His team's test framework drives the product, has an LLM judge it, and files the issues
His team runs a fuzz framework that drives a Migration Assistant with Playwright, has an LLM judge results, and files issues back into their system.
A random mix from the whole library. Shuffle for a new set.
His team runs a fuzz framework that drives a Migration Assistant with Playwright, has an LLM judge results, and files issues back into their system.
Jacob runs four AI products alone. He built Tracetify, a tool that rebuilds how a competitor got its first users from twelve public data sources, in about three days, and gave the AI one instruction for the whole interface: you decide. Users later asked whether he had hired a designer.
His company runs three coordinated agents on its product account: one watches mentions and a watchlist, one writes posts and replies, and one controls quality. Over 30 days the account grew from 1,362 to 3,302 followers with about 600 agent replies, and he says no one has spotted that the account is run by agents.
Gave the AI a classification task he had been doing by hand and it raised accuracy from 81.3/89.9/92.1 to 90.9/99.3/99.6 in a little over 2 hours.
His order was doubled, and the company support bot closed his ticket. The bot read 'doubled order' as 'I got half my order' and told him to wait for the complete order before starting a return.
A peer reviewer put a manuscript's reference list into AI and asked it to find a web link for each one. Two references were fictitious and 13 more looked suspect, so the reviewer rejected the paper.
The AI made a new OBS launcher in about a minute that opens OBS and refreshes and checks all of his sources.
A home services business was losing leads nobody followed up. The AI scores each new lead A, B or C against the owner's 4 rules, writes a short summary on the contact, and drafts personal WhatsApp follow-ups that a person checks and sends each morning. The poster says it brings in 5 extra contracts a month at about 9,500 euros each.
He is not a developer. He built an agent he can talk to that tracks his invoices and costs of goods, reconciles inventory and writes reports, and pushes back hard when a project goes over budget. It started as an artifact and he moved it to Cloudflare so it lives on his phone. He is now thinking about making it a tool for his whole industry.
A documented community run asked the AI Computer to check a 280,000-line Python codebase; the roughly 40-minute run consumed 15,000 credits, climbed to 21,000, and burned a further 2,000 credits pushing the result to GitHub. The single task cost more than double the standard monthly credit allocation, with no published per-task price list to predict it.
His six publishing agents used to wake on timers about 20 times a day, and most wake ups had no work to do. Switching to waking only when a task arrives cut the daily bill from $5.86 to $1.21, about 79% less, and September's four long form articles cost $50.53 total.
He had the AI gather his SPARQL training materials and present them on a website in a way that is easier to engage with. The site is live on the SIB Swiss GitHub pages.
Used the AI to build an omit needless words app and used it to edit a document he is working on, accepting about a third of its suggestions.
Team Data Platform Agent, evolved over 2 years, is widely used for SQL questions, on-call alert triage, lag monitoring, and scheduled workflows fixing bad data.
He rebuilt vdbx.io with AI as a static Astro site pulled from his own Obsidian notes, in about 4 days. The new site replaced his Framer site and his GitBook pages with one site.
Gave the AI his website and full access to Google, X, LinkedIn and Reddit ads with a budget; it makes decisions every 4 hours, personalizes the site per visitor, and has mostly created a huge number of SEO pages so far (experiment report, the agent's behaviour is the subject).
A solo founder runs his game key store with about a dozen AI sessions, one each for code, pricing, copy, bookkeeping and suppliers. They stay in sync through a shared record where facts are only ever added, each fact says whether a session confirmed it or just read it somewhere, and his decisions go into a numbered ledger the sessions look up instead of asking him again.
Andrew Connell moved 426 pages, 281 course pages, 217 videos and 8,911 comments back to his own site after closing his Voitanos business, with every old URL answering with a single 301. The AI did most of the work over six months in his personal time, and his site went from about 5 percent of the organic search share to close to 90 percent in two months.
Used the AI on the 20 dollar tier in a weekend to build good enough coverage on Floci features to make it useful to him; others then continued the work.
His AI server checks what the current shipped Micro.blog app versions are and whether an app is in review. It updates the public help page by itself.
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.