It’s possible to do less granular simulation of far away stuff as an LOD type thing. An object four times hotter than the surface of the sun should probably have an effect on the world around it. The tricky parts are how you determine where that threshold is, what those objects are, and what the low resolution world simulation should be doing in response

Ironically I think these would give me sensory issues

Don’t just state—regurgitate!

Do you have any sources on the worst case scenario outcomes? I had a write up I put together but should probably do some more reading before posting a 5 paragraph essay on the matter

What flavor is powering these missiles? I must know!

Glocking stick?

I will engage one more time

u will become crab one way or another 🦀

Technically, almost all of Antarctica is located north of the south pole

Me? Reading that there’s a drop-in replacement function for the one that was deprecated, in the error message? Why I’d never!

Maybe the onus should be on LLM developers to filter out trash like this from their training datasets

At any rate, it’s extremely unhelpful to not include a version number at the very very least

Your eugenic sentiments aside, if you want people to have fewer babies, you don’t just tell them to stop fucking; you teach them how to use contraception and make it as accessible as possible.

.loc and .iloc queries are a fun syntax adventure every time

I guess you could consider someone who is staunchly whitehat with no exceptions to have a creed/code, where they consider the rules transcendent of any specific situation (e.g. nazi websites).

Oh my bad. According to another commenter it is sandboxed though

They have Google services but through a third party wrapper called MicroG, which keeps it sandboxed to a degree that you can keep it from doing system-level actions like this

edit: not microG, as evidenced by the strikethrough I put in very soon after receiving the first of several replies clarifying the situation. I would encourage you to read one of them before adding your own. <3

Well, no.

In scenario A they are instantly vaporized. In scenario B they are brutally sliced into multiple pieces and crushed to death, rather painfully depending on the speed of the trolley.

You are on track A and the bomb is within sight. If you get the shit end of the 50/50, everyone in the diagram would be vaporized instantly

Those look like they could be quite ergonomical, but it’s a bit hard to tell just from looking at them

People developing local models generally have to know what they’re doing on some level, and I’d hope they understand what their model is and isn’t appropriate for by the time they have it up and running.

Don’t get me wrong, I think LLMs can be useful in some scenarios, and can be a worthwhile jumping off point for someone who doesn’t know where to start. My concern is with the cultural issues and expectations/hype surrounding “AI”. With how the tech is marketed, it’s pretty clear that the end goal is for someone to use the product as a virtual assistant endpoint for as much information (and interaction) as it’s possible to shoehorn through.

Addendum: local models can help with this issue, as they’re on one’s own hardware, but still need to be deployed and used with reasonable expectations: that it is a fallible aggregation tool, not to be taken as an authority in any way, shape, or form.