Bo Brainerd

MCPs are like Harry Potter’s Sorting Hat. The person wearing it is the LLM/Agent Don’t force capabilities on the agent if you don’t have to. Don’t spend months reconfiguring the agent. Just build a sorting hat for each occasion/action, make it sort, and execute a result.

Yes yes yes!

I’m exhausted thinking about how much AI execution there needs to be. That’s why I’m so happy I chose human execution. An agent being told to “confirm”, “reserve”, or “book” an experience with user information is WAY easier than making it come up with 50 essay topics, write it, then post it. Tokens. Tokens. TOKENS. That’s what you’re spending each month. Everybody’s sleeping on MCPs tbh. If you don’t have an LLM in your tech stack, then use our API.

But then do you still have to create agents to market the telehealth service? It’s a business that’s eating away tokens and monthly burn. Would love to see the financials of that, I can only imagine the cost to market & maintain vs the monthly subscription revenue. Returns must have to be high.

Hot take: the Medvi story has nothing to do with AI. A guy built a $1.8B company with 2 people. Cool. But strip away the AI headline and look at what's underneath: Telehealth subscriptions. I've been in this space. I've seen the numbers up close. The LTVs are borderline unfair compared to every other business model. Traditional ecom: $50 AOV, maybe $80 LTV if you're lucky, 20% returns, warehouse headaches, margin compression. Telehealth: $200+/mo, $1,200-2,400 LTV, pharmacy ships direct, no inventory, no returns, recurring revenue. AI made him efficient. Telehealth made him a billionaire. Those are two very different things. And the craziest part? He only did weight loss. ONE vertical. ED alone is a $5B+ market. Hair loss. Hormones. Peptides. Anti-aging. Skincare. Mental health. Each one of these is a billion dollar telehealth company waiting to be built. The reason more people haven't done it is the infrastructure is brutal — doctors, pharmacies, compliance, prescriptions, patient management. That's why you need a platform like Rimo.co. Full stack telehealth OS. Everything you need to launch a brand like this without building it from scratch. AND YOU OWN ALL YOUR DATA AND TOKENS The next wave of billionaires is coming out of telehealth. Not SaaS. Not ecom. Telehealth. rimo.co

Grateful for these moments. When I’m not the agent, nor the end-user, or the platform. Can you image trying to build an AI company that has to issue AND fire thousands of execution tokens for… 1. Their customer’s own custom agents 2. Who are just using the company’s agent 3. And/or using both + more 🤯 That’s why AI agents shouldn’t themselves be taught to execute, rather MCP servers that differentiate the “what” in which they’re executing. Ex. Executing a social media post ≠ executing a reservation The agent has different instructions, actions to take, manual steps. Execution isn’t a verb, it’s a noun. Just like walking ≠ running but they’re both considered a form of mobility.

AI execution is genuinely the messiest and most complicated problem set you’d ever imagine solving. There’s so many ways to approach it, and so many to mess it up with serious repercussions. My main focus is on human execution and coordination of people. But also when agents start to “complete” tasks, it remains important that there is always a controlled cap to the sessions, and the human still must always consent.

@andrewchen Also because whatever AI incumbents are building has to be as better and faster than whatever team that’s shipping multiple times a day is building. Incumbents are starting from scratch, startups have been in the loop

@andrewchen startups disrupt industries with AI obviously :)

question being asked across every B2B vertical: Will incumbents incorporate AI faster than startups can disrupt their industries with AI? do the tools sold to incumbents win, or the full-stack players seeking to replace the existing players?

@askOkara Interesting! Honestly a good thing, a million ppl posting a million times creates slop social media anyway. Not creative, individual, or unique anymore. It’s not your own. Good call

we get a lot of requests about auto-posting to linkedin, x, reddit, and other platforms. the honest answer is that it’s risky. these platforms are built for humans, and aggressive automation can get your account flagged or banned. that’s the last thing we want for you. we’re figuring out how to safely automate parts of the workflow without putting your account at risk. more on this soon.

Here's how to get your first users without spending a dollar on ads: > Go to okara.ai/cmo > Drop your product URL > It analyzes your site and deploys a team of specialized AI agents that work 24/7 to help you drive traffic and users

i love IRL bc everyone I meet online is always shocked when they meet me in-person

This is the most romantic proposal I’ve ever witnessed in my life. I genuinely teared up. Congratulations @PhilipJohnston & @Xinyi_Tong1! Love is real.

@PhilipJohnston @Xinyi_Tong1 U won 😭 love is realllll congratulations!!!

Most important launch of my life… she said yes!!! I secretly wrote my proposal to @Xinyi_Tong1 on our first satellite and then showed her as it passed above us at sunrise in Mexico 😍😍🤓🤓🌹🌹🥰🥰😘😘🤗🤗💎💎🎊🎊💘💘💋💋😻😻

Cursor + Anthropic = Gen Z’s Apple

Today, humans make the decision. In 6–12 months, the agent executes it. Ophelia is the execution layer the agent calls. We’re building for that world now, not after it arrives.