Benjamin Guo argues that modern software leaves people renting access to fragmented services while their data stays in systems they do not control.
2
Zo is a personal cloud and server with AI built in, where people can host websites, files, APIs, automations, and agents in one place.
3
Guo says cloud agents raise an ownership question: whether a person's or company's use of an agent improves their own system, or mainly improves the provider's system.
Summary
Benjamin Guo presents Zo Computer as a response to software subscriptions, fragmented services, and cloud systems that keep users from owning their data and tools. He calls this arrangement technofeudalism: people pay SaaS providers, those providers pay cloud companies, and the money flows upward to infrastructure owners. Zo gives each person a personal server in the cloud with AI built in. Guo describes how Charlotte runs websites, invoices, bookkeeping, and scheduling from it, while Anthia replaced services such as Squarespace and Calendly with her own Zo setup. The demo shows file storage, model access, scheduled tasks, integrations, browser actions, hosted services, and a personal website edited through chat. Guo then applies the ownership question to cloud agents. If an agent belongs to a provider, improving its use may improve the provider's system. He wants individuals and companies to publish agents, own them, and benefit from their continued use.
Guo builds his own software while waiting to speak
Guo opens by showing a speaker directory he made immediately before the talk. The page contains links, profiles, websites, blogs, and research about the other speakers. He put the page behind a QR code on his first slide. He presents this as a normal use of Zo Computer, which he describes as his computer in the cloud. The example establishes his preferred workflow: he tells a cloud-based system what he wants in plain language, and the system creates a working website or tool without a separate deployment process.
The old computer ideal was a person and a machine working together
Guo uses Apple's Finder icon, designed by Susan Kare, to explain his view of good computer software. The white face represents the human and the blue face represents the machine, with both faces shown in harmony. He says people have always liked machines and should feel at home with them. In his view, AI tools should recover that feeling instead of placing people in front of a confusing collection of separate products.
Modern software makes people feel surrounded by rented services
Guo contrasts the early internet and first personal computers with the current experience of moving among applications, websites, and services. He says prices are rising, sometimes because of AI features users did not ask for. He also describes regular people as upset, confused, or afraid of AI. The problem is practical as well as emotional: tools fragment people's work, lock up their data, and make the computer feel less like a place that belongs to them.
Guo calls the upward flow of software payments technofeudalism
Guo describes a chain in which users pay rent to SaaS providers, SaaS providers pay cloud providers, and cloud providers pay the companies he associates with the industry's kings, such as Nvidia. He says users become dependent because their data is locked into separate services. Those services, in his account, are rewarded for monetizing attention, taking data, and improving their bottom line rather than improving the user's life. Zo's stated mission is to give people a real home on the internet.
A personal cloud combines a person's data, AI, and hosted services
Guo defines Zo as a personal cloud, meaning a person's own home in the cloud rather than an account spread across unrelated cloud services. Users can put their files and AI there, then host websites, APIs, and other services from the same place. His personal cloud includes Zo, local devices, an agent, websites, and APIs that he can reach from a laptop, Mac mini, or phone. He compares this with the older practice of uploading website files directly to a server and treating that server as the place where work happens.
Anthia replaced a SaaS stack with one self-hosted business system
Anthia, a non-technical free-diving instructor, previously used services such as Squarespace and Calendly. Guo says she canceled those subscriptions and now hosts retreat websites, a personal workspace, notes, accounting, files, and a database on Zo. When a prospective customer asks about a retreat, Zo texts Anthia the person's number. She calls immediately, asks Zo for a payment link, and closes the sale while the customer's interest is fresh. Guo says this process has helped her book more retreat revenue.
Zo puts storage, models, automations, browsing, and hosting in one workspace
In the demo, Guo shows a website created during the talk, access to different models including Claude Code, a file system with cloud storage, and chat-based work over files. Zo can run AI tasks on a schedule and includes integrations, a skills library, and a browser that can log into sites and take actions such as buying products. Users can host arbitrary HTTP or TCP services, edit a personal website through chat, and create custom tools such as Guo's Calendly replacement.
Cloud agents raise the question of who owns the accumulating intelligence
Guo expects people to interact increasingly with agents, while those agents interact with websites and APIs. Many agents will themselves run in the cloud, which makes ownership important. He uses Claude Code as an example of a company-level cloud agent whose context belongs to the company providing it. In his view, improving that setup primarily improves Anthropic, so the intelligence flows upward. He calls this intelligence feudalism and proposes agents that individuals and companies can publish, own, and improve through their end users.
You run a small business across several SaaS products and want to see how one personal server could combine websites, scheduling, bookkeeping, files, and customer follow-up.
You are building personal agents and want a concrete view of hosting them in a cloud environment with storage, APIs, automations, and root access.
You are thinking about agent ownership and want Guo's argument about whether cloud-based agent use improves your system or the provider's system.
