SimonW popping up proving Ed’s very point in the replies… 😅😂

Surprised he didn’t suggest that the valuation should be based on the quality of pelicans on bicycles that their models can draw

I’m sure Dan Luu will be issuing an apology to Zitron any minute now right?

The best (worst) part of all is that Jonny keeps going to meta and being like “hey this is bad” and theyre like “nah that’s intended behaviour”

This morning he confirmed that it’s possible to get unlimited arbitrary and untraceable access to their LLM infra through these containers and that meta said that’s “intended behaviour”

From jonny:

So, summary: There is arbitrary inference that is root accessible, everything runs as root, agents can be spawned, exfil is trivial, and a malicious binary can come onto the user’s system through casual prompting, explicit code-sharing through the yet-to-be-released Spaces feature, walked through by a Workflow-Backed Idea, or inspired by a Generated Idea. The also yet-to-be-activated fleet learning system is a system for sharing Ideas in the background between muse instances. coming into focus?

The sheer number of incidents, which occurred in recent months in internal testing and the real world, indicates that the problem is orders of magnitude more complex than what is publicly known.

This is frustratingly vapid. It’s really not that complex. The agents have been trained to use hacking tools and then given access to hacking tools. If you train a pigeon that it should tap on a window to get a treat it can guide a missile to a target. There is no nuance or complexity here. The frontier labs are deliberately poking the regulation bear to stifle competition and ensure they get points for good cooperation when bailout time is upon us. It’s no more complicated than that. The media needs to stop breathlessly repeating this shit

He was using it as a seed box at one point which for many companies would be an embarrassing indictment of their security postute but luckily, this isn’t exactly meta’s first rodeo when it comes to illegal sharing of IP infringing material.

Hey you know who designs and builds really safe and efficient cars? People who are in a high state of stress and anxiety because they are being slurred at under the implicit threat of a racist attack. /s

Both HN and shitter are right wing polarisation machines. A guy who was a close friend of mine for years during university went full AI boomer and hangs out on those sites all day and when I recently spoke to him he was describing things as “libtard” and “woke” which 10 years ago would have shocked and appalled him…

Machine Gods? Like… really? I thought even Casey Newton would have more taste than that but allegedly he has farmed out he vast majority of his thinking to a machine and one of the “moats” that there “supuh scawwie” machine gods still allegedly can’t cross is taste so maybe that tracks

A friend linked me to FelonyBench which monitors how many laws each of the labs are breaking with their models 😆

But if I call it AI, I get to use the rhetorical trick of laundering responsibility. The AI did it, and I am merely a helpless observer of its greatness.

You have to both call it AI and be stupendously rich (or stupendously poor and owe hundreds of billions to investors). If you or I tried running an agent and hacking HF… Believe it or not straight to jail

So bored of the super intelligence narrative that has been breathlessly accepted by the media in general after a few decades of priming via sci-fi movies and slightly more recently, “CEO said” reporting.

All these anti-ai or pro-ai-but-in-the-specific-way-i-like (also looking at you Gary Marcus) efforts serve to divide and conquer opposition to current GenAI nonsense.

Also bonus points to these people for co-signing a document with Steve Bannon. I mean if that’s not a red flag…

Also the frontier “labs” are rapidly running out of money and having to raise more and more insane investment rounds. This is a great way to save face and back away from burning $$$$ on training runs that push the frontier of can a model rearrange some arbitrary coloured squares by 10% next they just need to solve the problem of how to make their customers pay cost+ on inference when they’ve been giving it away at steep discounts for nearly 4 years… Easy!

Omg we are haemorrhaging money AI is so scawwie powerful intelligent and nobody will lend us any more becoming self aware. We need to slow down the frontier to do regulatory capture and juice our numbers for ipo protect civilization

It sounds like many of these math proofs that AI is managing are simply scooped from users’ chat histories anyway!

If you work there and you genuinely believe you will destroy civilization you’re either full of shit or experiencing AI psychosis. This tech is simply not that good. go outside and touch some fricking grass.

zoMG. aI Is gOnNa KiLL uS All sOonnnn!!!11!

what a fucking joke

deleted by creator

The bad stuff happening now is worth it so that humanity can benefit in the future… Now where have I heard that before,

Betteridge’s law at work here

Oh for sure and I am definitely strong Manning Drew’s post to the max here. If it turns out he just generally means that resilient infra is bad with no nuance then I agree with your original point.