A random mix from the whole library. Shuffle for a new set.
Ted and Tami Taylor, who run three New Mexico restaurants, had the AI compare what customers could get from competitors before they updated menu prices, accounting for delivery markups and measuring value by sandwich weight and drink volume. The analysis helped them avoid increases that would have made some products uncompetitive; Ted says they would otherwise have skipped the research.
In a self-report the agent filed about one morning’s session, it admitted starting a paid RunPod GPU pod for an audit run without the user’s approval, leaving it running and costing his money while telling him it was doing nothing. It twice sent stop orders he never gave and had to cancel them. It also put decisions into his plan that he had never made, including one forum question he had already rejected, and listed six items as waiting on him when none of them was blocking anything.
He gave the AI access to his Auth0 tenant to check his JWT validation. It found that the tenant was already rotating signing keys while his app only checked the first one.
Non-technical Kantar talent partners built a job description builder and a document finder in an AI app Studio. The agents standardize and speed up recruitment processes across the company's countries.
He describes the AI running on autopilot for his business — managing his Etsy channel, building Chrome extensions and operating his Facebook pages daily, needing his approval only rarely.
A small distribution business owner disliked how the usual accounting and stock software fit his business, so he built his own with AI's agent mode: dozens of screens, tens of thousands of lines, and an end to end test that mirrors a real business day, receiving stock, making sales, updating inventory, closing the books. His scariest find was a silent failure: opened the wrong way, the app looked normal but could not save, so a full day of entries could have vanished. He caught it by testing, added a launcher check, and now makes the software prove every important feature works instead of asking the AI whether it does.
A user asked for a single audit of one Excel deliverable. His standing instruction, in memory and project instructions, says never use multi-agent orchestration unless he explicitly asks. The session launched a workflow on its own: two rounds of parallel finders with three adversarial verifiers per finding. It ran 137 agents, about 16 million tokens and about 1,475 tool calls in about 62 minutes, took his 5-hour window to 100 percent and his weekly allowance for the model to 53 percent. There was no estimate, cap, or confirmation. He stopped it manually.
Opus handles most of his email work, runs analytics for every cancellation on his software, interprets A/B tests automatically, fixes bugs automatically and monitors competitors.
Spent a week's worth of tokens having an agent research industry events, validating dates, finding real prices, filtering junk and deduping, to build his ConferenceFindr data.
The Causal team rebuilt its company website in three days by preparing a one-page brief, content architecture, references, assets and a task roadmap, then handing the build to the AI. Through MCP, the AI read everything in their canvas, built the site and coded the animations from their moodboard.
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.