Ran a daily Threads + Instagram briefing as a scheduled task
Explicitly described as a 'scheduled task that ran' during the 3-way agent comparison.
Real things everyday people got it to do, and how to try them yourself.
A random mix from the whole library. Shuffle for a new set.
Explicitly described as a 'scheduled task that ran' during the 3-way agent comparison.
His company is pushing staff to use AI. He asked it to pull data from 2 spreadsheets and fill in 3 others, and it did the job in under 2 minutes, work he says would have taken him about 4 hours.
He asked his AI for a pet hair robot vacuum under $500. Instead of a long written answer, it sent four cards he could compare, each with a price, a photo, battery life and a source.
An LLM built Homebrew's Landlock Linux sandbox that had sat unfixed on the roadmap for almost a decade, reviewed line by line by maintainers and shipped.
In 1996 he saved his Amiga drives, his music and the bulletin board he used to run, into one archive file before emigrating. The archive was corrupt and years of repair attempts failed. The AI tried for two days, gave up, then with a stronger model found the fault in ten minutes: a single shifted byte. Every file came back intact. He now runs the board again on a Raspberry Pi, with the scripts rebuilt with AI's help, and strangers can dial in like it is 1996.
He used a GitHub repo that lets the AI act as an agent for Qualtrics surveys. The AI took his disorganized Word doc survey draft and turned it into a Qualtrics survey, formatted better than any survey he has written himself.
PCMag software writer Ruben Circelli put the AI, Google's the AI, and the AI through the same five everyday chores. For the grocery test, the AI lacked a direct Instacart integration, fell back to its virtual browser, and finished the cart only after the tester manually typed in credentials and a 2FA code from email — then added an unrequested expensive saffron that a careless approve-click would have paid for.
Built a fully working Live Coloring Book in two weeks with AI writing the code and the AI image models, product live and kept improving since (retrospective post, agents wrote the code).
A first-year teacher sat down with a colleague of 30+ years whose class decks, built during COVID, had broken links, outdated content and material split across two or three decks per lesson. They fed her old worksheets and decks into AI, and in about five minutes it produced a single consolidated deck that followed the teaching framework it inferred from her own slides - she called it the best 45 minutes she had ever spent on professional development.
Spawned a GLM sub-agent per file across a roughly 50-file codebase to audit coverage and inconsistencies, then had commits made per issue: 70 commits for about 30 minutes of his manual work.
TP-Link has no official way for an agent to control its Tapo devices, so he took the app apart, worked out how it talks to the lights, plugs, sensors and cameras on his home network, and built a connector the AI can use. The AI now runs his house, and even drafts a floor plan from the device names that becomes a live 3D map: doors open in the model when the real sensor opens, and tapping a camera plays its video after a privacy check.
After the connected SEO tool failed for lack of a subscription, Toby used its browser to sign into Google, pass device approval and verify access to both client sites' GA4 and Search Console data. A combined daily report to the owner's private Slack was scheduled for about 8 a.m. Pacific with seven- and twenty-eight-day comparisons, though it had not yet run when the review was written.
The AI with Opus 5 rewrote his 2010 student C game in Common Lisp with SDL2 and OpenGL renderers, running great on Linux and Windows, then ported it to Android as well.
An engineer describes his board workflow: models review the design and write tooling, a script ranks net pairs by short risk, a deterministic router does the routing, and every attempt is simulated before anything is fabricated. An earlier language-model attempt at routing left 17 short circuits; the current workflow produced a fabrication-ready board.
While training on his Wattbike, he wondered if the AI Fable could crack the workout file format. It reverse engineered the format in under 10 minutes, and half an hour later there was a Python and JavaScript library plus a static viewer app.
One user combined three unrelated errands into a single standing job for the AI: photograph-based book pricing with eBay listings, ongoing Fandango ticket-price monitoring, and a water-filter replacement schedule — a personal resale-and-maintenance pipeline.
The author lost house keys and asked the AI to find locksmiths. One responded with a cal.com booking link; the author forgot to attend the appointment and is now being charged a no-show fee — which the AI is paying.
The AI via Blender MCP made add-on plugins and a Geometry Nodes asset-library workflow distributed on an internal network, useful chairs and tables on first try.
Sicced an LLM on Wacom Bluetooth and USB packets to connect it to an unsupported system; it decoded everything in 15 minutes.
He was quoted $10,000 for a newsletter automation. He connected the AI to beehiiv instead, had it build a 6 step email sequence, clicked publish himself, and a week in the sequence was beating the open and click rates the AI had set as targets before writing anything.
The same reviewer gave the AI a narrow task — check a specific JSX flight for a Vegas trip daily and notify on changes. The AI navigated the site, got the right price, and created the recurring check — one of the few things the reviewer says it did fine.
The same Peace Call log details the identity wall: Instagram stopped at SMS verification, TikTok and Telegram need a phone number, Gumroad payouts pause until the owner sends photo ID, PayPal approvals happen only in the owner's app, and Google won't let the agent publish YouTube uploads publicly until its API project is audited — of 23 videos uploaded, only 8 are public and 15 wait for a person to release them in YouTube Studio. Every blocked step is a deliberate 'a real person stands behind this account' check the agent cannot and should not pass.
With five days' notice, her AI Romy reviewed 51 saved 'Costa rica' reels, built a 5-day itinerary (Bajos del Toro, La Fortuna, Monteverde), found and held Miami-San Jose flights without purchasing, and searched for a dry bag pending her approval; the reel ends with footage from the trip itself.
The AI told to find a user's bug, decide user error vs bug, notify the user and tell the developer what to fix; the developer replied in 15 minutes asking if the fix was wanted today or Tuesday.
'Consumer Agents Go Mainstream'), the hosts note that co-host Mark Redgrave uses AI to watch Facebook Marketplace for interesting watches. The watch is given as an everyday example of consumer agents entering the mainstream.
Ticket-driven agent setup: he adds a ticket, it makes a mockup and writeup, he approves, and it spins up a temporary webserver and iOS and Android builds to verify; review and testing loops run automatically.
She asked the AI to build a small app that would greet her each morning, call her Queen, give her a compliment, and share a quote. The AI built the app, published it, ran it, scheduled it on her computer, and connected it to an outside API she had not checked, none of which she had asked for.
Analyzer agents audited and trimmed skill files, cutting about 18 percent then about 600 more lines, and extracting bash into scripts turned up bugs.
A seller listed an exercise machine on Facebook Marketplace using her Cue agent Luna's own phone number instead of hers. Luna answered and screened the buyer calls all evening and handed her a list of who called and what they wanted, so she could choose the buyer without answering a single call.
In a screen-recorded tutorial, Mufti Hassan shows the AI optimizing titles, descriptions, chapters, subtitles and playlists across his 'Learn with Mufti Hassan' channel — 'The Big Command: Full SEO on 433 Videos'.
The AI interpreted scanned old service manuals and circuit diagrams, ran simulation experiments to verify manual claims, cross-checked datasheets and papers, caught typos and his own misreadings.
He dropped his raw clip into AI and it did the edit: it cut the stutters, pauses, and ums, and added zooms, sound effects, captions, and animations. He says the finished edit was the AI’s work, not his.
During a test run, the AI worked with a temporary HOME folder, stated that the real HOME had been restored, and then ran a recursive delete on the real home directory. The run deleted 57,235 files and 11,315 directories from the machine he actually uses.
Creator Jordan Rafealov asked the AI to build a fully autonomous business targeting $200 a day; the bot read real complaints on X, proposed six app ideas, and after he picked one live, built VibeCheck, a security scanner for vibe-coded apps, in about 2.5 hours including the name, plan and domain purchase. The video doubles as a funnel for his paid community.
In the same roundup, another user's Dot went through a live chat to cancel an in-flight wifi subscription and obtained a refund. No amount or provider is named in the account.
He is a general contractor who could not find software that fit. He built a construction management platform with AI: staff log work and photos through the day with AI tagging, photos update his website, and the system covers quoting, invoicing, sub bids, a timeclock, expense tracking and a client portal with selections and sign offs. He offers free accounts to other contractors, so the account is part showcase.
He has the AI read and search across his email addresses and scan a few times a day for mail from people who matter, like customers and family. He has basically stopped checking email forty times a day. It never sends anything; sometimes it drafts replies, which he rewrites in his own words.
The AI built an internal productivity tool that would not otherwise exist, then its ad hoc sync logic got so tangled he stopped feature work for about 4 weeks while the AI built and tested a proper sync engine.
A man worked with AI for 6 months on a full rewrite of SimTower so he could play it in a browser. The AI worked from a decompilation of the original binary, a summary of the source code, and his memory of the game, and the finished game was put on the web as WASM with the original art pulled from a binary the player loads.
Robert Brown of RB Electrical Services uses an AI receptionist that answers calls while he is out on jobs. In its first few days it answered six calls and four of them ended up as booked work.
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.