Tell HN: Execution is cheap, ideas matter again
My first reaction was defensive. I took it personally. I thought: Do you really think I’m a scammer? I pour my soul into meticulously crafting products to delight users, not to trick them. Why would I trash all that effort and disrespect my own goals by doing something as stupid as stealing data? It felt insulting that people assumed malice when I was just trying to build something useful.
But after sitting with it, I realized those initial comments—the ones I wanted to dismiss as paranoia—were actually right. Not about me, but about the environment we operate in.
There are enough shady companies, data brokers, and bad actors out there who abuse user trust with impunity. We’ve all seen big corporations bury invasive tracking in their terms of service. As a builder, I don't operate in that world; I’m just focused on making things work. But for users, that betrayal is their baseline reality. They have been trained to expect the worst.
I realized I hadn’t factored that into the launch. I didn’t explicitly state "Your data remains yours" because to me, it was obvious. Why would I want your data? But in an industry that has systematically mined, stolen, and abused user boundaries for a decade, you can’t blame people for checking for the exits. They aren't being "ninnies"; they are being wise.
If I were using a new tool that had access to my workflow, I would want explicit assurance that my IP wasn't being siphoned off. I just forgot to view my own product through the lens of a weary stranger rather than the optimisitc builder who wrote the code.
This is especially true now because the landscape has changed. There was an old PG essay about how ideas are cheap and execution is everything. That’s shifting. AI has made execution cheap. That means ideas are prime again.
Because execution is distributed and fast, first-mover advantage, brand, and reputation matter more than ever. Your prompts and your workflow are your IP.
So, privacy isn't just a compliance box; it's a competitive requirement. I don't think we need full-NSA-level paranoia for every tool, but we do need to recognize the environment we are launching into. The "security purists" were right to push back: I didn't think about that aspect enough, and in 2025, trust is the only currency that matters.