
I remember the exact moment I realized a project I had poured two years of my life into was a ghost. It was a Tuesday, the office was quiet, and the metrics on my screen were not just flat, they were screaming for mercy. We had built something beautiful, something technically perfect, and something that absolutely nobody wanted to use in the way we intended.
Watching OpenAI pull the plug on Sora feels like watching a mirror of that same, hollow realization.
There is a specific kind of grief that comes with killing a product. It is not just about the code or the capital. It is about the identity you wrapped around the promise of what that product would become. When the news broke that the app was being shuttered, the industry focused on the money, the $1 billion Disney deal, and the compute costs. They missed the human cost of the pivot.
We love to talk about viral growth as if it is a proxy for value. Sora had its moment in the sun, a flash of brilliance that made us all wonder if we were witnessing the future of cinema. But viral attention is a fickle currency.
It buys you headlines, but it does not buy you retention. When the novelty of generating a woolly mammoth in the snow wears off, you are left with a product that needs to solve a real, recurring problem.
There is a perverse incentive in our industry to keep the lights on at all costs. We fear the optics of failure more than we fear the reality of stagnation. But there is a profound, quiet power in saying that something is not working. It requires a level of vulnerability that most executives are terrified to show. You have to stand in front of your team and admit that the vision you sold them was wrong.
When you stop chasing the ghost of a failed product, you finally have the bandwidth to see what is actually happening in the market. You stop looking at the competition through the lens of your own insecurity. You start building for the people who are actually showing up, rather than the ones you hoped would show up.
If you are currently holding onto a project because you are afraid of the post-mortem, you are already losing. The most dangerous thing you can do is let a failing product consume your best talent and your most precious compute resources.
True leadership is not about the products you launch. It is about the products you have the guts to kill before they kill your company.
No spam. One email with the asset, then occasional Spark updates.