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.
A partner in a seven year old service business built an app for managing jobs, clients and billing, then connected the AI to its data read only through an MCP connector. Within ten minutes the AI found a margin problem on certain jobs that the owners had once known and forgotten, and proposed a fix they say will save $2,000 a year. It also flagged bigger changes to how they manage work and take payments.
Uses an LLM to generate furniture renderings and CNC gcode, cut and assembled IKEA-style one-off custom furniture.
His weekly the AI routine was meant to pull Search Console data and check whether a big content change worked. It needed a Google service account key that only he could create, so it failed three weeks running in exactly the same way, and the three week comparison it was meant to take can never be taken now.
He used AI to make a Telegram channel for NHK's Documentary 72 Hours. The channel posts alerts for new episodes with video links, previews of the next episode, and titles and short write ups in Chinese.
A man who runs his own services at home sent his AI an article about a PHP security hole. The agent patched the services that had a fix, stopped the ones that did not, set up a watch for new fixes, and updated and restarted nearly everything as fixes landed through the day.
Peter Yang gave the AI Duet one short request and a link to his YouTube channel. Duet pulled a year of his published videos and their public data and built a report with charts, which found that personal AI agent videos are his breakout topic and that his solo videos now beat his guest videos unless the guest shows a real demo.
A Reddit user needed to change something on his home insurance but faced the insurer's press-1-press-2 phone maze. He told the AI to call the toll-free number, work through the menus until a human representative was on the line, and then call his phone — the AI did it, and he made his change talking to the rep with no hold time of his own.
Slack fix-it button tells the AI to fix an issue in his app, run tests, UI test with computer use, and deploy it.
The poster of the spam text case also had his bot remove his personal details from data broker listings. He lists both jobs among the things the bot has genuinely handled for him.
Working with AI on his NeoBox emulator, he found setjmp was eating 81 percent of frame time, switched to _setjmp for a roughly 130x per-call gain, and the fix was merged upstream in Geolith.
In the same roundup, Jason Zhang listed a bonus feature: sending messages to friends via connected Messenger and Instagram accounts.
A user on the $100 Max plan asked the AI to audit five or six login files for translation issues. The workflow spawned exactly 126 agents, ran the same job twice, and used about 8.4 million tokens, roughly half his weekly allowance.
Self-described non-cook (just learned what a skillet is) named her agent 'MOTY' (Mother of the Year). It surfaced a sheet-pan dinner recipe from her saved Instagram posts, built a grocery list from ingredients she already had, placed the grocery order pending her approval, and later gave her cooking-timing reminders while she cooked.
Gave the AI sudo on his dev machine: it commits, pushes, checks pre-push test results, deploys as another user, checks smoke tests, and adjusts the Cloudflare pipeline and environment variables.
He used AI 4 Argon in Antigravity CLI to run his network and system administration. It bought the domain and set up the full infrastructure, with hosting for a landing page, email, DNS, SSL certificates and hardening.
A self-described desperate vibe coder commented in an Reddit showcase thread that the SaaS he built with AI, which auto-generates landing pages and tracks who will pay, made $16k in 3 days with 876 paying users. It spread via Reddit and Twitter and got noticed in a hackathon Discord.
A man who self hosts an XMPP chat server had audio and video calls that worked only sometimes. He gave an agent access to his phone over ADB and to the server logs, and within 30 minutes the problem was clear: a setting for STUN and TURN discovery had been missed in an earlier move, so clients behind home routers could not connect. The fix was deployed.
Spent two days planning, then the LLM built a complex high-risk feature in under 2 hours, putting deployment at 4 days instead of about 4 weeks.
The same man took a photo of a clock this morning and asked the AI to buy replacement batteries. The AI identified the clock from the photo, found its instruction manual, worked out which batteries it takes, and ordered them from Amazon.
He had wanted an automatic provisioner for his boards for a long time but it was a whole project on its own, so he told the AI to write it. Some physical steps still need him, but each board now takes 2 to 3 minutes instead of a few stressful hours.
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.