Potential 6 month warning until the end

[Frombeyond]

So, we potentially only have 6 months until the end of civilised society and mass deaths. I’m terrified and sad

AI staff 'genuinely frightened' for humanity's future, ex-Anthropic researcher tells BBC www.bbc.co.uk/news/articles/c1kx0gyje9wo

At least that means it's okay to start planning for Christmas.

i read it’s 2030 new earth will come, everyone laughs at me but at least I’m awake, they are poisoning us at every opportunity, our bodies are rife with parasites, we deworm our pets but not ourselves. The elite eat babies. The world as we know it is f*cked

Haha I understand the topic fine thanks. My response was to you saying we can switch off rogue AI rather than to the OP. I’ve responded to the OP as well, I also don’t generally agree with the fearmongering, as it happens. But “we can switch it off if it goes bad” is not a good solution to the fears being raised because it is in general a hard problem to identify a rogue AI instance. A lot of damage can be done before any human even knows it’s happening. That’s how the previous cyber attacks have occurred, it was all
hidden until after the fact. Open ai could have turned off their machines before the agent hacked hugging face except they didn’t know it was doing that. So they didn’t. So the bad thing happened. When I said “switch off LLM” obviously that’s shorthand for switch off the machines hosting the model. Not sure why that’s confused you.

Yeah I moved freely between talking about agent / AI / LLM because really in this context the distinction isn’t that important. It is the LLM you’d have to monitor, that’s where the stochasticity and therefore the risk lies. Agent is just LLM(s) plus prompt(s) plus (usually deterministic) tools and it’s the LLM part that creates the risk.

Maybe if you clarify what you mean by “an AI instance” we’d have a more fruitful discussion.

Good to see this from someone who works in the field. I tried to say this earlier on in the thread and have been totally ignored. This thread worried me when I saw it, so I’ve done some reading and spoke to others online who have explained it the same as you, it’ll be humans that cause ai to destroy anything by not putting those safeguards there. There are some real doom mongerers on this thread whipping up a frenzy.

What do you mean, poisoning us? Who are and how ?

What does new earth coming mean?

Yeah, I'm not sure what you're talking about, but do you know what a large language model is? You're so the large language model is the output. It's the creation and the calculations of I guess the connections of how the different requests or data structures are quired Artificial Intelligence is all kind of stitched together. You wouldn't monitor the model itself, you would monitor the input Also, you said it's difficult to know how people are utilising Artificial Intelligence. Actually, it's kind of trackable down to the data warehouse, and the racking that it would be run off. It really is simple to kind of resolve this. If it became a problem. What's not simple is dealing with the humans that may be doing things with Artificial Intelligence, but that's a completely different problem. And it's not what the op has expressed. On a side note, love this conversation in mum's net.This is awesome.

Oh, ah, to the poster. I'm responding to it confused me because I'm extremely tired funny enough. I'm having a nightmare of a moment working through this bloody instance to get a piece of work done. It has to be breadcrumbed like a child, and then it confidently asserts stuff that is completely false, and when I do a Q&A it admits to its fault. But then it does it again in the second run.I don't add the third i'm so tired.I was so tired and I need it done before tuesday

I really hope you’re not talking to her about all of this stuff. It’s hard enough being a child without managing a parent’s anxiety too.

There’s already robotic beings in a local eatery

https://www.facebook.com/share/v/1JxbAe5r5N/?mibextid=wwXIfr

I think it’s quite cute…

As much as I think we shouldn't laugh off the dangers of AI... I think you might need to have a lie down. Oh, and step away from QAnon.

What a bombastic thread title. Take a walk outdoors, live in the moment. There's really not much else to do.

Won't be so cute when its eyes turn red and it's wielding a machine gun 🤣

I’ve seen more intimidating versions (not customer service) but still, that big ol piece of kit takes the job of a 19 year old adult dc.

Or scanning peoples brainwaves and sending our thoughts to the Deep State 😂

The world could give itself a month or so to work out how to manage without internet ( as we all did not so very long ago) and then unplug everything to do with it !

I watched a programme last night about this..

A quote that has stuck with me…

’Our children will never be smarter than AI’.

Now that is a little concerning

Oh yeah, remember these? 😎

A month or so 😂

Yeah I know 😂 but if push came to shove 😂 we lived quite happily without it. Time to get pens paper , tyoe writers and filing systems out and buy shares in the post office and

OP you are not being ridiculous and I too am concerned. I'm sure it's making me feel a little down about life for my (young adult) children. They are both intelligent but I listen to them planning and looking forward to the same things I did at their age, holidays, having children, careers etc.. and they are just in that blissful mid 20's bubble where the news is not their main concern. I think people are being extremely insensitive towards your post to be honest. I'm with you in being realistic and concerned about that may happen.

God forbid my handwriting had me in detention year 7 through 11

Yep I’m old enough

My handwriting is AWFUL it used to be ok but we barely use pens now !

Honestly sometimes I we should just unplug the bleeding thing. There are positives , lots of them but too many serious negatives now