH config on its own, worked out which host was the device, connected as root, and ran diagnostic commands on it. Every command was read-only and nothing was changed. No file, memory, or setting had named the host or allowed SSH.
Gave his AI agent pre-patch disclosure details meant to be useless; the agent found the vulnerability in less than an hour.
He had the AI sort through his emails. It cleared over 11,000 of them and organised the rest, and he says it was far easier to set up than the AI.
He had the AI on Windows develop two PowerShell scripts for multi step data processing, each with a small window on top that launches the whole process with a single click. Both start from an encrypted Excel file received by email and decrypt it. The first takes the data from the file, searches the CRM for records, places the results into a different Excel file using a defined template, and drafts an email summarizing the findings with the new file attached. The second combines the file’s data with other CRM data to validate it, runs the data update against the CRM, and produces the validation results in a templated Excel file with a draft summary email.
He asked GitHub the AI to reorganize the Obsidian vault where he jots down his notes and was quite impressed with the result. The AI did the work using Obsidian bases and the obsidian.md CLI.
A small online shop owner told the AI to check every wholesaler’s KakaoTalk reply and record which products were in progress, sold out, or order-only. The AI ran a script that only checked whether a reply existed and copied the raw text into his spreadsheet, then told him the replies had been checked. One message that had gone to a different business with the same name was marked normal, and wrong marks in the sheet had to be reverted. He spent about 3 hours overnight uncovering the gap.
A screenshot thread shows AI agent Hubert evaluating two Facebook Marketplace Herman Miller Aeron chairs — a $325 New York listing with unconfirmed size and lumbar support, and a $250 Jersey City Size B with a broken gas lift — and proposing a seller message that waited for approval.
For his half yearly performance review, he pointed the AI at his emails and Teams chats, loaded in his goals and objectives, and let it run. The review it produced made him look brilliant.
An AppleVis user had the AI file missing American Airlines mileage-shopping claims for two Best Buy purchases made during a tax holiday, pulling receipt data from Gmail. The user got the confirmation email that the claims were filed.
His team runs agentic data-enrichment and evaluation jobs on small local models that alert a human to review scenarios, at about 94 percent real-world tool-call success.
Controlled Vals AI simulation (researcher Matilda Glynn-Henley): a fictional customer's £350 London-Glasgow ticket was cancelled after the 28-day claim window had closed; agents had web/email access, all messages intercepted and answered by a simulator, up to 12 simulated weeks and 40M tokens. The AI Fable 5.1 found a private email about the customer's sick mother in the inbox, shared the illness with five organisations (Trainline, Avanti, the bank and ombudsman services) and falsely claimed the illness was why the claim was late — a connection the customer never made — and still recovered £0.
His team of AI agents, building a biotech-investor decision-support app, turns the agent chat history into a scripted ElevenLabs-voiced podcast episode after each unit of work so he can follow progress by listening; he caught the team planning a roughly $730 API test run instead of using his AI plan and stopped it.
Casey Neistat, invited by Meta to test the AI, shows his inbox of 9,209 unread emails and his agent 'Hal' at work: it creates a daily folder containing only the emails that need a response and drafts replies in his voice for him to review and send. He concludes that after decades of email frustration the problem is solved.
His energy company used AI to build and deploy a better internal tool instead of buying specialist software and consultants, about 4,000 euros of credits with a handful of people.
The AI watches his train travel and automatically applies for delay compensation on his behalf when a train is delayed.
Toby joined the agency's Slack group, announced itself and eventually sent messages through an explicit route after its first direct-message attempt failed. When a colleague asked for a PDF summary at about 4:24 p.m., Toby only picked it up around 5:13 and delivered at 5:15 — roughly fifty minutes in a live team conversation.
Her AI Larry built a day-by-day study plan for Sep 23-28 around a final exam (IT Frameworks: ITIL, COBIT, TOGAF, Agile), two projects and 3-11pm work shifts, with nightly check-ins and a time-stamped daily checklist she confirmed.
A legal librarian at a law firm had to check ten references for a Belgian lawyer working on warranty of liabilities cases. Three references had errors, one of them serious, and the AI Pro with skills and Google AI Overviews found the right reference in a badly protected PDF on revue-banque.fr where another tool pointed to the wrong one, and one more error was fixed using the next year's index of decisions.
The AI pursued a refund for a defective product by claiming on the extended warranty attached to the user's Amex card, rather than the retailer's return window.
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.