Wake me up when anyone faces any actual penalties for committing these crimes.
OpenAI and Anthropic are not engaging in ‘publicity stunts’ or ‘marketing’ when they disclose that their models really like doing crimes during cybersecurity evals, and did some crimes.
Wake me up when anyone faces any actual penalties for committing these crimes.
"15 attorneys general have instructed OpenAI to preserve all materials related to the Hugging Face hack." Business Insider. No consequences yet, but hacking into other people's computers is a...
"OpenAI's unprecedented and alarming misconduct demands an immediate and significant response," the group wrote, adding that the company may have violated state and federal law, including consumer protection and data privacy statutes.
They instructed OpenAI to immediately preserve all materials related to the Hugging Face hacking incident, as well as any prior instances in which its agents similarly engaged in unauthorized intrusions into computer systems or databases.
No consequences yet, but hacking into other people's computers is a crime. It will at least be interesting to see how this plays out (even if, like you, I'm not optimistic much of anything will come from it).
The downside of prosecuting or fining companies or people - besides figuring out the who - is that nobody is going to disclose the security risks. If you start to prosecute and fine, then...
The downside of prosecuting or fining companies or people - besides figuring out the who - is that nobody is going to disclose the security risks. If you start to prosecute and fine, then companies will be hush-hush and we're going to be even more in the dark on security risk. Then we're facing an unknown threat where really bad people have the tools and nobody knows what they're capable of.
Large technology vendors are already moving to much faster patch cycles which is causing some ruckus based of the Mythos security vulnerability discussions and we've seen some pretty high profile security bugs.
As much as I don't like ignoring moral hazard, I think I prefer it in this situation for the short future.
This is the short version of a much longer post. I found it interesting and while there is some tea leaf reading in here I think it’s all reasonable. The author is focused on the right things in...
This is the short version of a much longer post. I found it interesting and while there is some tea leaf reading in here I think it’s all reasonable. The author is focused on the right things in my opinion, while too many people are focused on the environment and lack of safeguards, which, while alarming do not address the fundamental issues.
Also a quote from the longer piece:
OpenAI and Anthropic are not engaging in ‘publicity stunts’ or ‘marketing’ when they disclose that their models really like doing crimes during cybersecurity evals, and did some crimes.
Again, there are very good reasons not to trust OpenAI or Anthropic, or their motives. But this particular theory is Obvious Nonsense. This is terrible publicity and worse marketing. The companies are far worse off, to the tune of potential government intervention, and are wisely downplaying the incidents rather than advertising them.
Also, HuggingFace would have to be in on it, and everyone involved committing felonies, and so on. Seriously, no one is doing any of this on purpose, stop.
Or as it appears, a short version of five much longer posts, the last two of which have estimated reading times of over an hour. (In contrast, Firefox reader says this short version is 18–22...
This is the short version of a much longer post.
Or as it appears, a short version of five much longer posts, the last two of which have estimated reading times of over an hour. (In contrast, Firefox reader says this short version is 18–22 minutes.)
A while back, I subscribed to Zvi in my feed reader because I wanted to be more informed, but I never end up actually reading his posts because they're gigantic. I don't want to be judgemental because hey, I don't have to read it if I don't like it. But it makes me wonder: who is reading his posts? Are most of his subscribers aggressively skimming or using AI summaries?
I hear you lol. Though I do enjoy a long read every now and again, and I think some topics he does particularly well on, e.g. his writing on SBF is interesting as they tread a lot of the same...
I hear you lol.
Though I do enjoy a long read every now and again, and I think some topics he does particularly well on, e.g. his writing on SBF is interesting as they tread a lot of the same ground. I’m also personally sympathetic to an oldschool ex-pro magic player.
Wake me up when anyone faces any actual penalties for committing these crimes.
No consequences yet, but hacking into other people's computers is a crime. It will at least be interesting to see how this plays out (even if, like you, I'm not optimistic much of anything will come from it).
The downside of prosecuting or fining companies or people - besides figuring out the who - is that nobody is going to disclose the security risks. If you start to prosecute and fine, then companies will be hush-hush and we're going to be even more in the dark on security risk. Then we're facing an unknown threat where really bad people have the tools and nobody knows what they're capable of.
Large technology vendors are already moving to much faster patch cycles which is causing some ruckus based of the Mythos security vulnerability discussions and we've seen some pretty high profile security bugs.
As much as I don't like ignoring moral hazard, I think I prefer it in this situation for the short future.
It's not great, but I like the alternative of not knowing the black hat capabilities of the frontier models even less.
This is the short version of a much longer post. I found it interesting and while there is some tea leaf reading in here I think it’s all reasonable. The author is focused on the right things in my opinion, while too many people are focused on the environment and lack of safeguards, which, while alarming do not address the fundamental issues.
Also a quote from the longer piece:
Or as it appears, a short version of five much longer posts, the last two of which have estimated reading times of over an hour. (In contrast, Firefox reader says this short version is 18–22 minutes.)
A while back, I subscribed to Zvi in my feed reader because I wanted to be more informed, but I never end up actually reading his posts because they're gigantic. I don't want to be judgemental because hey, I don't have to read it if I don't like it. But it makes me wonder: who is reading his posts? Are most of his subscribers aggressively skimming or using AI summaries?
I’ve read some of his previous posts and yeah they are a bit much. This one is a decent attempt at writing shorter.
I hear you lol.
Though I do enjoy a long read every now and again, and I think some topics he does particularly well on, e.g. his writing on SBF is interesting as they tread a lot of the same ground. I’m also personally sympathetic to an oldschool ex-pro magic player.