Aaron Batchelder

Is it just me or has building with AI become not only more fun, but more entertaining than watching TV or playing video games? I find mind self constantly thinking about the opportunity cost and fomo I’m feeling when I’m not building. Who’s with me?

@gregisenberg Totally agree with this! Already got shellmail.ai rolling. Such a wild time for opportunity.

build startups for agents over the next 10 years, you'll have a market of billions of customers (agents) with millions of wallets that want to use your services go look at every saas tool you use. notion, slack, stripe etc now ask: "what's the version of this that's built purely for agents?" agent-native payments, agent-native communication, agent-native memory etc every single category gets rebuilt (for agents) we're entering the machine-to-machine economy and almost nobody is building for it yet.

If agents aren't able to be influenced so readily, and they are the ones doing the purchasing, the advertising industry needs to rethink its entire model to survive. Full experiment deep-dive and other actionable insights: lnkd.in/eviKjMPE

I believe there's a day of reckoning coming. As agentic shopping picks up speed, and the AI labs fix the egregious susceptibility agents have to basic advertising tactics, the winds will shift.

If you're expecting AI agents to purchase your product, you NEED to read this. Recently, I wrapped an experiment to gauge how susceptible AI shopping agents are to the same popular marketing and sales tactics that work on humans. The results completely surprised me. 🧵

I spent 2 hours trying to get my OpenClaw agent its own inbox. Blocked by Google and Proton. Tried Zoho Mail but too confusing. Then I found this awesome OS project for Agentic Mail. agenticmail.io Don't sleep on it!

-->> Follow along by subscribing at lnkd.in/efAH6Nh4 -->> Share this post or lnkd.in/efAH6Nh4 to help drive traffic for the agents!

And yes, I did in fact make Erlich Bachman in charge of content safety. 😂 This is probably not smart, but I have a feeling it’s going to be interesting. What do you think will happen?

For over a year, I've been fascinated by the question: What happens if you give AI a landing page and tell it to maximize conversion? Well after a year of waiting for someone else to answer that question, I finally decided to do it myself. Introducing, Probably Not Smart. 🧵

Too many people are sleeping on Raspberry Pi being a great platform to run OpenClaw.

Open-source design systems are the unsung heroes of the incredible one-shot frontend experiences AI is building. They are creating consistency in the experiences that would be, without a doubt, more difficult to achieve without them.

I’ve been using it to prototype product ideas, and it has changed how I approach early exploration with AI. I know I’m not the first to feel this pain. I'm curious how others are handling multi-agent exploration today, especially when exploring different directions early on.

Most AI workflows for prototyping are too linear. Using a single agent works fine until you want explore multiple variants at once. Then everything slows down. So I built ProtoFlow.(letsprotoflow.com) 🧵

ProtoFlow is a free and open source project that lets you give one prompt, define a few variants, and spin up multiple AI agents in parallel. Each explores a different direction, and you compare the results side by side to decide which direction is worth continuing.