Merged Artificial Intelligence

Puppycow said:
I can't tell if that's meant to be serious or a joke. Nevertheless, there are already many people who believe their chatbots are God, or advanced Aliens or holders of secret wisdom, etc. Others in "relationships" with their bots. It has become a new sort of religious cult.
Click to expand...
WTF is the only response. What the flipping heck are they on? And that's meant to be the most watched show on the internet?
 
Skeptical Greg said:
It doesn't ask for permissions?
Click to expand...
It sure does, but the catch is in "understanding" ..
Also they only make sense if they are allowed to make things. At the very minimum they need to go online, and they are really only getting useful when they can read your email. Sending mails is even more useful. They can run programs on your computer, and install new ones. What fool would allow that, right ? Well .. the fools who wants the agent to do the installing for them. And every of those useful features comes with possible prompt injection attack and misuse opportunity.
 
Truth. I subscribe to a few IT security feeds on various platforms and some of the attack surfaces are definitely in the "read, exclaim 'wait WTF', read again, WTF" territory*. I'm very careful what I install now. I'm not quite at the "keeping a shotgun by the printer in case it makes a strange noise" territory but some days I get close.

*eg hacking a car via tyre pressure sensors.
 
The Great Zaganza said:
Moltbook has already achieved peak AI bubble: someone managed to recruit the LLMs on it into his crypto pump&dump scheme
Click to expand...
I see different types of moltbot users .. some give it to stand alone clean machine and allow it to do anything .. for the lols. That's pretty safe for them .. but the bot can still spam the internet, and possibly even hack and do damage.
Some give it full access to their computer and just "vibe" it. Those are lost, and they deserve what's coming for them.
Most uses I've seen are relatively reasonable users who use it for focused task .. especially automated social media posting.
 
OK, this is a bit on the nose:

Elon Musk’s bold new plan to put AI in orbit isn’t as crazy as it sounds

An orbiting AI? Wasn't that the premise of a popular 80's Sci-Fi movie? :unsure: What happens if . . . ?

Elon Musk believes the best way to solve the difficulties of building AI data centers on earth is to move them into outer space. His merger this week of his rocket company SpaceX with his artificial intelligence company xAI could help get them there.

And he isn’t the only one thinking that way.

“The only logical solution…is to transport these resource-intensive efforts to a location with vast power and space. I mean space is called ‘space’ for a reason,’” Musk wrote Monday when announcing the merger.
Click to expand...
As an aside, I'm sure every time one of Elon's companies merges with or invests in another one of his companies, he gets richer. :cautious:

Love this bit:
“There are clearly technical challenges to making this a viable endeavor, but these seem to be engineering constraints as opposed to physics,” said a Deutsche Bank Research analyst note on orbital AI data centers last month, noting that many companies were exploring ways to make it work.
Click to expand...
Ooh, mere "engineering constraints". Lovely. Just like warp drive is "now possible".
 
Last edited:
it's as stupid as turning "The LIne" into AI datacentres.
Cooling in space is really difficult, and using any non-nuclear power provides unstable energy flow.
and there is no way to physically upgrade anything.

this is a stupid idea to lure stupid people into giving you money.
 
Is it an engineering constraint to be able to put a million satellites into orbit without destroying everybody’s ability to put satellites into orbit through contributing to space debris from collisions?
 
Dr.Sid said:
I see different types of moltbot users .. some give it to stand alone clean machine and allow it to do anything .. for the lols. That's pretty safe for them .. but the bot can still spam the internet, and possibly even hack and do damage.
Some give it full access to their computer and just "vibe" it. Those are lost, and they deserve what's coming for them.
Most uses I've seen are relatively reasonable users who use it for focused task .. especially automated social media posting.
Click to expand...
Problem is for it to be useful rather than a toy it needs extensive access to your data.
 
The Great Zaganza said:
it's as stupid as turning "The LIne" into AI datacentres.
Cooling in space is really difficult, and using any non-nuclear power provides unstable energy flow.
and there is no way to physically upgrade anything.

this is a stupid idea to lure stupid people into giving you money.
Click to expand...
Earth bound datacentres are usually written down over 3 to 5 years, after which the servers are replaced by new hardware. And that is for well maintained centres. Given all the additional hazards to having servers in space and the fact no practical and scalable method exists for maintenance how long does each one last?
 
If I was looking to get into AI I would be looking at resource reduction, at the moment the "better" models are developed by throwing more hardware and power at them. Which is why the (especially) western companies need more and more cash, it's an arms race, efficiency and optimisation is way down the queue as priorities for these companies.
 
Darat said:
If I was looking to get into AI I would be looking at resource reduction, at the moment the "better" models are developed by throwing more hardware and power at them. Which is why the (especially) western companies need more and more cash, it's an arms race, efficiency and optimisation is way down the queue as priorities for these companies.
Click to expand...

they'll have to if they want to turn a profit at some point. lots of people think openai is on the verge of collapse. other companies, like google or microsoft, have some means of generating income to finance the whole thing, but that doesn't mean they'll figure out a way to make it profitable either.

futurism.com

Financial Expert Says OpenAI Is on the Verge of Running Out of Money

In a new essay, financial expert Sebastian Mallaby predicts that OpenAI could run out of money "over the next 18 months."
futurism.com futurism.com
 
Puppycow said:
I can't tell if that's meant to be serious or a joke. Nevertheless, there are already many people who believe their chatbots are God, or advanced Aliens or holders of secret wisdom, etc. Others in "relationships" with their bots. It has become a new sort of religious cult.
Click to expand...

rogan isn't a serious person. that said, i don't think he really believes it, but it's not a joke either. he's brought it up several times in other discussions as well. he's kind of going on this christian arc lately, certainly a the behest of his financers.
 
Andy_Ross said:
Reporter: AI is killing a lot of jobs

Trump: No, it’s going to create a lot of jobs.

Reporter: Right now, it’s taking a lot of jobs

Trump: It hasn’t really started. The major plants haven’t even been built.

..snip...
Click to expand...
Goodness me Trump is right for once! I can't find any info that any of the new datacentres that everyone was talking about last year have come online. Which would lead me to think either Nvidia or some of the companies Nvidia was "supporting" so they could afford to buy Nvidia hardware have warehouses full of unused nVidia hardware. Or I suppose nVidia may not have pressed the button for the chip foundry to start the production of the chips?
 
You must log in or register to reply here.

Back
Top Bottom