Raw LLM Responses

Respect for using Manus as the reference for agent work. I have been using them for years now, they are way ahead of the curve. Especially now with their own Agent Browser extension. Here is an example of a situation I used Manus for this week that just blew me away. I 3D printed a planetarium and wanted to put it together, but needed some hardware. So I tasked Manus with going to find all the parts I needed, added them to carts, checked out and prompted when I needed to put in payment details. Now this is nothing too new, RPA and Software Automation like Playwright and Beautiful Soup have been doing this for a while now, and I have deployed RPA solutions and BPMS solutions at scale, but they are often prone to layout shift, which causes errors, and even perfect executions can be caught out by a small change in browser reaction speed. Manus found situations where it couldn't find the exact product I needed, so it looked elsewhere to find it and suggested it as an alternative for me, or it would notice that it couldn't add products to the cart - realised it was in the trade portion of the bunnings website, then logged into the non trade version of the website and added them to the cart. I watched this happen in real time, and it was wild. But after 30 minutes, I had 3 carts with exactly what I needed, I put in my card details myself, then let it go. Products all arrived within a day. I could have done this myself of course, or I could have gone to the stores myself, but that costs much more in wasted time than the tokens I used for this task.