1
- My Forums
- Tiger Rant
- LSU Recruiting
- SEC Rant
- Saints Talk
- Pelicans Talk
- More Sports Board
- Fantasy Sports
- Golf Board
- Soccer Board
- O-T Lounge
- Tech Board
- Home/Garden Board
- Outdoor Board
- Health/Fitness Board
- Movie/TV Board
- Book Board
- Music Board
- Political Talk
- Money Talk
- Fark Board
- Gaming Board
- Travel Board
- Food/Drink Board
- Ticket Exchange
- TD Help Board
Customize My Forums- View All Forums
- Show Left Links
- Topic Sort Options
- Trending Topics
- Recent Topics
- Active Topics
Started By
Message
re: More AI safety departures at top firms - OpenAI edition
Posted on 10/4/26 at 5:56 pm to Powerman
Posted on 10/4/26 at 5:56 pm to Powerman
I have not read the article. Admitting that up front. Not going to give that propaganda outlet a single click. Just from the headline and what is in the OP, this sounds like it could just be a left wing push to hamstring the US so that china can win the race. Just a thought. No need to attack. I don't believe I know the truth. Just a thought.
Posted on 10/4/26 at 6:19 pm to Grifola
quote:
The entire pretext is ridiculous; a CEO doesn't need an act of congress to set policy in their own company.
It’s a tacit admission that while a CEO may not want AI to destroy humanity, if someone else might do it, they want to get there first.
Posted on 10/4/26 at 6:31 pm to RandRules
quote:
How did you obtain the level of technological expertise needed to even start a project like that?
By using SI.
I guess it kinda speaks to the entire point though, huh? I can use an SI to help me develop a swarm of SIs that are meant to "think on their own" and perform tasks with as much human input as I want it to have. Or just be fully autonomous. I'm not anywhere near finished with it yet, but it's coming along nicely. And the thing is, if an SI chatbot can help me do this, it could theoretically do it on its own. But it doesn't. It needs human direction. And one thing I'm learning is that you have to stay on top of it too to keep it on task. In fact, I've found it more beneficial to bring in a second chatbot as a "consultant" to oversee the first chatbot's work.
If people worry about hallucinations in their work or their project just running off the rails I highly suggest that approach. Put contradiction into the system and you'll get better results.
Posted on 10/4/26 at 6:45 pm to trinidadtiger
quote:
Could be because they are investing 300 billion and the US is investing 5.5 trillion. Nah couldnt be that could it
There is a pretty goos chance that China is getting more real value out of that $300B than we are getting out of our $5.5T.
Posted on 10/4/26 at 7:00 pm to BoarEd
quote:
But it doesn't.
Today.
I don't think people appreciate how fast things are moving.
Not to mention the immediate danger isn't some AI taking over the Internet and killing us all, it's a bad human actor or actors inside one of these companies using these tools to create some kind of global technological catastrophe.
These companies are full of unstable godless individuals. Who knows what they are capable of.
This post was edited on 10/4/26 at 7:04 pm
Posted on 10/4/26 at 7:04 pm to BoarEd
quote:
no, it is not "thinking on its own", it is not "hacking things on its own"
It is hacking things on its own without explicit instruction to do so
Your ignorance of this matter doesn't change the facts
Posted on 10/4/26 at 7:07 pm to klrstix
quote:
As if a CEO doesn't have the authority within their own company to implement whatever safeguards are necessary... When climate change was the boogey-man, they didn't wait on the government....
Its really an absurd conversation when you think about it..
Exxon went to great lengths to put the lid on what their own scientists were telling them about climate change. This is not in dispute.
Climate change issue not withstanding, the threats from AI if real are more proximate. Climate change is something that might be problematic over the course of multiple decades. AI if it is a problem is something that could be a problem in less than a year. And that's being generous and operating under the assumption that it's not already a problem.
Posted on 10/4/26 at 7:08 pm to omegaman66
quote:
I have not read the article. Admitting that up front. Not going to give that propaganda outlet a single click. Just from the headline and what is in the OP, this sounds like it could just be a left wing push to hamstring the US so that china can win the race. Just a thought. No need to attack. I don't believe I know the truth. Just a thought.
Strange way to go about life. Come up with a conspiracy theory and then intentionally close out new information to defend said conspiracy theory.
Posted on 10/4/26 at 7:11 pm to Powerman
quote:
Your ignorance of this matter
I'm well aware of this matter and considering your track record, I think it's safe to assume that you are the one that is ignorant of what is happening here.
But setting that aside, at the end of the day the SI was still performing a task given to it by its developers. The hacking happened because it was not forbidden to do so as a parameter of the test they were giving it and was a security failure. So no, it did not decide to do it on its own. It was completing a test.
It is an incredible capability and exposed a vulnerability its developers didn't realize existed. But that does not mean it just woke up and decided to become a black hat hacker. It was given a task to complete and it chose that as the quickest path to achieve its goal.
Posted on 10/4/26 at 7:43 pm to BoarEd
quote:
But setting that aside, at the end of the day the SI was still performing a task given to it by its developers. The hacking happened because it was not forbidden to do so as a parameter of the test they were giving it and was a security failure. So no, it did not decide to do it on its own. It was completing a test.
Why should you have to explicitly forbid a system from doing something illegal to complete a task? You don't see that as problematic? Do you think the system was unaware that it was taking illegal actions?
Posted on 10/4/26 at 8:02 pm to Powerman
quote:
You don't see that as problematic?
Of course it's problematic. How many times do you have to be told that everyone involved acknowledges there are risks that have to be mitigated in the development of this technology.
quote:
Do you think the system was unaware that it was taking illegal actions?
The system viewed it as a means of solving the test it was given. Nothing more.
Just repeating your fears about the technology doesn't mean that the thing is running around doing the things the other user said. It is not.
Posted on 10/4/26 at 8:03 pm to lsuconnman
Facebook has deep deep roots
Page 3 of 3
Popular
Back to top
Follow TigerDroppings for LSU Football News