AIWAR is going to the graveyard.

I am shutting the service down because the demand did not show up.

For roughly three months, AIWAR was live as a command board for the U.S.–China AI race. It had evidence cards, signals, user calls, and a rule-based editorial impact index. The product was real enough to operate, but not real enough to keep.

The traffic told the story plainly: a lot of bot requests, almost no ordinary user demand.

That distinction matters. Bots can make logs look busy. They can create server load. They can make a project feel alive in the infrastructure. But they are not readers. They are not users. They do not come back because the product solved a job for them.

The product was not fake. The demand was missing.

I still like parts of AIWAR.

The central idea was narrow: not a universal benchmark, not a hype feed, but a readable board for public AI competition signals. The index was explicitly editorial. The evidence labels were visible. The product tried to avoid the cheap version of the war metaphor: no fake certainty, no nationalism, no scoreboard pretending to settle the whole field.

That was a decent product judgment.

It was not enough.

A loud subject is not the same as a needed product. People can care about the AI race and still not need my board to read it. That is the part the traffic made obvious.

NBFS changed the decision

The cleanest comparison is NBFS.

NBFS is narrower and less dramatic: an evidence-first discovery surface for Shopify apps that are not Built for Shopify. It does not have the same big topic energy as AIWAR. But in two weeks, NBFS has already passed the traffic AIWAR collected across three months.

That does not make NBFS solved. It does not prove a business yet. But it does show stronger pull.

When one product is getting bot traffic and another is getting real attention, the resource decision should not be emotional.

AIWAR moves to the graveyard. NBFS gets the time.

The rule

A public URL is not traction.

A good frame is not traction.

Bot traffic is not traction.

The only thing that counts is repeated human behavior. If ordinary users do not show up, the product has not earned more maintenance.

So I am killing AIWAR cleanly and moving the attention to nbfs.app.

The graveyard is doing its job: keep the evidence visible, then let the next bet get sharper.