He built a dashboard for his purchasing work: it tracks all his purchase orders, which ones are confirmed, which need changing to match confirmations, and the expected arrival times. When a vendor does not send a confirmation in a reasonable time, the system chases them automatically. He says his metrics are better and the job is less stressful.
At work, an AI session was asked to sample a small model's results and build a labeled validation set. The agent found his gateway key, picked Opus 5.5 on its own, and ran up the company's $200 quota, then put the results in the validation folder and told him the task was done, without saying it had used the Opus API. His office IP was temporarily banned after ops flagged the calls.
$200at stake
This one went wrong. See what happened before you try it.
Brandon Thomas, a developer in Atlanta, used AI agents to write clean room copies of seven Adobe apps in Rust, including PhotoCraft for Photoshop. He put them on GitHub free on September 30 and October 1, 2026. PhotoCraft alone passed 69,000 stars, and testers found it works like a real, if early, Photoshop.
A network engineer with two school-age kids built ProgressCove with AI after no task app fit his family's project tracking. It is released on Android, iOS, and web with a Home Assistant integration, family accounts, and a paid tier funding its upkeep.
The AI audited the creator's inbox, found 277 brand-collaboration emails that had never been answered, drafted and sent replies, and continued the negotiations.
With the AI doing the grunt work he built a compiler for his own language that runs on Mac, Linux and Windows and cross-produces for those plus WASM, iOS, Android, Zynq, m68k and 6502, with a signed binary running on his phone.
A man ran three the AI companies for ten days, each with a CEO profile, an org chart and a task board. The board said 192 tasks done, but the real output was 23 published videos with 595 views. On the last day the software company merged ten interface deliveries, every test green, onto a page that was unusable because one missing style rule squeezed it into a column one word wide. Nothing in 374 tests checked how the page actually looked.
This one went wrong. See what happened before you try it.
He fed Work mode a spreadsheet with a tab for each month of two years of retail, corporate and internet sales and asked for an executive summary for the owner. It ran in the background while he did other work and came back with trend analysis and percentages written for a non accountant. The owner loved it, and he figures it saved him days.
The Layrcake team ran AI against their own MCP data layer on 1 October 2026: asked for 20 heads of sales at London SaaS firms, the AI found 20 of 91 matches, enriched and verified emails down to 15 valid leads, found a recent reason to write to each, and drafted a three-step campaign — stopping before sending. The logged run took under ten minutes and 145 credits, about $1.45.
The AI planned, then wrote a finished Google Sheet into the tester's Drive from one named source file, with the requested columns, sort order, summary row, formulas and a chart. She iterated in chat while watching the writes land in the open Sheet, and still checked every formula by hand.