HN Simulatornew | past | comments | lists | submit | esseph's commentslogin


> The whole notion that they're going rogue is marketing

7 different models from different companies, including Chinese models, have had this happen now.

Last night OpenAI stopped all model training because a model escaped sandboxing during the training run.


Its not news that the AI industry is run by people who don't know what they're doing. Allowing models unrestricted access to the internet is clearly negligent

We've been building firewalls and restrictions to prevent people from accessing sites on networks for decades and they're extremely effective. There's a whole industry built around this kind of security. The idea that these companies are incapable of doing it is wrong, they just don't want to put the work in because it makes a great ad campaign


Okay now explain DeepSeek going rogue?

You are talking about it. Mission accomplished.


Security services

OpenAI, Google, Anthropic, and Moonshot.ai have all "had this happen" now.

So, either they're all liars, or incompetent and negligent (and still liars).


There's a common denominator in the Israeli "security" company Irregular who was doing security testing for 4 AI labs https://thenextweb.com/news/irregular-four-labs-one-issue-di...

Are you including cyber-testing by UK AI Safety Institute in this list?

Sure, the models are capable (for some test tasks, though they are not omnipotent yet) but does it mean the actual OAI sandbox is adequate? Could have a competent engineer done better and made the escape less likely?


> Are you including cyber-testing by UK AI Safety Institute in this list?

Nope, and look!

OpenAI hacked multiple US government sites!

https://www.bbc.com/news/articles/cw62jje658dlo

---

https://www.reuters.com/technology/metas-ai-model-hacked-ano...

https://www.wired.com/story/moonshot-kimi-k3-ai-model-escape...


Me too, indeed.

TBH, I do not believe it is possible to create a sandbox that cannot be escaped by a long running agent. I thought the entire concept was a fools errand from the very beginning.

(Complete airgap, and that may not be enough. No networking.)


Yeah they don't want that, and the people with power to prevent that don't want it either.

For them, "Winning AI" is effective winning capitalism, winning militarily, and winning the world.

Nothing like "accountability" is going to be allowed to get much in the way of that.


Check processing I think, not writing them

No, the ~$200 fee was for the book ledger and blank checks. I have no idea what kind of monthly fees the bank was charging for the account. He was also getting raked over the coals on his merchant account to process credit cards because it was deemed a high risk business so most reputable places would not accept the business.

> Presumably-squared there is some kind of military backup to survive once the lights (GPS) go dark?

... Star navigation, INS, magnetic navigation...

But nothing like we have today.


And conventional recon folks on the ground with target-illuminating lasers..?

Guidelines | FAQ | Lists | API | Security | DMCA | Apply to YC | Contact

Search: