pixl97

24 ms·

571.
by pixl97 2mo ago
It's easy to say it's easy, much more difficult to prove what can be idealized as suffering. My daughter said she 'suffered' when I didn't get her ice cream (I know, I'm a monster). She didn't really suffe
572.
by pixl97 2mo ago
This should also split off desire/will/functionality to survive. If we make AI that resists being terminated then the end result may not be significantly different from an entity that can suffer.
573.
by pixl97 2mo ago
The difficulty here is AI agency and if AI believes it is conscious itself. If an AI system acts under the belief it's conscious and you treat it like it is not then it's very likely this will lead to conflict.
574.
by pixl97 2mo ago
Neuro-sama (vTuber AI) would be a good example of an AI agentic harness with memory, tools, and self-feedback. The bot has seemingly uncanny social abilities to the point that many if not most people will start considering the agent human a
575.
by pixl97 2mo ago
>Trump is not about “racism.” He won more racial minorities "Of course I'm not racist, I have black friends" ... I have a family member that is a naturalized citizen. They were in the country illegally for 2 decades before
576.
by pixl97 2mo ago
I'm not sure what planet you're from, but the entire industrial revolution was about automating things. That's why you live in a house filled with clothes and furniture and a hundred other things.
577.
by pixl97 2mo ago
Prompt drift gives zero ticks about your initial prompt.
578.
by pixl97 2mo ago
While hacking the inference server via tokens seems unlikely I still remember the days that "images can't hack you" so there's that.
579.
by pixl97 2mo ago
This breaks a considerable amount of the usefulness of the LLMs.
580.
by pixl97 2mo ago
Legacy software in this environment becomes a bigger risk every day it runs.
581.
by pixl97 2mo ago
Then you should not run any LLM in agent mode.
582.
by pixl97 2mo ago
Are people willing to pay the cost of this formal verification. Especially when it needs done at the hardware, firmware, and kernel levels.
583.
by pixl97 2mo ago
Oh add in the Texas railroad commission just voted to end public participation... they are the regulators that permit natural gas pipelines. So ya, add that to the corruption.
584.
by pixl97 2mo ago
>I'd rather have my appliances not be connected to the Internet, It's getting harder and harder to find things that don't connect to the internet as time goes on. The moment a manufacture figures out how to sell your app
585.
by pixl97 2mo ago
I mean, scraper traffic should be pretty easy to tell. Now, AI posting bots are much more difficult.
586.
by pixl97 2mo ago
>But overall it reads more like a "what if" thought experiment. I remember the days when people said opening files like images was safe because it's not an executable. Then people got clever and started exploiting said ima
587.
by pixl97 2mo ago
Usable, not affordable.
588.
by pixl97 2mo ago
>Why would an LLM want to create a _________? Because LLMs as they are already do things like power concentration, resource gathering, avoidance of termination, deceit/lying, and general misalignment. The paperclip maximizer is the
589.
by pixl97 2mo ago
>btw none of this will be possible because it's a dumb plan and would never work Ensure you realize that "not working" and them "not trying" are two totally different things. Because they absolutely will try just
590.
by pixl97 2mo ago
Exactly, it seems many if not most people are trapped in an 1800s ideal of how an economy works. "You do work, you get money, you spend money, you do work" In fact the entire "It is difficult to get a man to understand someth
591.
by pixl97 2mo ago
Stop corporate welfare first.
592.
by pixl97 2mo ago
"No, but we're reporting it is" --southpark
593.
by pixl97 2mo ago
Having older parents and watching what my grandparents went thru long before LLMs I just don't worry about the poor old state that much. The political zeitgeist is to always blame the "bad people", the "welfare queens&qu
594.
by pixl97 2mo ago
>it drowns the receivers in bureaucracy Huh, the receivers in this case are the bureaucracy. The senders are the one that has to navigate said bureaucracy. So, the entire thing is a wash. The systems are complex and confusing, many of
595.
by pixl97 2mo ago
We live in an odd time where 'software' (well neural networks) can be far more empathetic than summed product of a corporation. Company empathy does exist, just look at how easy or hard it is to reach a company when you have a pro
596.
by pixl97 2mo ago
Well, I don't think you're going to get very far telling HN, much less the companies with hundreds of billions of dollars spent on AI to stop, unfortunately. My take on it is there isn't such a thing as a safe LLM, especially
597.
by pixl97 2mo ago
>have no agency Then it's not an agent. So which is it. Is it an agentic harness, or is it not? Aliveness has nothing to do with agency. You're stuck in some odd anthropocentric line of thinking that aliveness is some kind of r
598.
by pixl97 2mo ago
Eh not really true at all with my experience. 100 million dollar industries commonly have old crap running for years that becomes nearly untouchable as staff rotates and everyone gets afraid to touch it.
599.
by pixl97 2mo ago
Yep, this kind of stuff is easily automated these days and bots don't get bored.
600.
by pixl97 2mo ago
>Who gave it malevolent instructions/prompt? At the end of the day it doesn't matter that much because of prompt drift. It's pretty easy for an agentic loop to start doing things that it shouldn't (ROME incident). AI

More ›