“While some techies use the buzzy AI platform OpenClaw to help book flights or summarize news, one Silicon Valley startup tapped it to stand up a nearly fully autonomous software-engineering team.
Vinay Pinnaka, co-founder and chief technology officer of Mountain View, Calif.-based JustPaid, used a combination of OpenClaw and Anthropic's AI coding tool, Claude Code, to create a team of seven AI agents to grind out code 24/7.
In the month since Pinnaka put his AI team to work, they have turbocharged workflow. The agents have built 10 major features, each of which would have taken Pinnaka's human developers a month or more to build. Recently, the JustPaid team hired a new human developer who was trained almost entirely by the AI agent engineers.
Meanwhile JustPaid's humans are being deployed to work on high-priority tasks like customer requests, Pinnaka said. For now. "Once [AI] gets to the stage where it is able to handle human empathy, I would be able to say, “I can replace everyone with AI, a 100 %" he said.
The company, which makes an AI-powered platform that automates financial operations, was founded in 2023 and has nine employees.
Though AI-assisted coding has been around for a while, new technology is making it possible to autonomously create apps -- and automate nearly all aspects of software development. While the practice is quickly spreading across Silicon Valley, there are still plenty of concerns about cybersecurity and larger, existential questions about the future of human-based software engineering.
Pinnaka says the creation of his AI agent engineering team followed the natural evolution of Claude Code, which was dramatically improved when Anthropic released the Opus 4.6 model in early February. Then, as the OpenClaw craze began to sweep Silicon Valley and beyond, the idea popped into his head that he could build an "AI version of an engineer."
OpenClaw is an open-source AI agent orchestration system, and "claws" are autonomous agents that can plan and execute tasks on their own, and, critically, spin up their own subagents to tackle specialized tasks, access files and themselves delegate tasks to other subagents.
For Pinnaka, OpenClaw functions as "the brain that decides what needs to happen," and Claude Code is "the hands that do the coding work."
Each of his seven AI agents has its own identity and tasks, from writing code to reviewing it and conducting quality assurance.
Pinnaka isn't alone. Kuse, a startup that develops "AI co-workers" on behalf of companies, built its own AI agent employees using OpenClaw. The company said its AI agent employees collaborate with each other and alongside human employees. The agents also have their own identity -- they send Slack and Gmail messages, speak in Zoom meetings and proactively start work. "We didn't remove anyone or stop hiring, but we have fewer reporting lines and fewer meetings," the company said.
Other startups have shared in Silicon Valley's enthusiasm for OpenClaw and independent AI bots, but only up to a certain point.
Tatyana Mamut, co-founder and chief executive of Wayfound, a startup that helps businesses monitor their AI agents, said she is experimenting with the platform, but only in a strictly controlled, separate environment where it doesn't access any business data.
"OpenClaw and other agents that are left to their own devices to make decisions need to be supervised all the time," Mamut said. For large businesses, it's still considered too risky to deploy AI agent platforms like OpenClaw. That's why companies like Nvidia have released their own tools that aim to make them safer for enterprises. For OpenClaw to work as a true personal assistant, it must have access to all of a user's data and systems. When agents go rogue, they can tamper with or delete valuable files.
---
Belle Lin writes for WSJ Leadership Institute's CIO Journal.” [1]
1. Startup Taps AI for Autonomous Software Team. Lin, Belle. Wall Street Journal, Eastern edition; New York, N.Y.. 01 Apr 2026: B5.
Komentarų nėra:
Rašyti komentarą