Fable/mythos are the first models from anthropic that hide 100% of reasoning tokens. So it seems to me like we're about to get a lot more data about to what extent Chinese model progress has been a consequence of distillation techniques.
How so? The fact that China is going to launch a comparable model soon is the whole point of what’s happening now. Everyone knows there are going to be open models soon that have the same capabilities - Anthropic has literally said so. The restrictions now are to buy time to patch the security issues before those Chinese models are made available to the whole world.
The Wired headline reframes the issue in a way that’s misleading. SK Telecom was a previously resolved issue (as in prior to Fable launch).
It may have been a contributing factor, but the crux of the shutdown was the industry reporting of Fable jailbreaks (reportedly spearheaded by Amazon CEO Andy Jassy). The more interesting and honest angle is that the industry which has taken the seriousness of Glasswing at face value felt blindsided by Fable release and totally exposed by the residual risk, when they know they still have a months-long bugfixing backlog exposed by Glasswing and are desperate to buy more time.
This misleading looks deliberate on Wired’s part, to appear as though they’re getting a scoop when they’re really just being dishonest. Shameful.
> honest angle is that the industry [felt] exposed by the residual risk [and] have a months-long bugfixing backlog exposed by Glasswing
Two problems with this theory.
1. Amazon complaining to the White House wouldn't have been the opening salvo. Amazon and Anthropic would find it much easier to talk to each other than go through the White House. We'd need evidence that Amazon (and probably others) already asked Anthropic to not release a Mythos-class model but Anthropic released it anyway. Are they on record saying this?
2. The jailbreak Amazon found needs to be real. Maybe the White House staffers are not AI experts and they don't really understand what a jailbreak is... but it's much harder to make that claim about Andy Jassy. For the jailbreak to be the real reason for the export control order, the jailbreak would need to be significant and cause material harm to Amazon. Then Jassy might pass it along to the White House assuming he already was refused by Dario.
But there is no evidence the jailbreak was real. There is one story that it amounted to a request, "fix this code." In any case, Anthropic is on record saying the so-called jailbreak didn't enable any vulnerability work that couldn't already be done by other models.
Is there any interpretation where Amazon is involved but for a reason besides trying to screw over Anthropic? I don't see why Amazon would act against Anthropic when they have a deal to offer Anthropic models on AWS.
Many of us know coders and cybersecurity professionals who are even better than Claude Fable or Mythos. It's outstanding how much praise and careful consideration it gets. At the same time, humans with even more expertise are discarded and fired.
> SK Telecom has poured capital into Anthropic several times, including a $100 million investment in 2023 that coincided with the formation of a commercial partnership to develop an AI model tailored to the telecommunications industry.
> the White House asked Anthropic to revoke SK Telecom’s access to Mythos, according to a person close to the AI lab. The company immediately complied,
The articles and discussion around this and the Amazon story all seem to me to be an earnest tech press and community searching for a genuine reason for the administration blocking Anthropic’s models.
However, thinking back to the spat with the DoD and more generally how the administration is much more supportive - and supported by - OpenAI and XAI and it’s easy to imagine this is just another escalation in the fight between a “liberal leaning” company and its competitors and the administration.
There might have been something said by someone at Amazon or something but I’d guess Occam’s razor the administration just leapt at the chance after their supplier sanctions fell flat?
its very real, chinese capital is being intertwined/injected into all facet of south korean society through its own domestic companies and politicians.
Over a decade ago, I was going to a tech group with a supposed security leader (that ended up getting a huge buyout from Cisco a few years ago).
A student wanted to create a project that bypassed the GF of China. He basically told him he was an idiot, it was impossible, and discouraged the student from even trying. This was a hacker (the security leader) at one point in his life, that claimed to support the freedom of speech and expression.
The actual reason he said this to the student is because he supports the government in China. Many other comments I heard after this proved to me that this is true. My guess is that his startup even received funding from the Chinese government.
It's really disappointing that tech leaders can publicly support murderous regimes like the Chinese government for decades with no blow back. They are far worse than Putin/the Russian government and have 100X the money and technology.
If he supported Russia, he would probably get cancelled and never be able to be involved in the tech industry again.
Looking at the comments on Korean political articles, the point is not so much whether SKT is actually related to China, but rather that the switch to cut off AI vendors is now outside the realm of model performance specs. Now, when foreign companies integrate AI into their workflows, they probably need to add a category for vendor continuity in their evaluation criteria. From a corporate standpoint, uninterrupted continuity is what matters, and with this case, Anthropic is likely to lose a lot of trust, contrary to its own stated policies.
22 comments
[ 2.7 ms ] story [ 42.1 ms ] threadWhy do you think China will?
I am quite certain the gap will only grow
It may have been a contributing factor, but the crux of the shutdown was the industry reporting of Fable jailbreaks (reportedly spearheaded by Amazon CEO Andy Jassy). The more interesting and honest angle is that the industry which has taken the seriousness of Glasswing at face value felt blindsided by Fable release and totally exposed by the residual risk, when they know they still have a months-long bugfixing backlog exposed by Glasswing and are desperate to buy more time.
This misleading looks deliberate on Wired’s part, to appear as though they’re getting a scoop when they’re really just being dishonest. Shameful.
Two problems with this theory.
1. Amazon complaining to the White House wouldn't have been the opening salvo. Amazon and Anthropic would find it much easier to talk to each other than go through the White House. We'd need evidence that Amazon (and probably others) already asked Anthropic to not release a Mythos-class model but Anthropic released it anyway. Are they on record saying this?
2. The jailbreak Amazon found needs to be real. Maybe the White House staffers are not AI experts and they don't really understand what a jailbreak is... but it's much harder to make that claim about Andy Jassy. For the jailbreak to be the real reason for the export control order, the jailbreak would need to be significant and cause material harm to Amazon. Then Jassy might pass it along to the White House assuming he already was refused by Dario.
But there is no evidence the jailbreak was real. There is one story that it amounted to a request, "fix this code." In any case, Anthropic is on record saying the so-called jailbreak didn't enable any vulnerability work that couldn't already be done by other models.
> the White House asked Anthropic to revoke SK Telecom’s access to Mythos, according to a person close to the AI lab. The company immediately complied,
Lesson learned - don't invest in US companies
However, thinking back to the spat with the DoD and more generally how the administration is much more supportive - and supported by - OpenAI and XAI and it’s easy to imagine this is just another escalation in the fight between a “liberal leaning” company and its competitors and the administration.
There might have been something said by someone at Amazon or something but I’d guess Occam’s razor the administration just leapt at the chance after their supplier sanctions fell flat?
A student wanted to create a project that bypassed the GF of China. He basically told him he was an idiot, it was impossible, and discouraged the student from even trying. This was a hacker (the security leader) at one point in his life, that claimed to support the freedom of speech and expression.
The actual reason he said this to the student is because he supports the government in China. Many other comments I heard after this proved to me that this is true. My guess is that his startup even received funding from the Chinese government.
It's really disappointing that tech leaders can publicly support murderous regimes like the Chinese government for decades with no blow back. They are far worse than Putin/the Russian government and have 100X the money and technology.
If he supported Russia, he would probably get cancelled and never be able to be involved in the tech industry again.
Anthropic may discover a lack of availability for HBM memory in the near future.