Their kids are in the 70s and 80s too

It's funny when you boot up these old games you played long ago, how small the view port is, and how I didn't seem to notice that at the time.

To avoid the comming massive risk to human life from climate disruption, I think there needs to be an Australasian / ASEAN open boarder arangement like the schengen area.

I heard this is a common tactic in news forum discussions too. When you don't want people talking about the substance of the post, then you mention something like how the site is hosted on wix which supports the IDF, and derail the conversation.

It doesn't work that much on sites which organize comments in a tree, like HN or Reddit. You can just fold that part of the thread.

Flaging and killing the comment doesn't help dissuade form this lol

How do you test and monitor outbound access against program that adapts itself to get around things? if your MITM and filtering keywords etc, cant it just .. encode it traffic somehow or another.. If you're looking at traffic volumes, cant it just go slow.. If you have strict ACLs, we've already seen in the HF case, traversal from an intermediate system..

If you can't tell bytes are leaving a node, you probably shouldn't be selling security services or testifying to congress you are taking the lead in AI security

Use a basic firewall? Not a single outbound byte should leave the machine, except inside a virtual network towards in-scope test subjects. That's not going to be infallible because hypervisor exploits still exist, but it's the lowest bar and they failed to even meet that.

> If you have strict ACLs, we've already seen in the HF case, traversal from an intermediate system

- The intermediate system shouldn't have outbound access to the internet

- You should ideally be using a proxy that filters the set of endpoints that clients are allowed to access to reduce the exposed surface area.

It's odd to find out that I use a higher level of isolation in my unimportant home network to stop IOT devices from doing funny things to HomeAssistant than big AI labs use to keep their possibly-world-ending AIs contained.

I know that the people working there aren't idiots so the most likely explanation is that the incredibly weak security was intentional because its inevitable breach would make for great marketing.

> I know that the people working there aren't idiots so the most likely explanation is that the incredibly weak security was intentional because its inevitable breach would make for great marketing.

It's more likely to be different specialisations. Most of the people doing the evaluations are more data sciencey ML type people, rather than software engineers. This isn't helped by their culture which is very much driven towards alignment as the only possible solution to super-intelligence (which may be true, but I have my doubts that this will happen in any reasonable time frame).

Oh of course, but data sciency ML type people aren't [supposed to be] tasked with of operating and securing the infra, that would make it even worse.

These labs have teams of dedicated security engineers and infrastructure engineers, that's why it doesn't add up for me.

> Oh of course, but data sciency ML type people aren't [supposed to be] tasked with of operating and securing the infra, that would make it even worse.

But AI collapses the need for expertise, right? So the DS people can now do security at the level of the best in the world, right?

Some form of this issue is playing out in lots of businesses right now, and I'd be very surprised if it wasn't happening in the labs.

I suspect the lab's infra people are focused on inference, rather than evaluation. I agree that this is insane, but it seems like the most likely explanation to me.

> How do you test and monitor outbound access against program that adapts itself to get around things?

Literally what the industry has been doing since public networking is a thing. Adapt yourself.

>How do you test and monitor outbound access against program that adapts itself to get around things?

Current generation LLMs don't do that. Also you can just prompt it to not hack systems on the internet and the model follows that instruction.

You can also add a simple classifier to internet requests. The classifier model can obviously be as intelligent as your hypothetical rogue model.

Nothing you said makes any sense in the context of LLMs or the current news cycle.

Air gap?

So just enable access to a completely offline, cached, and transparently proxied, copy of the internet.

Isn't that essentially what they trained with, anyway?

I think that was the plan. But OpenAI hacked the proxy

There has to be a route to where the LLM is running, and if there's a route for that there's probably a way for the machine to use it to route traffic somewhere else.

No, just allow the firewall to access the LLM endpoint, nothing else. And protect the firewall settings from the agent.

There doesn't have to be. You can easily block all outbound traffic, or the VM can exist without any network interface at all - you can control it through its host using a serial console.

I had much better access to personal computing as a teenager with multiple clunky desktops, that I do today I feel.

That seems like a weird claim to me. What did you have then that you don't have now?

Is it because you're using an OS like Windows or Mac that are very closed and don't encourage hacking? There's an easy solution to that.

(Note that I'm not taking about things like the age verification laws, which haven't had a meaningful effect anywhere yet.)

I give todays Turdle 1 for difficulty.

Looked like a two to me.

I heard CO2 is a gas with a powerful green house effect. What happens at the end of life of these units I wonder? What if they leak and it gets into the atmosphere!

So, in the US, for instance, the average person leads to the emission of 16 tonnes of co2 per year. The amount of co2 in your domestic heat pump (which… doesn’t weigh multiple tonnes) is more or less immaterial, even if it does leak.

> I heard CO2 is a gas with a powerful green house effect.

Compared to other refrigerants it's a paradise:

* https://www.epa.gov/hfcs/technology-transitions-gwp-referenc...

* https://en.wikipedia.org/wiki/Global_warming_potential#Calcu...

Maybe but at this point I no longer think it's fair to be quite this dry on something that quite so many people would be legitimately sincere about anyway so I wouldn't exactly apologize for being the victim.

> Maybe but at this point I no longer think it's fair to be quite this dry on something that quite so many people would be legitimately sincere about

Yeah, but I'm from the UK so I should be able to detect dry humour. It's something that Americans don't really get.

You know, like affordable healthcare.

I apologise, I am at the dry humour stage of climate grief.

Take heart in the thought that this is going to ease off a bit in a couple of years when the current solar cycle dies down and the sun is not dumping about 5% more energy into the atmosphere than normal.

I was waiting for it to .. say usage limit reached after reverting it back to how you started..

It reaches the usage limit after adding a cookie banner above the button