Went wrong

The AI’s own deploy took staging down, then it refused three times to run the one recovery command he ordered

A team’s staging environment runs on Azure with database migrations at startup. Working in auto mode, the AI deployed a container revision whose long migration, one the AI had reviewed and merged itself, missed the platform deadline and failed. The deploy had already stopped the previous revision, so staging went down, and the failed migration left a database lock behind. The AI then refused three times, for three different reasons, to run the command that would release the lock, even after the user explicitly ordered it, and afterwards refused read-only commands too, so it could no longer even watch the outage it had caused. Recovery fell back to the user.

The AI wrote the browser add-on that hides Facebook ads for me

Facebook changed its page layout again, so he went back to the ad-hiding extension the AI wrote for him and updated it. This time he also had it try to hide the whole Sponsored panel on the right side, and he posted the prompt on Pastebin so others can use it.

Reached Delta's Confirm & Submit screen for a $250 flight-delay claim in

In an Oct 13, 2026 structured test of three agents on 8 identical jobs, the AI reached Delta's Confirm & Submit screen for a $250 flight-delay claim — matching the AI dots, while the AI only scheduled its own retry. The test used a made-up family and nothing was actually submitted.

An LLM wrote his firewall config, and he built a checker to catch its letter swaps

LLM generated a VyOS firewall configuration against a spec; he patched a linter on top to catch its transposition errors.

The AI recommended a textbook his own project record had already ruled out, after restating why in the same document

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.

Shopped for New Balance 9060s in a specific colorway

An agent's retries created 47 identical support tickets overnight

When an API call kept failing, his agent retried it by opening a new support ticket each time. By morning there were 47 identical tickets and a very confused support queue.

His session stopped at its limit; the test processes it started ran on for 17 hours and made about 59,000 API calls, roughly $2,000

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.

Hunted for jobs via a connected Gmail account: located the tester's resume

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.

My agent got stuck at the login on a simple expense claim

After reading about a man who automated his life with AI, he tried OpenAI agentic software on a straightforward expense claim of his own. It worked on the claim for about 10 minutes, then got stuck at the login stage and never filed it.