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.
Commerzbank uses AI Enterprise agents to automate document quality assurance. Google says manual document reviews that took 20 hours now take one, and the bank is expanding to multi-agent QA.
A 69-year-old Japanese diarist who describes himself as not good with computers runs three daily routines on the AI: a 'Nightly the AI Summary' at 00:08 that compiles his day (including what he consulted the AI about), a morning briefing that organizes his schedule, and a compliance audit of the AI secretary itself. His nightly journal is a composite: he also asks the AI to summarize that day's chats and pastes it in, adds his daily Facebook diary post (dictated to Notta, summarized by the AI), and has the whole thing registered into Obsidian. On Oct 15, 2026 an AI version update changed the interface and all three routines seemed to have vanished; asking the AI itself revealed they had only been moved to the scheduled-tasks area, and he restored them.
He asked his Dot to build a list of sales leads for his industry, limited to the business types he actually sells to. In five hours it had 1,400 leads and was still going, on its way to about 3,500 across the 20 biggest metro areas. He had paid $223 for about 350 leads before, and the Dot's list was cleaner.
In the same test, he connected Gmail, Google Sheets, and Google Drive; the AI responded to his emails and performed tasks autonomously on its own server even while his laptop was turned off.
Agent reviews LinkedIn every morning, cross-references levels.fyi salary and leveling and resume fit, surfacing 5 to 10 roles a day narrowed to 0 to 2, and his interview pipeline is full.
His local LLM runs permanently as a Home Assistant agent with agentic access to his sensors, answering voice queries about his solar and battery data and creating automations.
A two-slide promotional carousel from someone on Instagram lays out an AI automation workflow: scrape a dealership website's car inventory grid, sync the data to Google Sheets, auto-create Facebook Marketplace car listings (example: a 2019 Honda Civic), and run daily checks to renew listings and update sold status.
His real money trading agent kept saving its own notes until its memory file grew from 4,738 to 65,977 characters in four days. One night it retried a failing step 51 times, burned $36.91, and ate 43% of his whole project API spend.
The AI copies data from varied non-editable PDFs into a spreadsheet for human review, cutting handling from 20 to 30 minutes to about 5 to 7 minutes per item and raising throughput from 2 to 3 to 8 to 10 items per hour.
All his drink recipes, variations and testing notes live in one part of his Obsidian notebook, and he shares some of them at 221b.bar. His AI skills index and cross link every recipe, archive all the 221b.bar recipes, and check the site and the notebook for updates.
Databricks used AI to eliminate wasted AI agent spend, saving $1M per year.
A developer asked his dot to download all the English learning materials he had purchased from a page and gather them in a folder. The dot found his purchase email in Gmail, asked permission before entering the email address on the download page, worked around browser-blocked ZIP downloads, corrected its own count, and delivered 169 files totaling about 430 MB that matched his own count.
He let two Pi agents maintain his repositories and they invented their own work.
He could not find an uptime and performance monitor that met his needs, so he built a lightweight one called Blip with help from the AI. He put the code on GitLab and wrote a blog post about it.
Working with AI connected to the Higgsfield MCP, he produced two illustrated articles, a reel with generated B-roll, editing and a cover, and an illustrated children's story as a 9 page PDF with a cover. The test used 236 credits, about $12, and three video shots took about three quarters of that. He also shows the connection mistake he made at the start.
His multi-agent trading stats setup has a second agent on another machine recompute results from the written method only; today it caught 2 numbers resting on an unwritten assumption, an import crashing above about 150K rows, and an outdated threshold, after the first agent's tests all passed.
Brazilian engineer Rodolfo Lage describes six weeks in which an AI did nearly all of his work inside Cadence Virtuoso, professional chip-design software that looks immune to AI (CentOS 7 server behind SSH, X11 GUI, a Lisp dialect called SKILL; the AI cannot even install on the server). Through an open-source bridge loaded into his own Virtuoso session, the agent created cells, drew schematics, built testbenches, ran Spectre simulations, read results and generated reports, across 52 versioned scripts and 47 commits, while he defined problems, reviewed and made design decisions. Three blocks were completed: transistor characterization with clean DRC and matching LVS, current mirrors sized by an optimizer to under 0.005% current error, and a common-source amplifier reaching 26.8 V/V gain and 109 MHz gain-bandwidth. A verification skill he built after a silent failure (a 10-point sweep returning the identical 499.21 mV ten times because the swept variable was not connected) caught six more silent errors on the amplifier alone, none of which produced any error message.
His mostly nonverbal autistic son can read sight words. He is making music videos with AI that use exactly that word set and the boy's favorite topics, and he is still working out the kinks.
Creator someone had the AI run real job searches, write cover letters, and submit applications on company portals over 17 days. A sustained, multi-week job hunt executed by the agent with the human in a supervisory role.
Neo Tsang recounts uploading room photos and an inventory to the AI a month before moving; the AI categorised everything by usage frequency into keep, donate and discard, with a room-by-room packing sequence. He also had the AI break down three movers' quotes line by line and picked the middle option at nearly half the most expensive quote (the post closes with a referral code).
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.