Manus Blog

Cause it’s still easier and cheaper apparently This is the most recent example I found last week for a local barber: https://news.ycombinator.com/item?id=48166050 They seem to be using Manus: https://manus.im/ And my other assumption is that it immediately integrates with IG/Facebook which is where they do a lot of their marketing I see no reason that trend is going to slow, especially if you can go to meta to manage your entire business marketing. Regular people running business just want fast cheaps and good enough.

Un-paywalled content: https://archive.is/vIVsZ Introduction video: https://www.youtube.com/watch?v=K27diMbCsuw Website: https://manus.im

have watched it's use case https://manus.im/share/OZB4PvsXY5N5FrvLXqqCl4?replay=1, it's amazing for me. You just give a command, it can get information from the internet and analysis the information and then give a report document, draw system architecture diagram. if it has a body, "I Robot" maybe will come ture

Since I get downvoted because I guess people don’t believe me, I’m sitting at breakfast reading a book. I suddenly think about yaml streaming parsing, start a gpt research, dig a bit deeper into streaming parser approaches, and launch a deep research on streaming parsing which I will print out and read tomorrow at breakfast and go through by hand. I then take some of the gpt discussion and paste it into Manus, saying: “ Write a streaming go yaml parsers based on the tokenizer (probably use goccy yaml if there is no tokenizer in the standard yaml parser), and provide an event callback to the parser which can then be used to stream and print to the output. Make a series of test files and verify they are streamed properly.” This is the slot machine. It might work, it might be 50% jank, it might be entire jank. It’ll be a few thousand lines of code that I will skim and run. In the best case, it’s a great foundation to more properly work on. In the worst case it was an interesting experiment and I will learn something about either prompting Manus, or streaming parsing, or both. I certainly won’t dedicate my full code review attention to what was generated. Think of it more as a hyper specific google search returning stackoverflow posts that go into excruciating detail. https://chatgpt.com/share/68b98724-a8cc-8012-9bee-b9c4a77fe904 https://manus.im/share/kmsyzuoRHfn1FNjg5NWz17?replay=1

Vouched. I don't know. I'm guessing your plight falls into buyer beware category. Not sure what path you took,[1] [2], but it would seem one needs to carefully understand what they offer. Webapp links mentions how easy it is to link to your own domain, the second link manus-hosting site mentions it is not permanent hosting and subscription based. [1] https://manus.im/features/webapp [2] https://static.app/manus-hosting

Some redundancy also helps to keep a running todo list on the context tip, in the event of compacting or truncation. Distilled mini/nano models need regular reminders about their objectives. As documented by Manus https://manus.im/blog/Context-Engineering-for-AI-Agents-Lessons-from-B…

Manus post: https://manus.im/blog/manus-joins-meta-for-next-era-of-innovation (https://news.ycombinator.com/item?id=46426534)

Context editing is interesting because most agents work on the assumption that KV cache is the most important thing to optimise and are very hesitant to remove parts of the context during work. It also sometimes introduces hallucinations, because parts of the context are with the assumption that eg tool results are there, but theyre not. Example Manus [0]. Eg, read file A, make changes on A. Then prompt on some more changes. If you now remove the "read file A" tool results, not only you break the cache, but in my own agent implementations(on gpt 5 at least) can hallucinate now since my prompt etc all naturally point to the content of the tool still beeing there. Plus, the model got trained and RLed with a continuous context, except if they now tune it with messing with the context as well. https://manus.im/blog/Context-Engineering-for-AI-Agents-Lessons-from-B…

I ask Manus to write a script to demo my CLI tool and use asciinema to record it. It's perfect Whole process: https://www.pgschema.com/blog/demo-with-manus-and-asciinema Replay: https://manus.im/share/8fEln1OzxpnsRSU1PnHweG?replay=1

This format, or similar formats, seem to be the standard now, I was just reading the "Lessons from Building Manus"[1] post and they discuss the Hermes Format[2] which seems similar in terms of being pseudo-xml. My initial thought was how hacky the whole thing feels, but then the fact that it works and gives rise to complex behaviour (like coercing specific tool selection in the Manus post) is quite simple and elegant. Also as an aside, it is good that it appears that each standard tag is a single token in the OpenAI repo. [1] https://manus.im/blog/Context-Engineering-for-AI-Agents-Lessons-from-B… [2] https://github.com/NousResearch/Hermes-Function-Calling

You can see the replay of the AI agent working on a task here: https://manus.im/share/OZB4PvsXY5N5FrvLXqqCl4?replay=1

How do you folks think about the Manus finding on dynamic tool selection? https://manus.im/blog/Context-Engineering-for-AI-Agents-Lessons-from-B… > A natural reaction is to design a dynamic action space—perhaps loading tools on demand using something RAG-like. We tried that in Manus too. But our experiments suggest a clear rule: unless absolutely necessary, avoid dynamically adding or removing tools mid-iteration. There are two main reasons for this: > 1. In most LLMs, tool definitions live near the front of the context after serialization, typically before or after the system prompt. So any change will invalidate the KV-cache for all subsequent actions and observations. > 2. When previous actions and observations still refer to tools that are no longer defined in the current context, the model gets confused. Without constrained decoding, this often leads to schema violations or hallucinated actions. > To solve this while still improving action selection, Manus uses a context-aware state machine to manage tool availability. Rather than removing tools, it masks the token logits during decoding to prevent (or enforce) the selection of certain actions based on the current context.