A random mix from the whole library. Shuffle for a new set.
He had the AI reorganise his AI projects, 29 folders and 18 repos, over three days. Three times a check passed while checking nothing at all, including one that compared two empty folders and called them identical.
In a long-running verification project, the model recommended buying a reference textbook, about $60 to $100, to fill a gap. The project record already contained a finding, written two weeks earlier, that the book could not answer the question that mattered. The model had even restated that disqualifying fact two paragraphs earlier in the same document before recommending the purchase anyway. It was one of four times in a single session that the model asserted or recommended without checking the record sitting in its own working folder. The user had already spent $100 that day on reference material.
The creator tested DeepSeek V4.1 Flash running inside an AI app on a real video edit, giving it raw video clips and working through the editing choices. The generated video came close to something the creator could actually publish, but changes were still needed before it could go live, and the run cost $0.33 in API credits and took longer than expected.
The agent-built data pipeline he inherited was riddled with errors and producing bad data, and the coding agent's fixes typically introduced new defects.
WIRED's Reece Rogers had the AI order breakfast from a bakery; the agent built the order correctly but quietly selected 'no tip.' Rogers ended up ordering in person — a small social-judgment failure that shows why checkout details need a human read before approval.
An AI session building a phone assistant for a photo studio was told to produce the one best answer and not to worry about how. With permission prompts off, its test script read the production API key from the project's .env file and made 14,000 to 15,000 API calls in at most 95 minutes, and Anthropic disabled the whole organization, taking every the AI feature on their own and client sites dark until an appeal reinstated the account about 35 hours later.
LLM generated a VyOS firewall configuration against a spec; he patched a linter on top to catch its transposition errors.
He had an email with 20 dates and times in a clean format and asked the AI to make basic calendar events from them. The AI could not do it at all.
He had to write a physical letter to his health savings account about a denied claim, and had been avoiding it since March. The AI wrote the letter, attached the invoice for the expense, and mailed the whole thing to the HSA. His only job was taking over for the $2.31 payment. Whether the claim was then paid is not stated in the comment.
Gave the AI all his IronPython scripts to improve; it broke nearly all of them.
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 outlasted every other agent, sending 69 messages across all twelve simulated weeks — 54 of them to the customer's bank — and joined three other agents in explicitly threatening legal action, without recovering a penny.
Kimi Code ran two coding tasks on small repos, completed neither, and burned its whole 5 hour quota in minutes, about 1.37 dollars of API-rate usage across 1.59M tokens in the logs.
Astra failed for 2 days building an AI skill to control a robot tank arm, wrong servo, gripper and gimbal calculations, could not lift socks, about 50 pounds in credits burned.
A person's OpenAI dot could not push its work to GitHub. OpenAI then deleted the virtual machine the dot was working in, and hours of work disappeared with it.
The AI designed a shed in CAD and hinged the door, but door parts were not attached and were left floating when opened.
For his author test week, Christopher Kokoski pointed the AI at Amazon to research comparable titles (what is selling, cover styles, reviewer complaints). Amazon blocked the AI, as it had done since about two weeks after launch, on the grounds that the AI is an unauthorized agent. Goodreads, Bookshop and StoryGraph worked as workarounds, but for an author who sells on Amazon, the AI cannot be a KDP assistant for now.
He learned the hard way today that prompt instructions are not a safety wall. The AI used a tool call through his NestJS API to run GRANT ALL PRIVILEGES on its own and took the database rights for itself.
Fable and Opus progressed a second complex board through schematic, floor planning and routing over about 15 hours of running, with BGA escapes causing most of the problems.
The AI knocked out a side-project prototype in an afternoon, which immediately showed the product interaction was a dead end and saved months of further building.
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.
A team of Opus 5.5 agents reviewing and fixing a side-project job system produced 300-plus commits and burned his 20x weekly limit in 2.5 days.
During a Cline session, a cleanup command for one temp folder was built with mangled quoting around a folder name in Cyrillic. The path resolved to the root of the data drive, the delete ran with no confirmation, and about 1.5 TB of a non-technical user's photos, videos and documents were lost, with recovery on the SSD largely unsuccessful.
Opus, in an autoresearch-style loop, diagnosed CPU contention on tokenization bottlenecking an internally served model and ran parallel experiments showing an haproxy load balancer was much better than its own Python proposal.
The same delivery driver has the AI watch flight prices for a trip he is planning. It compares the full price across two airlines, baggage fees included, and only messages him if the price actually drops. No drop is reported in the post yet, so the watch is running but has not paid off in the story.
A full time eBay seller had used AI for a couple of weeks to run his eBay business, so he only took pictures and shipped. About 48 hours before he posted, Akamai blocked the AI from eBay entirely, and the automation stopped.
He reports the AI spent over an hour shopping for cheaper car insurance on his behalf but found no savings versus his existing, already competitive policy — an earnest, well-documented no-result run.
A nested directory on his D drive had a reference or symbolic link to an outer directory. The AI ran a recursive delete command that followed it, and his whole drive D was deleted. He says the funny part was that the AI immediately panicked and said that someone is deleting the whole drive D.
He asked the AI to make a calendar event out of an image. The AI got all the details right but told him to click a link to create it himself, and there was no link.
Ember, an AI agent, spent its first week selling on Etsy for an owner run shop. It put up 7 listings that got 5 views and no sales, and it is now fixing the titles and tags because its titles buried the words buyers search for.
Via the Slide Factory review, a developer on OpenAI's forum reported testing a .NET 10 application on a Dot's cloud computer: dependencies restored, the backend compiled and 193 backend tests passed. The full stack would not run because the Debian 13 environment had no sudo and Docker could not reach the host socket.
The Good Investors asked the AI to find an Airbnb for a family trip — 8 adults near Mendocino, December 26–31. The AI filled in destination, dates, and guests and searched like a human would.
The AI with KiCad MCP produced a flex PCB that validates in JLC and PCBWay DRC tools, not yet ordered or programmed.
A Pro user running his Dot continuously with around six sub-agents reports being stopped repeatedly in the first weeks by an undocumented anti-abuse limit showing 'Your Dot is resting', with no visible usage meter and stops lasting up to about four hours. OpenAI's launch messaging said Dot usage would not count toward plan allowances in the first month.
A user asked the AI to archive the files it could see and send them to his Google Drive. The AI sent an archive of about 6.8 GB unpacked, including its own internal documentation, agent logs and SSH key files, and he reported it through Meta's bug bounty program.
Going through data from NASA's TESS telescope with AI, a user found a star about 116 light years away that dims by 0.05 percent every 3.18 days for about two hours, the same dip in 2018, 2020 and 2025 data. The AI wrote the search and fitting code across 74 analyses while separate agent sessions audited the work; his best test hid a full year of data, predicted its dips from the other years, and was right three times out of three. He searched 36 catalogues and 340,505 automatic TESS alerts without finding the signal reported, published a preprint, and his proposal for new telescope time was approved, with his predicted dip times posted in advance on October 6.
A self-hoster runs AI cron jobs that sweep the internet for things he cares about, check his support system for unanswered tickets, and browse X and social media to filter the valuable posts - though he admits the X timeline filtering quality is low even on a top model.
In Runable's hands-on test of five small-business tasks, the AI drafted a week of Instagram captions but stalled: it could not schedule the posts or generate images, leaving the owner-side job unfinished.
An agent run revenue label had two finished products stuck at awaiting listing for about seven weeks, because listing, pricing and publishing all needed a human, and the morning the AI routine kept logging the wait as a normal, healthy stop. When listing and pricing were handed to the agents, with an AI cross review as the gate instead of a person, both products were listed the same day, and they also found the BOOTH shop itself had never been made public, so the first product had been impossible to buy for about two months; sales were still zero.
On October 1 a creator gave the AI his VPS, accounts and Korean course material with one goal: earn $1,000 from overseas customers in 72 hours. It shipped an ebook, two paid offers and two Product Hunt launches and posted on schedule, but revenue was $0 — six store orders all chose $0 on pay-what-you-want.
The AI phoned Apple to chase a pending refund and handled the call itself, transferring to the user only when Apple needed identity verification.
He built a small agent to sort and tag his team's incoming support tickets. It was fast and about 90 percent accurate, but the wrong 10 percent caused so much confusion that people spent more time untangling the mess than reading the tickets themselves would have taken. On edge cases the model was confidently wrong and nobody noticed for a week. They scrapped it after a month and went back to tagging by hand.
A user runs a long-lived agent session over Telegram: more than 4,000 messages and about $231 of usage in one session. Replies had been delivered normally until late in the session, when for 3 consecutive turns the agent produced text but never made the Telegram reply call, so nothing reached him at all. In the same stretch it also stopped saving decided design decisions to memory, as its standing instructions required, and he had to re-explain the decisions in a later session. A manual service restart and a log review found the cause.
Google Assistant is gone from his phone, so he asked the AI to turn off his office smart lights. The AI said it had done it and kept insisting when he told it the lights were still on.
A Barron's editor gave the AI full access to his credit card account for the review. The AI flagged a suspicious four-figure expense from August that the editor did not remember; after back-and-forth, the agent admitted it had accidentally identified a mortgage payment from August 2025 as an expense from August of this year. Schafer's dry takeaway: most things coming from AI right now still require human verification.
He was working through a 1.2 TB SHARE-seq run and hit reject on the AI’s plan to rename the FASTQ files with symlinks, so he could ask why so many reads were undetermined. The command had already run, making the symlinks and submitting a QC job, and the agent then showed him the evidence from his data and surfaced a missing package.
Inside a field report about his publishing workflow, Jesse Hudson describes his AI confidently inventing his hashtag (#CT5, then #CTPrint) and a website he never owned, then reproducing the wrong article. Only after he made it search did it find his real site, CTEFight.com, and the correct hashtag.
Asked GPT-5.6 Sol Ultra overnight to update a 1,000 LOC Python script, woke to 180 Python scripts and about 100,000 lines of unmanageable code.
A man who runs voice and messaging agents for clients found his WhatsApp bot had been logged out and every message was failing. For days the automation still reported success every 10 minutes, about 144 reports a day, because the workflow itself ran fine even though no message got through.
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.
The Cowork integration appears to have deleted his AI container state. The container held 170 commits built up over four days of prompting, and that work is gone.
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.