This Week in Startups: How These 3 Founders Are Building on Open Claw | E2248
Date: February 12, 2026
Host: Jason Calacanis
Guests: Presh Dinesh Kumar (Wellness Company), Sean Liu (Cobot/OpenClaw Integration), Vishnu (Agent37)
Episode Overview
This episode delves into the world of agentic AI platforms, focusing on OpenClaw, a cutting-edge open-source system that enables advanced automation, replication of work roles, and seamless integration into startup workflows. Jason interviews three founders—Presh Dinesh Kumar, Sean Liu, and Vishnu—who are already harnessing OpenClaw to accelerate product development, automate complex business processes, and build the next generation of AI-first startups. The discussion covers real-world use cases, product demos, technical insights, and implications for startup economics and fundraising.
Key Discussion Points and Insights
1. OpenClaw and the Agentic Platform Revolution
- What is OpenClaw?
OpenClaw is an open-source, agentic AI platform (formerly “Claudebot”) that enables users to create “Personas,” “replicants,” and autonomous agents to automate business and personal tasks. - Why is this a big deal?
Unlike sandboxed LLMs, OpenClaw agents can interface with multiple tools and APIs, act independently, and execute multi-step tasks, making these AIs “team members” rather than chatbots.
“We basically took AI and we put it in a little closet. Now you let the AI out of the dungeon and you say, here's the keys to the kingdom.” — Jason (06:07)
2. Founder Use Cases: Presh Dinesh Kumar and the Power of Automation
Automating Operations and Scaling Experimentation
- Presh shares how, thanks to AI tools (culminating in OpenClaw), his wellness company shipped three to four iOS consumer apps in a year—at speeds unheard-of previously (00:00–02:56).
- Automation through agentic AIs enables startups to prove multiple theses simultaneously, run more experiments, and reduce required headcount drastically.
- "A startup is just a handful of experiments... now you can prove out multiple thesis at the same time because of these agentic tools." — Presh (00:23)
Real Examples:
- AI Assistant “AWA”:
- Automatically compiles and shares use case documentation, connects email and messaging apps, and executes multi-step business processes such as gathering analytics, emailing users, and preparing due diligence docs.
- "It’s like having an assistant. It’s like having a chief of staff." — Jason (04:35)
- Connecting Product Analytics (Posthog):
- AWA retrieves active user lists, processes and exports data, and emails selected users for references in diligence rounds—tasks “done in about five minutes.”
- Agent Mail & Security:
- AWA has its own work email and only executes tasks after owner approval—introduces a security layer against prompt injection and spoofing (07:57–10:12).
- Cron Job Content Generation:
- AWA surfaces research, drafts blog posts, and (potentially) generates social images for distribution (10:44–14:55).
- “Who wants to do that? Why not have a replicant do it and check their work?” — Jason (14:15)
Impact on Costs and Team Size
- Jason and Presh agree: What once required 10 people and a $1M/year burn may now need just 3–4 people and a third of that budget.
- “That job from five years ago is now going to be replaced... with an OpenClaw replicant.” — Jason (07:21)
- “You can start to just automate out rules or... tasks. That number is just increasing every month.” — Presh (20:10)
3. Product Demos
Sean Liu’s Integrated Smart Glasses Agent
- Live Demo: Sean, wearing Meta Ray-Ban smart glasses, streams a live use-case:
- Looks at a product, verbally asks Gemini (via OpenClaw) to add it to his Amazon cart, and the agent executes the shopping process seamlessly (23:05–25:07).
- “What I’m building... is integrating the visual understanding to the OpenClaw.” — Sean (24:06)
- Sean has created a two-layer system: Gemini Life handles live perception (audio, video) and OpenClaw delegates task-execution.
- "You could probably walk through [Costco]... just talk out loud, stream this, and manage the store." — Jason (28:10)
- The biggest personal use? “Adding things from the fridge to my Amazon cart, hands-free.” (29:50)
Vishnu’s Hosted OpenClaw (Agent37.com)
- Motivation: Wanted to run OpenClaw securely and affordably; realized a VM in the cloud is safest, but setting up is still non-trivial for most (36:42–40:12).
- Aha Moment:
- “I was working out in the gym... wanted to use my voice. [The agent] researched, set itself up, and handled audio I/O.” (38:58–39:21)
- Providing Access to Friends, Now a Business:
- Shared OpenClaw agent as a service for friends; now Agent37.com, offering hosted agents for $1.25-$3.99/mo, growing daily.
- Jason Invests On-Air:
- Jason offers both Vishnu and Sean $125k for 7% equity and accelerator admission, sealing the deals live.
"I just spent $250k. You're both in the accelerator. This is how we do it, folks." — Jason (43:46)
- Jason offers both Vishnu and Sean $125k for 7% equity and accelerator admission, sealing the deals live.
Notable Quotes & Memorable Moments
- On the Scale of Automation:
“What’s the percent you trust your agent to get it to? 80%? 90%?”
“Yeah, 80, sometimes even 90. I’ll do final approval or edits and then send it out.” — Presh (14:55) - On Agent Integration and Security:
“AWA will never perform a task without confirming with me through Telegram first.” — Presh (09:12) “It’s basically rooted your machine.” — Jason, cautioning on running agents locally (37:26) - On Startup Economics:
“You could skip venture capital completely... That is what we’re going to see more of. My best advice to founders is to skip rounds.” — Jason (35:42) - On Real-World Implications:
“This makes you the Terminator, it makes you RoboCop.” — Jason, on real-time AI augmentation (32:14)
Timestamps for Important Segments
- [00:00–02:56]: Introduction, Presh on AI’s acceleration of shipping products
- [03:38–04:40]: Presh explains examples of AWA agent in use
- [04:41–07:03]: Automating complex business processes with agents
- [08:09–10:12]: Email integration & security for agents; risks & mitigations
- [10:44–14:55]: Cron jobs for research monitoring, content/blog auto-generation
- [16:14–17:52]: Customer support workflow—from email to Slack, bug fix automation
- [19:03–20:21]: The shrinking cost and headcount for building startups
- [23:05–25:07]: Sean’s live smart glasses demo
- [29:50–31:42]: Sean’s examples, from grocery shopping to perfume formula tracking
- [36:42–40:47]: Vishnu’s hosting experience and Agent37.com
- [42:07–43:44]: Jason offers/invests $125k each in Sean and Vishnu
Tone and Takeaways
Jason’s upbeat, future-forward style frames the discussion: dramatic drops in startup costs, the rise of autonomous agents, and the dawn of AI as a true team member. The founders’ stories illustrate practical, immediate impact, while technical caveats around security, control, and infrastructure remain front-of-mind.
Final Insights
OpenClaw and agentic AI platforms are fundamentally changing how startups operate—enabling solo founders or lean teams to experiment, build, and scale at speeds and costs previously reserved for large, well-funded companies. The episode closes with Jason proactively investing on-air in Sean and Vishnu, underscoring the breakneck, opportunity-rich pace of the current AI founder wave.
For full demonstrations, detailed technical discussion, and Jason’s signature energy, listen to the complete episode.
