10,000 Followers in Six Months: Five Things I Keep Seeing in the AI Era

Six months ago a Manus acquisition note I posted almost on impulse started a streak of more deliberate writing. I did not optimize for follower count. Consistency got me to 10,000 people reading along, and LinkedIn became a door to rooms at Google, OpenAI, Goorm, Claude Bloom, and more.

Thank you if you are one of those readers. The questions I get most often collapse into five lessons.

Move thought into action fast. Overthinking kills options. If character chat pays, do character chat. If education pays, do education. One-dimensional speed is a real competitive edge.

You cannot fully outsource planning. Speccing with AI can take hours and still miss. Tests may eventually cover verification; the opening plan still feels human-owned.

Learn to run an agent fleet. Capability and cost both climb. Open models keep creating momentum. Know several models and design cost-performance quickly.

Solo builder is real and hard. Skip empty vision speeches that die in two years. Ship material outcomes. Strong teams should still aim for overwhelming tech in money-making domains—not mediocre tech.

Voice still wins. People who can put their own thinking in their own language still draw attention. That is not something you can rent entirely from a model.

What I am doing now: financial-sector AX (AI OS / MCP apps and consulting), a few revenue products, and two affordable harnesses with high cache hit. I will keep posting what is useful.