Keep pulling the thread on Ryan Carson.
According to Ryan Carson, the traditional startup advice of minimizing system and documentation work for an MVP has reversed, and founders must now invest heavily in these areas upfront to leverage AI agents.
Ryan Carson has raised a $2 million seed round for his startup but does not plan to hire any human employees for a while, opting to use AI agents instead.
Ryan Carson had his Devin agent build a custom command-line interface (CLI) for the Google Ads API, allowing him to manage and optimize ad campaigns through conversation.
Ryan Carson believes AI agents are fundamentally composed of cron jobs and markdown files.
Ryan Carson states that it is significantly easier and faster to onboard and train AI agents compared to human employees.
Ryan Carson built a system called ClawChief, consisting of Markdown files and cron jobs, to configure his OpenClaw instance as an executive assistant named R2.
Ryan Carson's OpenClaw agent, R2, performs a sweep of his inbox, calendar, and priorities every 15 minutes and sends him updates via Slack.
Ryan Carson's OpenClaw agent, R2, conducts daily business development outreach by researching contacts, adding them to a spreadsheet, and sending emails on his behalf.
Ryan Carson estimates he receives $2,000 to $3,000 worth of tokens for his $200 per month ChatGPT Pro subscription, believing OpenAI is currently subsidizing token costs.
Ryan Carson uses a secondary agent within VS Code, such as Codex, to monitor and fix his primary OpenClaw agent when it breaks, as OpenClaw often cannot repair itself.
Ryan Carson has found that OpenClaw's native "heartbeat" feature is unreliable for time-triggered events and recommends using cron jobs for deterministic execution.
Ryan Carson predicts that companies like Evry, which offer "OpenClaw in the cloud" services, will become billion-dollar businesses by providing on-demand digital employees.