US ban on Mythos is related to a jailbreak research by Amazon researchers
maxloh
12 points
7 comments
June 13, 2026
Related Discussions
Found 5 related stories in 119.3ms across 10,416 title embeddings via pgvector HNSW
- Amazon security research reportedly led to the White House's Anthropic Fable ban kikibobo69 · 15 pts · June 13, 2026 · 67% similar
- Our response to the US ban on Fable 5 and Mythos 5 ubutler · 67 pts · June 13, 2026 · 67% similar
- Statement on US government directive to suspend access to Fable 5 and Mythos 5 Dylan1312 · 1713 pts · June 13, 2026 · 62% similar
- Has Mythos just broken the deal that kept the internet safe? jnord · 37 pts · April 10, 2026 · 61% similar
- Apple's Security Has Been Tough to Crack. Mythos Helped Find a Way In jbredeche · 13 pts · May 14, 2026 · 60% similar
Discussion Highlights (3 comments)
ryanschaefer
> According to a Wall Street Journal (WSJ) report, the technique in question was tested by researchers at Amazon, Why not link that? Is this a dupe? Found the primary source: https://www.wsj.com/tech/ai/amazon-ceos-talks-with-u-s-offic...
heroprotagonist
We should probably ban all of those Amazon models that can also be jailbroken.
DivingForGold
The 1st amendment, if that's what Anthropic claims, only goes so far. Releasing Mythos or Fable is much like giving away Javelin missiles for free to everyone and anyone, but then crooks and hackers are part of that group. Severe damages will result, even losses of life from 2nd and 3rd hand kinetic events, not missile explosions. Because of this danger I reluctantly favor government teams scrutinizing them and issuing approval before models are released, sorry. Anthropic guys step over the line, they do need to be arrested to show them you just don't behave like this... It's really just one or 2 steps away from a "ready to hack" exploit. OMG, it already happened: As reported, within hours of Fable 5's public launch, a researcher known online as Pliny the Liberator posted on social media claiming to have "liberated" the model. The method required no code exploitation, no reverse engineering, no software vulnerabilities — just carefully crafted prompts. The documented techniques included: Unicode and Cyrillic character substitution, well, there's the tagline, cyrillic substitution indicates the Russky's already have it. More drain of millions of wealth from corporations into hacker accounts via crypto.