Transcription
What other AI stops at generating ideas? Manus delivers results. We see it as the next paradigm of human-machine collaboration and potentially a glimpse into AGI.
That was co-founder and chief scientist of Menace II, which has gone viral on social media, on claims that it has launched the world's first so-called general agent. Let's get a little bit more about what some are calling China's next deep-sea moment with our tech reporter, Annabel Jewellers.
Annabel, tell us who is behind Manus. Why is it causing such a stir?
Yeah, well, maybe we can start with why it's causing such a stir because it was only released just a couple of days ago at the end of last week. And and the access right now is really limited. So it's still in its beta phase. And so you can access it only by invitation as a discord channel that you can follow as well. But what we know so far is that Manus works like a general AI agent. So it's important to to make the distinction here between what Manus can do and a sort of everyday chatbot, because a chatbot like ChatGPT, for instance, is human-led. You put in the text prompts and then it gives you the answer. This is about taking it to a different level where essentially you have what's called Agenda AI, and and which was seen as one of the big trends in AI for 2025. That agenda is when the AI can act autonomously and it's able to make decisions based on the information that's available to it.
So Manus, on the website, lays out a lot of different use cases for this, like analysing financial statements, screening job applicants, even deciding or proposing where you should live based on your setting requirements in terms of how much you can spend, the kinds of facilities that you're looking for. It's basically, it seems quite limitless at this point in time. And that is also why, again, it's causing such a hype online as well as really where it's coming from. And we understand that it's made by a group of researchers in China. And so it again points to that really stiff competition between the US and China, who can be in the lead on AI, especially when Manus is saying it can outperform OpenAI's own advanced agent, Deep Research. Such a deep bench in China, I think, is one of the aspects.
I think what do we know right now about the limitations or the limitations? I think generally go back to the access because as I said, it's in its beta phase. So right now you can only access it by invitation. So you can submit a request on Manus and give a reason for it. So I've tried as a Bloomberg journalist to get access to it, but we we do have a Discord channel as well. And they're putting out some invitation codes on that. It's jumped to around 150,000 followers. So again, the the the emphasis on that. But there's a lot of different glitches that people are pointing out online, things like Manus crashing, giving incorrect information. But as I said, it's still in that beta phase. It's not fully open to the public just yet. We still need to get more details. How exactly has this technology been developed?