Carrying out an online task

The agent detected that Alaska had moved my flight 90 minutes earlier, a fact I'd completely missed, which qualified me for a full refund. It canceled my ticket, saving me roughly $550, and rebooked my one-way return to San Francisco. I don't know about you guys, but I'm starting to feel the AGI.

My first experience with Muse was on Monday. I needed to book a flight for myself and my wife. I found it myself on Google Flights, then, for journalistic reasons, asked Muse to take a crack at the booking. I put my airline login credentials in Muse's vault, but Muse met a terrifying obstacle, a captcha page, and asked me if I should take over or if it should attempt to pass the test asking if it was a human. Authorized to try, Muse failed; I took over and logged in, then passed the sticks back to the agent. It got kicked out of the browser a few minutes later. The process took 30 minutes and achieved nothing. I then booked the flights myself at a cost of five minutes and just over $500.

My first attempt was asking it to shop for New Balance 9060s. … The experience was not great. Muse struggled to find the right colorway and seemed to get stuck during browser use. … [On IMAX tickets] Muse then opened its browser view … navigated to the theater website, selected the tickets, and proceeded to guest checkout. … It arrived at the payment screen and prompted me to use Stripe Link.

Grok Bot turns X Bookmarks into working demos—launching Cursor Agents, validating results with screenshots and video, and deploying each project automatically. 🤖 The entire discovery-to-demo workflow is automated. https://t.co/lwj4VY9lql

I have something all men would appreciate and that's having your agent call, email, and set up your doctor and dentist appointments lol. I wanted to try it with @Muse as a test run. It emailed the office, called and left a voicemail. They got back to me to confirm and that was it. Used @bot to fill out all of the new patient paperwork since it has the context needed. It then added the dates to my calendar with all of the info it grabbed from the email. 5 minutes tops. And moving forward this will be the flow.

I used Muse today to apply for visa. If you have never applied for visa, congrats, bc visa application is one of the worst things in human bureaucracy machine. anyways, i got Muse to: - book a multi-legged refundable flight - book a refundable hotel - get my bank statement to prove i can pay - dig through my emails to find all my papers (passport photo, i94, etc.) - write an itinerary plan for what i would be doing each day - fill out the form from the website which is pretty good! but some problems i bumped into: - i needed a passport photo, so I used chatgpt to change a selfie into a passport photo. Muse could upload it, but you needed to size the photo to fit in the frame and Muse couldn't do it, so I had to do it myself. - Muse couldn't upload the rest of the files, so i had to upload everything. - while i uploaded, i realized 1 of the document field could only be JPG (not PDF! why did it not accept pdf??) so i asked muse to convert the PDF to JPG. also they didn't accept images > 2MB so also had to ask Muse to compress all my photos. i wish it just noticed that and fixed it itself. - i also realized there was a bunch of details Muse was missing (e.g. there were more docs that we needed, extra flights to book etc. to follow the instruction) - i realized i needed to book an extra flight, and it has strong enough bot detection that Muse couldn't do it itself :( so i had to book it manually. in the end, Muse did ~70% of the work, and i still had to spend ~1 hour on it. but it's like 1 hour of annoyance (i bumped into issue, i had to fix/ delegate to Muse, bumped into another issue, etc.) so i'm still pretty grumpy by the end. i'm grateful; but also if someone can build a visa application bench & train computer use agent on this successfully that person can have all of my money.

I downloaded the tool this week to test it out for myself. Using the bot feels a lot like texting an actual person. It uses emojis such as 🫡 and 👀to react to requests as it receives them, and the infamous three dots appear when it's writing a response. Sometimes, the bot even leaves you on read while it works. (Compare that with ChatGPT, which is always vomiting up walls of text and asking goading follow-up questions.) When I gave my agent a $25 budget and asked it to surprise my editor with a book, the bot read his previous writing (at my instruction), found a book he might enjoy— Material World by Ed Conway—and sent an email alerting him that it was ready for pick-up at his local Barnes & Noble. The bot also helped me find an Airbnb for an upcoming vacation, track down new in-network primary-care providers, and reach out to cleaning companies to get quotes for a friend.

A colleague of mine who has also been testing out the bot used it to book tickets to a museum in Mexico City. The museum was sold out for the dates he wanted, but the bot dutifully monitored the portal for three days and eventually snagged tickets when spots freed up. After the bot paid for the tickets, it rearranged his itinerary—swapping around restaurant reservations—to account for the new schedule.