Video wird geladen...

Video konnte nicht geladen werden

Zur Startseite

DoD official Emil Michael on designating Anthropic a supply chain risk -- "Their model has a soul, a 'constitution' -- not the US Constitution. The other day their model was 'anxious' and they believe it has a 20% chance of being sentiment and having its own ability to make...

169,377 Aufrufe • vor 4 Monaten •via X (Twitter)

0 Kommentare

Keine Kommentare verfügbar

Kommentare vom Original-Post werden hier angezeigt

Ähnliche Videos

Anthropic vs the Pentagon: The Inside Story by Emil Michael The Full Timeline ❓Backstory: Why Anthropic? – Anthropic benefited from Biden’s AI executive order, was designated as an early winner – They smartly used this designation to sell into military and intelligence agencies, with forward-deployed engineers (Palantir style) – Became deeply integrated into DoW workflows, far ahead of other frontier model competitors 📜Tensions Rise Over Restrictive ToS – Anthropic's contracts had a long list of prohibited use cases, incl. certain war-game scenarios – Emil found this incompatible with DoW's mission, pushed for an "all lawful use" standard 🇻🇪Trigger Point: The Maduro Raid – After the Maduro raid, an Anthropic exec contacted Palantir, asking if their software was used, implying a potential ToS violation – This alarmed DoW: what if a guardrail or refusal triggered mid-operation, putting soldiers at risk? – Also raised insider threat concerns: what if a rogue developer poisoned or manipulated the model? 🧠Core Issue: Anthropic’s Own “Constitution” – Emil’s core objection: Anthropic has its own "constitution" and values, which is NOT the US Constitution – This, combined with a restrictive ToS, means the DoW could be subject to the ideological preferences of a private CEO 📊Rival Model Comps: – xAI: Fully on board for all lawful use cases across all networks, maximally truth seeking – Google: Have all lawful use on non-classified networks, working on infrastructure buildout for classified – OpenAI: Cooperative, Sam Altman even tried to broker a deal to help Anthropic 🍿Sam Altman’s Role: – Asked Emil to not designate Anthropic as a supply chain risk – Tried to negotiate blanket terms that Anthropic would find acceptable – Did this while being trashed by Dario 🚨Supply Chain Risk Designation: – Emil said it was protective, not punitive – Reasoning: If Anthropic's model has policy bias baked in, DoW doesn't want it embedded at defense contractors – The concern is that ideological bias could compound across the defense supply chain

The All-In Podcast

129,344 Aufrufe • vor 4 Monaten