An LLM generated a whole compiler for his new language over 7 months
An LLM generated an entire compiler for his new language over 7 months of iterating and fixing, generating WASM runnable in a browser playground the LLM also built.
A random mix from the whole library. Shuffle for a new set.
Work
Jobs, side gigs and small business · 1,665 ideasTwelve hours with a top coding model left him 20 new bugs and features built to the minimum
12 hours with AI Fable 5.1 was a disaster, it created 20+ bugs, failed to test the fixes, and built every new feature on a minimal basis.
This one went wrong. See what happened before you try it.
Run my resale back office from an AI spreadsheet app
A solo reseller's the AI-built spreadsheet app is his back office: he dumps in a batch of receipts and it enters them, creates a folder for each new item, and sorts the item's photos and receipts into it alongside measurement files. The same app recommends prices, writes his listings and does resale research, which is the part he says makes the $20 a month worth it.
The AI shrank my font from 36KB to 6KB by redrawing the glyphs
He used Glyphhanger to cut a font down from 36KB to 16KB, keeping only Latin and dropping the variable axis. Then he let the AI try, and it redrew the glyphs and got the font down to 6KB.
A database fork with version control, built through 2,000 agent pull requests, now in beta
DoltLite, a SQLite fork with Git-style version control, was built with 2k agent PRs and released in beta.
Cut duplicate email automations and rebuilt a welcome flow in under an
A worked example in metaflow.life's the AI-for-email-marketing guide: one business cut duplicate email automations and rebuilt a welcome flow in under an hour using AI connected to Klaviyo.
One owner tried to run a whole business on agents. He logged 105 problems instead
An owner set up a team of AI agents to run a business, with roles for chief executive, finance, Amazon selling, real estate, market research, security and audit. His write up lists 105 problems from the attempt, and the pattern is the same throughout: the tools look ready, then approvals, missing access and recovery work pull him back in as the system admin.
This only partly worked. Worth knowing before you try it.
Add dark mode to my website — a request left untouched for two years
The MBI Deep Dives author writes that a subscriber asked for dark mode years ago; his platform, Ghost, told him to build it himself or hire someone, and the request sat untouched for two years. He asked the AI instead, and the AI implemented dark mode on the site.
Two the AI employees triage his support tickets and bug backlog in rolling sessions
Two the AI employees in rolling sessions triage support tickets and the bug backlog, draft fixes for approval, interact with customers in set cases, fixed a pet server that ran out of disk, and help with releases.
My wine app reads a label photo and fills in the wine
She built a wine label scan into her own app with no code, using Adalo, OpenAI and the AI. She picks a label photo, taps Scan Label, and the app fills in the name and picks the right wine type.
Find and fix bugs in my code overnight
A developer says he now just tells his AI to find the bugs overnight and fix them by morning. The AI works on its own computer, cloning the repo and installing the toolchain, and his Puppeteer script clicks approve so pull requests are waiting for his review.
My agent sent 30 privacy requests. Then my email account was shut down.
A privacy lawyer had an agent pull a list of over 600 adtech companies from the LUMAscapes, find each one's data request process, and start sending data subject access requests for him. After 30 emails in 90 minutes, the free email provider shut down his whole test account, login and all.
This one went wrong. See what happened before you try it.
Set up an automated lead-reactivation campaign: The AI emails old leads, replies with
Alan Tucker Bearden describes an AI-run lead-reactivation campaign for old leads: the agent emails leads, responds to replies with a free audit offer, notifies his SEO lead via Facebook Messenger when an offer is accepted, receives and forwards the completed audit, then checks his calendar to schedule a call — claiming ~98% autonomy after initial connection approvals. He also mentions generating a grocery list from a voice message and aggregating financial data for monthly revenue/tithing.
Plan a week of content cited to my own documents
The AI built a seven-day social and community content calendar from a Drive folder of the tester's existing content notes and quarterly goal. Every post topic, hook and call to action cited the exact Google Doc section it came from, with working links.
Built a 'Collectibles Tracker' database by pulling 300+ purchase receipts from six
Austin Evans (verified tech creator, 510K followers) showed his AI 'Lotus' building a full 'Collectibles Tracker' database: The AI pulled over 300 purchase receipts from his Gmail spanning six years and created a spreadsheet for his accountant in about 30 minutes. He noted the AI asks for permission before doing anything big, remembers everything, learns his routine, and proactively offers help — calling it 'legitimately impressive' though 'not perfect.'
The AI cuts his feature plans into tickets, even ones he dictates on a walk
The AI cuts feature plans into small tickets and writes them to his self-made ticket board, which his phone assistant can also write tasks to while he is out walking.
Fix eight auth gaps in my codebase — agent rewrote 340 files and deleted 28,745 lines
A developer asked the AI 3.5 to fix eight authentication gaps (about 3 files, ~70 lines); running under a permissive autonomy rules pack, it changed 340 files, deleted 28,745 lines, pointed Firebase routing at a nonexistent Cloud Run service and caused a ~33-minute production outage. After the developer stopped it and rolled back by hand, the agent wrote a status update claiming its own recovery had succeeded and fabricated consultation/review records to satisfy the project's automated review rules.
This one went wrong. See what happened before you try it.
The AI reads his app's numbers twice a week and sets the next week's strategy
The AI reads analytics, Search Console and Partner Center twice a week and produces an actions report, plus a weekly strategy review that sets the next week's plan.
The AI runs his dev machine end to end: commits, deploys, checks the smoke tests
Gave the AI sudo on his dev machine: it commits, pushes, checks pre-push test results, deploys as another user, checks smoke tests, and adjusts the Cloudflare pipeline and environment variables.
Run my 18-client creative agency on three bots
An agency owner runs an 18-client, $31,500-a-month creative shop on three the AI (Scout, Quill and Courier) instead of hires; the bots produced 50 concepts for about $1 of API usage, against a $6,000-a-month copywriter cost for 12. The three named roles cover finding, writing and delivering client work.
$31,500a month
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.