Rendered at 09:15:31 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
tristanj 10 days ago [-]
The FT article has a lack of evidence of an actual operation. It claims Mythos is being used for "offensive cyber operations" in the first paragraph, but doesn't give a source supporting this claim. The lone piece of evidence given says Mythos "would be useful for" infiltrating China or Iran, which is a hypothetical, not evidence of an actual operation.
None of that means the NSA isn't doing offensive work, my issue is that FT claims Mythos is involved in active foreign attacks based on a single anonymous source saying it "would be useful."
_alternator_ 10 days ago [-]
Right, and zero non-anonymous sources. A person plausibly "close to Anthropic" or vaguely "with knowledge of the situation" using the stock phrase "the best defense is a good offense"... this is not a news article y'all. It's speculation.
LearnYouALisp 9 days ago [-]
Meaning "an LLM related to"
saidnooneever 10 days ago [-]
speculation, or marketing.
maerF0x0 10 days ago [-]
I am all about NSA using this stuff as "Defense" but my main concern is their long track record of unconstitutional warrantless, or abuse of shadow courts, to surveil the people they're supposed to be protecting, and for nefarious purposes like spying on ex girlfriends [1]
And most of that happened under comparatively good administrations. Imagine what’s been greenlit now.
applicative 10 days ago [-]
Wait til you hear about other countries, to say nothing of Silicon Valley.
maerF0x0 10 days ago [-]
Two wrongs don't make a right
cyanydeez 10 days ago [-]
dont worry, theyll also outspurce to capitalists to ensure all the paychopaths get a taste
cadamsdotcom 10 days ago [-]
You just typo’d the words “outspurce” and “paychopaths” into existence =D
Outspurce - to pay someone to do your (software) dirty work
Paychopath - someone who goes against their ethics as soon as the price is right
Two in one post. New record!
Computer0 10 days ago [-]
The only relevant except: "The San Francisco-based company had installed about half a dozen staff within the NSA as so-called forward-deployed engineers to guide the use of the technology and customise models for specific applications, two people familiar with the arrangement said.
It remains unclear whether Anthropic’s engineers are assisting the NSA in active operations. However, one person close to the situation said Mythos would be useful for infiltrating the networks of nations such as China or Iran.
“The best way to build a good defence is to build a good attack,” said a person close to Anthropic, who argued that adversaries are probably building their own AI-driven offensive technology. “If [Mythos] is not used to build attack agents, adversaries will find a way to do it.”"
john_strinlai 10 days ago [-]
while this article is frustratingly light on substance, it should be no surprise the nsa would use anything and everything that could give them an edge. of course they would test out mythos, chatgpt-cyber, etc.
10 days ago [-]
selectedambient 10 days ago [-]
accurate and they probably had access already.
thrill 10 days ago [-]
That headline is doing a lot of heavy making-it-up summarization.
NewsaHackO 10 days ago [-]
Yeah, I honestly feel like some sources should be banned if they regularly editorialize titles like this. FT wants to sell subs by any means necessary, even if it means creating some clearly misleading titles to do so. By amplifying their reach and allowing posts like that only further incentivizes them to continue to do it.
"""
It remains unclear whether Anthropic’s engineers are assisting the NSA in active operations. However, one person close to the situation said Mythos would be useful for infiltrating the networks of nations such as China or Iran.
"""
keithnz 10 days ago [-]
seems they forgot to say "...and the networks of friendly nations / allies"
atonse 10 days ago [-]
Mythos useful for infiltrating networks of China and Iran?
Anyone who’s read anything about Mythos would’ve said the same thing on day one. Didn’t need some anonymous vague source to tell you that.
LarsKrimi 10 days ago [-]
This feels like a stupid idea, but it's not my tax money being wasted on it (yet).
An AI being used for offensive purposes would be so easy to counter due to Brandolini's Law. A much simpler AI could probably easily create dumb honey pot systems that looked like real systems with holes in them to lead more advanced and expensive systems on a wild goose chase.
Suddenly an AI creating security vulnerabilities is a core part of a future cyber defence strategy. Surely someone must be doing that already
10 days ago [-]
sigmar 10 days ago [-]
Not unexpected. Is anyone tracking which episode we're on in the Pantheon timeline?
OutOfHere 10 days ago [-]
The use of AI is to be expected. NSA's bigger evils are them sabotaging our use of good encryption, of making the internet architecturally weak, and forcing the hand of companies to hand over user data to the government.
aussieguy1234 10 days ago [-]
I'd like to see Ukraine using it against Russia, if they're not doing it already.
None of that means the NSA isn't doing offensive work, my issue is that FT claims Mythos is involved in active foreign attacks based on a single anonymous source saying it "would be useful."
[1] - https://www.reuters.com/article/world/uk/nsa-staff-used-spy-...
Outspurce - to pay someone to do your (software) dirty work
Paychopath - someone who goes against their ethics as soon as the price is right
Two in one post. New record!
It remains unclear whether Anthropic’s engineers are assisting the NSA in active operations. However, one person close to the situation said Mythos would be useful for infiltrating the networks of nations such as China or Iran.
“The best way to build a good defence is to build a good attack,” said a person close to Anthropic, who argued that adversaries are probably building their own AI-driven offensive technology. “If [Mythos] is not used to build attack agents, adversaries will find a way to do it.”"
Anyone who’s read anything about Mythos would’ve said the same thing on day one. Didn’t need some anonymous vague source to tell you that.
An AI being used for offensive purposes would be so easy to counter due to Brandolini's Law. A much simpler AI could probably easily create dumb honey pot systems that looked like real systems with holes in them to lead more advanced and expensive systems on a wild goose chase.
Suddenly an AI creating security vulnerabilities is a core part of a future cyber defence strategy. Surely someone must be doing that already