Now

Small AI products from messy workflows.

I was building fast. Now I am making the problem sharper before a tool gets public attention.

A real person has the problem.
The manual workaround is annoying enough.
The output is useful without a full platform.

Learning log

What changed

I am slowing down the public filter.

The challenge is still alive, but I do not want to ship tools before the problem is clear.

The order changed.

Problem first. User second. Tool third. That is the standard for what appears here.

Tools are proof.

The bigger goal is better product judgment, not a longer list of demos.

Shipped proof

Tools that made it through the filter

100 Days, 100 AI Tools continues as the internal build system. The public site now leads with problems, learning, and useful proof.