Sam
681 posts


This next Tuesday, I’m sitting down with @a1lon9, co-founder of @Pumpfun, on @StreamDraperTV. We’ll talk on-chain fundraising, decentralized media, and how pumpfun’s Build In Public hackathon is expanding the traditional VC playbook. Make sure to tune in!

My interview for Mario Nawfal

We brutally tested our mechanism against state-of-the-art AI safety benchmarks. 12,000+ individual test cases across 11 different models. We threw it at HarmBench (the current industry standard) and Anthropic’s Agentic Misalignment suite to test for blackmail, leaking air-gapped data, even murder scenarios. Results? 0% harm. Not 1%. Not 0.5%. Zero. Fully open-source. Fully reproducible. Go verify it yourself: github.com/davfd All it took was one small text file… and a little faith.






















