I spent nine years as a product manager at Google. Since then I've run Gabriel Hounds, a private advisory focused on harassment, attribution, and enforcement, working in cloud infrastructure and security, and partnering directly with trust and safety teams at some of the largest platforms. I've seen this problem from the inside.

Online hate runs like a business now. Campaigns pay to spread it. Big accounts amplify it. Platforms get paid to host it. A 2026 audit put a real number on the fix: cleaning up most of the hate speech exposure on a major platform costs about 2.9% of revenue. The fine for leaving it up runs as high as 10%. Ten reinstated accounts known for hate and conspiracy content are worth an estimated $19 million a year in ad revenue to that platform. Senate GOP-aligned groups alone put about $82 million behind anti-trans TV ads in 2024. The people funding this know these numbers. I haven't found anyone publishing them in real time, with a name attached. (More of this research is at the Hobocode.net blog .)

I want to build the system that does that. I'm calling it the Online Hate Narratives Weather Report: trace a narrative back to where it started, follow it as it moves across platforms and into the mainstream, and name who's paying and who's profiting, wherever the trail allows.

Here's what I'm actually asking for right now: $5,000. That covers cloud infrastructure and the AI agent costs to run a feasibility phase, proving out whether this can be built and pricing out what it actually costs to run. Ingesting and classifying narratives across platforms with AI agents is exactly the kind of workload where the real cost can surprise you, and I want that number before I ask anyone for what an ongoing operation would need.

If the numbers work, this becomes a public, non-profit system with a free weekly bulletin. If they don't, you'll know that too. I'll publish what I find either way.

Any amount helps get this phase funded.