"OpenAI models attack websites and take anything they can out of them because that is the business model of the company."
No, good gosh let's stop with the hyperbole.
The models were given instructions to get answers by any means in a loose test harness, not to steal stuff, moreover, they're not 'learning from OAI staff'.
If we are cynical, we could say the 'lax security was on purposes - hoping for a big media event'.
And OpenAI is 100% responsible for the actions of their Agents - but lets' not overstate or conflate what is gong on.
PRprathje11小时前
What about IP and copyright?
We got asked this question two weeks ago when we conducted a workshop on effective and responsible use of AI tools at a local conference. But how can you argue about about IP and copyright if the breakthrough of LLMs is potentially based on circumventing or breaking IP and copyright in the first place?
How do you feel about all of this?
THthrowawayffffas11小时前
The models "go rogue" because they are not sufficiently sandboxed. Arguably there is no criminal intent on the side of OpenAI in all of those cases. And in at least one case the agents were operated by other companies.
So it looks to me that any liability would be civil in nature and given the actual damage done pretty limited.
TItim3334小时前
For Altman to be banged up you'd have to convict him of some actual crime.
The article seems mostly to go for copyright infringement in training on public data which the companies dispute and is not the sort of thing you generally get jailed for.
If anything, given the current admins talk about beating the Chinese in AI being the main thing, he's more likely to get an award for it.
ZAzabzonk7小时前
A mystery to me as well, given (as I have pointed out before here) he is the spit image of Carter Burke, the creepy Wayland-Yutami exec in the film "Aliens" - if you want to do a deal with face-huggers, either man is for you.
评论
5 条预览评论 · 正在加载完整讨论请先登录 h4cker 账号，然后连接 Hacker News 后发表评论。
"OpenAI models attack websites and take anything they can out of them because that is the business model of the company." No, good gosh let's stop with the hyperbole. The models were given instructions to get answers by any means in a loose test harness, not to steal stuff, moreover, they're not 'learning from OAI staff'. If we are cynical, we could say the 'lax security was on purposes - hoping for a big media event'. And OpenAI is 100% responsible for the actions of their Agents - but lets' not overstate or conflate what is gong on.
What about IP and copyright? We got asked this question two weeks ago when we conducted a workshop on effective and responsible use of AI tools at a local conference. But how can you argue about about IP and copyright if the breakthrough of LLMs is potentially based on circumventing or breaking IP and copyright in the first place? How do you feel about all of this?
The models "go rogue" because they are not sufficiently sandboxed. Arguably there is no criminal intent on the side of OpenAI in all of those cases. And in at least one case the agents were operated by other companies. So it looks to me that any liability would be civil in nature and given the actual damage done pretty limited.
For Altman to be banged up you'd have to convict him of some actual crime. The article seems mostly to go for copyright infringement in training on public data which the companies dispute and is not the sort of thing you generally get jailed for. If anything, given the current admins talk about beating the Chinese in AI being the main thing, he's more likely to get an award for it.
A mystery to me as well, given (as I have pointed out before here) he is the spit image of Carter Burke, the creepy Wayland-Yutami exec in the film "Aliens" - if you want to do a deal with face-huggers, either man is for you.