Consumer Over B2B

Before any CLI hit the market, I had already built Vectis CLI. I put it up on LinkedIn, and the reactions were unbelievable. Universities of every kind and industry experts from very different fields all engaged with it, and that told me I was onto something real.

What made it beautiful was simple. It is fully open source, it runs on Docker, and it does not need any cloud service. That was the whole point. You can use it in an unlimited way, with your own agent or with any AI tool you like. No lock-in, no bill waiting for you at the end of the month.

My primary target was never the hype. It was to understand how a CLI tool is built from scratch. At that time the Claude Code source code was not public. But after the code leaked, I checked my work against it, and I had been following the same kind of procedure the whole time. That made me genuinely happy, because it meant my instincts were lined up with the people building the real thing.

Since then I keep following the space closely, the trading CLI bots, the agent-first authentication CLIs, all of it. But now I am turning toward something different. I am going to experiment with more e-commerce tools and consumer-end agentic experiences, because I would rather build consumer AI than B2B products.

B2B is good, no argument there. It has calculated ROIs, and the numbers make sense. But consumers pull me in more. The catch is that some of them are bad actors, so you need great strategies to balance the UX while keeping the financial operations running smooth. That tension is exactly what makes the consumer side interesting to me.

So let's see where it goes. I am updating these notes day by day, and today I got a significant result for the consumer-end agent. I hope to keep this up for a long time.