A conversation with Manus AI's cofounder and CPO Tao Zhang
TL;DR
Tao Zhang, cofounder and CPO of Manus AI, explains how their autonomous AI agent went viral by demonstrating executable outcomes rather than chat responses, and shares their unconventional product development approach where functional prototyping precedes design and prompts replace traditional interfaces.
🤖 Product Architecture 3 insights
Virtual sandbox gives AI autonomous 'hands'
Manus provides each AI session with its own cloud computer (virtual machine), allowing it to browse, write code, manage files, and execute tasks independently without local safety constraints or constant human permission.
Transparency builds trust through visibility
The interface displays every action the AI takes—opening browsers, writing Python, accessing files—so users can observe the process and build confidence in autonomous operations.
Asynchronous execution enables true autonomy
Because Manus runs entirely in the cloud, users can close their laptops after assigning tasks and receive notifications upon completion, eliminating the need to babysit the AI or repeatedly click 'accept' buttons.
⚙️ Development Philosophy 3 insights
Prompts replace interfaces as the new PRD
In AI products, conversation becomes the interface, making prompt engineering the equivalent of UI/UX design where product managers craft the experience through text rather than mockups.
Prototype-first workflow reverses traditional sequencing
The team uses Manus to build fully functional prototypes (complete with frontend, backend, and AI capabilities) before engaging engineers for productionization or designers for polish.
Targeting non-technical 'long tactical users'
Rather than competing for engineers who already have Cursor and Devin, Manus focuses on average users who lack technical backgrounds but need AI to execute complex, multi-step tasks autonomously.
🌍 Company Building 3 insights
Global expansion at 100 employees
Manus opened offices in Tokyo and San Francisco immediately after launch to engage face-to-face with international users and understand local business cultures rather than remaining a 'cloud-only' company.
Physical co-location drives innovation
Product managers, engineers, and researchers sit together in the same physical space to enable rapid experimentation through daily 'small talks,' avoiding the silos that slow down traditional departmental structures.
Authentic demo video generated 3.5M signups
The viral launch was a self-shot screen recording created in six days for $0, demonstrating real use cases like resume screening and coding rather than polished marketing, leading to 2 million waitlist signups in the first week.
Bottom Line
Build AI products by using the AI itself to create functional prototypes immediately, treat prompt engineering as your primary interface design discipline, and seat product, engineering, and research teams together to accelerate innovation through rapid iteration.
More from Stripe
View all
Nat Friedman and Daniel Gross in conversation with John and Patrick Collison
AI leaders Nat Friedman and Daniel Gross join Stripe's Collison brothers to discuss how we're in the 'slow' beginning of the singularity, where human bottlenecks still constrain model improvement but will soon give way to AI self-improvement, creating profound economic uncertainty and a new golden age of personal AI agents that fundamentally alter human-technology relationships.
Stripe Sessions 2026 | Indexing the economy
John Collison and Emily Sans present Stripe's economic data revealing a surge in AI-driven business dynamism, debunking myths about a K-shaped recovery while showing how solopreneurs scale faster than ever and commerce shifts toward autonomous agents.
Sam Altman in conversation with Patrick Collison
Sam Altman discusses the recent 'takeoff' moment in AI capabilities driven by coding models crossing subjective thresholds, while outlining OpenAI's evolution into a low-margin infrastructure provider and sharing untold stories from the secret eight-month period when GPT-4 existed only inside the company.
What comes after smartphones, with Evan Spiegel of Snap
Evan Spiegel describes Snap's pivotal 2026 'crucible moment' as the company approaches one billion users and profitability while launching Specs—AR glasses that aim to shift computing from isolating screens to shared, heads-up spatial experiences after 12 years of development.