Hacker News Reader

Comment by s1artibartfast | Hacker News Reader

I think it is thuggery. If DOD doesn't want to use anthropic as vendor, thats fine.

Prohibiting contractors and departments from using it to produce products well within anthropic terms of service has no risk.

It is petty retaliation and strong arming.

Imagine if Ford said they weren't interested in building tanks, and then the DoD said employees and contractors can't use Ford for any purpose, because they are a supply chain risk

View on HN

If Ford told the DoD you can't use the tanks they bought however they need to, and had a kill switch in it in case they wanted to cut the DoD off, then that would obviously be a supply chain risk.

The military and it's suppliers need mitigate risk, and that includes suppliers intentionally causing harm to soldiers on the front lines, or in the backend logistics supply chain.

Honestly, every company should be treating Anthropic and OpenAI as supply chain risks, and should completely avoid dealing with them.

> If Ford told the DoD you can't use the tanks they bought however they need to, and had a kill switch in it in case they wanted to cut the DoD off, then that would obviously be a supply chain risk.

This is a misrepresentation of the situation. Anthropic did NOT sell something to the DoD and then restrict usage - the DoD purchased their offering fully aware of this clause.

The ability by the supplier is the problem, regardless of there is an upfront clause or not. The upfront clause is the problem

That's true for every company and every individual involved with the dod. They can all put guns or pencils down and refuse at a future date for a unspecified reason.

That has nothing to do with supply chain risk designation. That comes from sabotage

They won't because Trumps family has billions invested in OpenAI.

Reply on news.ycombinator.com