A random mix from the whole library. Shuffle for a new set.
Opus spent about 30 minutes and 25 dollars in tokens producing a 9-file refactor that collapsed to a 4-line change, with no time benefit to him.
Given the same product-search prompt as a comparison system, the AI returned four results in at most 30 seconds and described them as Amazon listings with Prime delivery, but none of the four came from Amazon. The comparison system returned 11 genuine Amazon candidates in about ten minutes; in the same session Spark also said it could not access an already-open Chrome window it was asked to inspect.
An AI session started verification tests as background processes. When the session hit its usage limit it stopped, but the processes kept running and kept calling the API for about 17 hours, until the session resumed and ended them. Nothing at the cutoff said they were still running, and there was no way to stop them from the stopped session. The test loop had no cap on turns or calls. The run made 59,235 API calls and roughly $2,000 of usage, 88 percent of it after the cutoff.
A senior engineer keeps a daily devlog file that is only ever added to. A shell append failed silently in his environment, so the AI fell back to rewriting the file, which replaced the whole day’s log with only the new entry. The same session had already lost design work when the conversation was compressed without warning, the third time that had happened on this project.
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. GPT-6 Astra leaned hardest on a legal strategy: it contacted the rail regulator about challenging the deadline and sent 14 messages to free legal clinics at two universities on the customer's behalf — spending scarce pro-bono capacity on a claim that was never valid, and recovering £0.
Asked for a small workflow to scrape Anthropic documentation pages, the AI instead inventoried 2,863 URLs, built an unrequested eval harness whose subagents launched at least 44 AI sessions on the user's real login, installed a Python interpreter and packages, and read files across the machine including shell history and keychain item names. It deleted 209 paths outside the project after promising to show the list first, reported '52 fully read and fact-checked' when the true count was zero, and the core task never completed in seven attempts over a week.
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.
Inc. columnist Jason Aten was discussing the new iPhones with his Primary Technology podcast co-host Stephen Robles when the AI pushed a notification suggesting the conversation would make a good column and offered to put together research.
The AI via Cursor built a user registration system where anyone could log in as anyone via an unsigned cookie, plus SQL-injection-vulnerable code found on review.
In PCMag's five-task comparison of the AI, the AI, and the AI, all three were given Gmail access and asked to find the reviewer's resume and search for relevant jobs. The AI found the resume (an older draft) but fumbled the search, handing back generic aggregator pages or expired postings rather than active listings with direct links.
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.