Meta dropped $2 billion on Manus and closed the deal in 10 days.

They have unlimited resources, and have previously built Llama.

But instead of entering the agentic space by building a tool in-house, they paid $2 billion for Manus.

I spent some time testing it to understand what they saw. But before that here’s what happened in AI this week:

1. Manus Meeting Minutes

The feature uses intelligent speaker recognition to identify participants and assign action items with attribution accuracy, while supporting offline recording that continues even when internet drops.

2. Pane

Pane's AI agent writes real formulas, edits cells directly, and builds charts step-by-step in the grid with human-level precision. The tool supports CSV, Excel, and PDF imports with automatic table extraction, and includes full formula support.

3. Unloop

Users draw out triggers, thoughts, emotions, and behaviors on a visual canvas to identify repetitive behavioral loops, with AI surfacing hidden patterns and connections they might have missed.

With ChatGPT and Claude, it’s an interaction and you have to be involved in the process. With Manus, that’s not the case.

If I ask for something, ChatGPT and Claude give me an answer. Manus opens browsers, clicks through sites, writes code, builds files, and delivers outputs I can use immediately.

The tech

I use it the most for structured outputs like spreadsheets, presentations, and reports.

I tested Manus against ChatGPT and some manual work on three tasks.

Competitor Research

What I asked for: Research 3 AI agent tools. For each: what it does, pricing, and one key limitation. Output as a table.

What I got:

What I had to fix: One pricing detail was outdated by a month. I now verify all time-sensitive data before using it.

When I use this: Before client meetings, when I need competitive intel fast, I can scan 3 competitors in 5 minutes instead of losing my morning to research.

Research to Presentation

What I asked for: Take the competitor research and create a 3-slide presentation: overview, comparison, and key insights.

What I got:

I now do research briefs at 2 pm and show up to 3 pm meetings with polished slides.

Bulk Research with Wide Research

What I asked for: Research 12 SaaS tools in project management. For each: pricing, target user, key feature. Output as a spreadsheet.

What I got:

When I use this: Market scans or bulk competitive analysis. I can now research 50 potential partners in an afternoon.

What Manus Can't Do (Yet)

Few days into testing, I finally understood the acquisition.

Meta could have built this, but they would have spent a year guessing what outputs people need, which tools to integrate, and when tasks should stop.

Manus already knew from two years and 80 million tasks. They learned by watching people use it wrong, then fixing it.

Meta acquired that operational knowledge.

If you need a personal intern who never sleeps and costs less than a coffee, try Manus for a week and reply with how it saved you hours of work.

Until next time,
Vaibhav 🤝🏻

If you read till here, you might find this interesting

#AD 1

Discover what drives conversions for your competitors with Gethookd. Access 38M+ proven Facebook ads and use AI to create high-performing campaigns in minutes — not days.

Start your free trial

#AD 2

Stop losing nuance when you type prompts. Wispr Flow captures your spoken reasoning, removes filler, and formats it into a clear prompt that keeps examples, constraints, and tone intact. Drop that prompt into your AI tool and get fewer follow-up prompts and cleaner results. Works across your apps on Mac, Windows, and iPhone. Try Wispr Flow for AI to upgrade your inputs and save time.

Try Wispr Flow