How Manus AI found Claude’s weakness and made it a trillion times better
I agree I sound a bit insane and dramatic here, but let’s dive into how Manus AI’s launch decided to wreck the internet — and yes, it’s like Deepseek all over again
Look at how clean the UI is. A very nice first touch on the landing page and not too much on the technicalities. Just straight up to results.
oh and also, even their intro video script is one of the use-case, never fails to surprise me tbh.
Multiple real workflows
No blatant surface-level workflow like what other GPTs claim. Straight up offering use-cases, they even offer the replay capability. Sounds like a very simple capability but they managed to impress the user on their first launch.
Not to mention, insane amount of use cases.
They managed to pull out the ‘Public, but still private’ card by asking users to join the waitlist, making the wait even more worth it just by seeing the use case gallery.
Let’s get deeper on how Manus AI is different
From a layman’s perspective, I primarily use ChatGPT for its cleaner look and easier to read for a long time if I decide to grind for my tasks. When I find myself stuck with ChatGPT not being able to give me what I want, I will switch to Claude, then go back to ChatGPT. Given the short amount of attention I paid to Claude, one of the most repulsive parts of Claude AI is its limited usage, as well as its serif display.
Get 데보라.’s stories in your inbox
Join Medium for free to get updates from this writer.
Don’t get me wrong, Claude’s artifact works insanely great. This is proven by how ChatGPT started to incorporate artifacts, but Claude remains original.
However, back again, Claude remains with its serif look, making it harder to stay on that platform for a long time. Their multiple artifact requires scrolling which makes it harder to navigate.
And these points I mentioned above, are exactly what Manus AI is solving. It lets you browse the whole artifacts/docs in one line, including the input files instead of forcing the user to go back and forth just to refer back to the original file. With their sleek-looking design which is similar to what we have on ChatGPT already, it’s intuitive enough for the user to onboard, implementing Jakob’s Law (UX).
Manus AI does sound more human, compared to other GPTs. Maybe there is a separate fine tune to make them sound human, but the answer quality they offered (just from seeing the use cases) is giving the answer I expect from ChatGPT that requires me 3–4 continual prompts.
Which made me wonder,
What model do they use?
Let’s put our thinking cap on, there is no official statement on what model they use. But they are founded by Monica.im, which deeply integrates Claude 3.5 & Deepseek.
iykyk, I’ll stop here.
Final Note
As someone working in an AI-heavily influenced industry, I’m excited about what Manus AI can offer further and genuinely impressed with how AI has come so far.
