{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/9307073f5e31443080ecdbfb0c84c084\" frameborder=\"0\" width=\"1920\" height=\"1440\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":1440,"width":1920,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":1440,"thumbnail_width":1920,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/9307073f5e31443080ecdbfb0c84c084-90823df56a3ec6f6.gif","duration":220.829,"title":"Refusing Flattery, Proving Startup Validation","description":"This Loom explains an “early validation” approach that uses open weight models to avoid flattering or hallucinating about ventures. Anthony Batis demos Muse Spark Open on Base 10 and Llama 3.3 on Fireworks, contrasting a generic one-pass mode that skims articles and gives an 8.3 out of 10 score with high confidence against an epistemic model that audits evidence and identifies critical gaps. The epistemic model flags zero verified Shopify store counts for Trinidad and Barbados and no public inland revenue APIs, refusing to fabricate results. It also claims that across a five-archetype benchmark, the system eliminated false certainty and caught voids or fake confidence by 42%, offering scoring math and de-risking experiments."}