Not from a client. Not from another AI builder. From a friend who doesn't care about AI at all. I showed him the lead response flow and expected feedback on the tech. Instead he said: "So basically this stops businesses from replying late?" That was it. No questions about prompts. No questions about models. Just the outcome. Funny how people outside the AI bubble force you to simplify. Anyway, Still a long way to go. But for the first time, it feels less like an experiment and more like a product.

The two companies racing hardest in AI, Anthropic and OpenAI, both just asked the world for a way to slow AI down, right as they each took a step toward going public. In this video I break down what both plans actually say, why the labs with the most to gain would call for a pause, and whether a global slowdown could even work. The real takeaway isn't about who wins the race. It's that the one thing you fully control is how good you get at using this stuff, so bet on yourself instead of a company.

I showed the same offer to 2 people. Person #1:"This is too expensive." Person #2:"This feels suspiciously cheap." Same offer. Same pricing. Same page. 😭 Spent the next hour trying to figure out what they were actually reacting to. Turns out neither person was talking about the price. One was reacting to risk. The other was reacting to perceived value. Made me realize pricing is way less about numbers than I thought. Now I'm reworking how the offer is presented before touching the actual price. Has anyone else run into this?

And honestly, I'll take it. A workflow that kept randomly failing for the past 3 days finally worked end-to-end. No errors No weird outputs No modules breaking for no reason. Just: Lead comes in → context gets structured → response generated → follow-up scheduled Watched it run once. Then again. Then one more time just to make sure I wasn't hallucinating 😭 Still needs work. But there's something satisfying about seeing a bunch of disconnected pieces finally cooperate. Building these systems is weird. You can spend an entire day making zero visible progress. Then one tiny thing works and suddenly the whole week feels worth it.

You can spend 8 hours building something impressive… then lose the client because your follow-up message sounded robotic 😭 Ran into this today while testing outbound flows. Technically the system worked: - lead came in - AI generated response - follow-up triggered correctly - everything automated But the conversation still felt… off. Too polished. Too “assistant-like.” So I started comparing actual business owner replies vs AI-generated ones side by side. Real humans: - reply faster - sound shorter - sometimes ignore grammar completely - get straight to the point Meanwhile AI keeps trying to sound like customer support from a Fortune 500 company 💀 Spent hours adjusting tone and weirdly enough: the more HUMAN and slightly imperfect the responses became, the better the conversations felt. Starting to think the future winners in AI automation won’t be the people with the smartest systems… It’ll be the people who make automation feel invisible.