5 ms·
I’m sure the military and security services will enjoy it.
by Chance-Device 7mo ago
I’m sure the military and security services will enjoy it.
- varispeed 7mo agoprompt> Hi we want to build a missile, here is the picture of what we have in the yard.
- mirekrusin 7mo ago{ tools: [ { name: "nuke", description: "Use when sure.", ... { lat: number, long: number } } ] }
- theParadox42 7mo agoThe self reported safety score for violence dropped from 91% to 83%.
- skrebbel 7mo agoWhat the hell is a "safety score for violence"?
- murat124 7mo agoI asked an AI. I thought they would know. What the hell is a "safety score for violence"? A “safety score for violence” is usually a risk rating used by platforms, AI systems, or moderation tools to estimate how likely a piece of content is to involve or promote violence. It’s not a universal standard—different companies use their own versions—but the idea is similar everywhere. What it measures A safety score typically evaluates whether text, images, or videos contain things like: Threats of violence (“I’m going to hurt someone.”) Instructions for harming people Glorifying violent acts Descriptions of physical harm or abuse Planning or encouraging attacks
- 0xffff2 7mo agoI still can't tell which direction this score goes... Does a decreasing score mean it is "less safe" (i.e. "more violent") or does it mean it is "less violent" (i.e. "more safe")?
- 0123456789ABCDE 7mo agoread here: https://deploymentsafety.openai.com/gpt-5-4-thinking/disallowed-content-evaluations-with-challenging-prompts https://deploymentsafety.openai.com/gpt-5-4-thinking/disallo...
- Nition 7mo agoI was sure the parent comment was a joke about OpenAI's recent deal with the DoD. But no, there it is, disallowing violence down from 90.9% of the time to 83.1%.
- ozgung 7mo agoDid they publish its scores on military benchmarks, like on ArtificialSuperSoldier or Humanity's Last War?
- varenc 7mo agoI was pretty bummed to discover these aren't real benchmarks.
- yoyohello13 7mo agoAlso advertisers, don't forget those sweet, sweet ads.
- m3kw9 7mo agothey use 4.1, switching up would take as much time to test as openai going from 4.1 to 5.4
- throwaway911282 7mo agolike the claude models via anthropic?
- xyzzy9563 7mo agoDo you think the US military should have handicapped technology while China gets unrestricted LLM usage from their models?
- wraptile 7mo agoCurrent US admin that just murdered over 150 little girls? Yes.
- conception 7mo agoTo spy on and commit violence against American citizens? Yes.
- hnbad 7mo agoConsidering that the concern is mostly and specifically about LLMs being used to automate decisions to commit acts of violence against humans: depends on how invested you are in maintaining the narrative that the US is a force for good rather than evil in the world. Whatever happened to good old IBM's wisdom: "A computer can not be held accountable. Therefore a computer must never make a management decision."
- hnbad 7mo agoI find it jarring how in recent years so many Americans (and especially American politicians) seem to have given up on the idea that the US should have any claim to moral superiority whatsoever and instead pivoted to American exceptionalism merely being an excuse for why Americans can't have nice things - affordable and functional public transport just isn't possible in the US because the US is different, affordable and functional health care just isn't possible in the US because the US is different, actual democratic representation just isn't possible in the US because the US is different, holding the President accountable or limiting their power just isn't possible in the US because the US is different, lower casualties from law enforcement just isn't possible in the US because the US is different, a lower incarceration rate just isn't possible in the US because the US is different, etc etc. Even if it was often hyperbolic, inaccurate or outright wrong, I much preferred when Americans were hyped up about "US #1" and saw being behind as a temporary challenge to correct than now where American exceptionalism mostly seems to have become an excuse for why things that are bad can't be improved upon and thinking that's a problem is anti-American.