dpoloncsak

If I clean my gun (tool) while it's loaded (stupid idea) and it goes off, who's to blame? The 'stupid user causing the problem', right? I personally wouldn't blame the gun...

If it falls into the wrong person's hands, it's STILL my responsibility as the owner.

If you're not going to take time to learn to use and be responsible with the super sophisticated and all-powerful tools, don't play with them. I'm not arguing for the death penalty every time someone makes a mistake, but I think it's very important to accredit responsibility and blame correctly. We've learned these tools are potentially as dangerous as a loaded gun. Be responsible.

Why draw the line at the person with the brain capable of free thought, instead of their great great great grandparents? Am I understanding your question correctly?

Legally, we draw the line at the human causing the action (Insanity pleas withstanding). I don't see why we treat this any differently.

>99998 years ago when it became legal for an AI to own an LLC

There's no way to provide a good faith rebuttal here. My entire argument is an extension of "LLMs can't be held accountable, so they must never make decisions". If governments start letting them own LLCs without a human in the middle, we're in more trouble than "Who do you blame for this shitty code" or "Who's responsible for this compromise"

I don't think so...? Maybe I'm understanding the article wrong? I get the overlay, but I'm being specific to LLM chains... They require something to kick them off, they do not act autonomously. A human does something, types, speaks, presses a button, whatever. A Cron Job firing off is the result of a prior prompt, by a human.

I don't follow the chain further back than that, as I believe humans have free will. I get that that's debated, but thats why I draw the line at human action.

Context overload can cause strange results, yes. Hence why a HUMAN needs to be held responsible for the output of their tools.

EVERY breakout that's hit mainstream news has been because of a single 'Security Firm', Irregular. Maybe I'm unaware of some less-headline-grabbing ones, but they all seem to stem from being 'unaware the environment wasn't sandboxed'

I think you go back to the original instance, yes.

In your example, who's paying for it? Whether by providing the hardware + power or paying a LLM service. Whoever is paying the maintenance cost is responsible, in the event of your demise. These things run on physical hardware owned by someone at the end of the day, it's not a deity in the atmosphere.

You're starting the autonomous harness, you're responsible for any output it provides. I don't get how this is a foreign concept.

If I jump out of a moving car that I'm driving, I'm not suddenly absolved from damages because "the car did it"

Right. Even when an LLM prompts an LLM, what prompted the first one in the chain?

> Why does it matter so much whether there was a human who typed into a chatbot or another LLM invoked a sub-sub-agent?

Responsibility. Someone needs to be held responsible for any damages done, plain and simple. You can't take an LLM to court, you take the prompter. Asking an LLM to ask a sub-agent to break the law can't suddenly absolve you of any wrong-doing.

>If I now tell a machine.....

You/your estate is still responsible, or atleast whoever is paying for the power for the machine, or renting the space in a data center...whatever.

Sure, but it's still not skynet-level 'the AI just started doing things'. It does what it finds it needs to do to achieve the goal defined in the prompt.

It's very important to not personify these tools and remember that the tools are acting on behalf of real people. In the same way the AI didn't 'go rogue and hack HuggingFace'. It was an oversight made by a human.

Built by whom? Acting as an 'agent' on behalf of whom?

An LLM can prompt an LLM when first prompted by a human.

I think OP is trying to convey the idea that LLMs do not take initiative to do anything, and these are not 'beings' capable of doing things. These are tools being used by humans.

Everyone gets 'got' once

I remember a while ago the discussion kinda steered from "The leading LLM provider will be the one with the better model" to "...will be the one with more user history".

Tools like OpenClaw and Pi seem to remove that from the equation, letting you keep your 'history' and customization while using whatever inference provider you want. If Muse takes off, which seems to be built upon or atleast arch'd similar to OpenClaw, I think we'll see the rise of on-device harnesses.

This would further the efforts to "resist all attempts to.... lock them into your service, and will drop you if your competitor makes a 3% better product, and will reject every upsell for actually profitable services.", imo. In a model-agnostic harness all you care about is speed, accuracy, and price.

Yearly audits seem fair, no?

>We've seen a lot of companies pay off ransomware gangs rather than invest in the kinds of robust backups that would make recovery possible/less painful than rewarding the hackers.

Experienced Ransomware gangs will set the price target as something high, but not cost more than the price of being down a few days while you rebuild, making paying them seem like the most cost-effective solution

It's a zero day that has been released to the public with full exploit code.

You're missing the important part there.

Throwing it out there, YouTube is top-of-their-class and ran mostly from ads

There are a lot of people who would say that is exactly what Reuters is. Any American news outlet, really.

Obligatory relevant xkcd: https://xkcd.com/538/

Sorry, I may has misphrased what I meant. Can I use this with a local model? Or is it locked to Claude

...with the risk of getting banned, you can't reasonably use this for anything 'important'.

In order to make my life easier, it probably needs to be doing important stuff.

You can't hide behind 'other apps are doing it too', if my 'important business account' or whatever gets banned from using your tool, I'd be upset at you.

How do you guarantee your uptime SLA if my account may get banned any second?

This can be used locally, no?

It's well detailed and easy to follow.

If you're targeting beginners, do you think it's valuable to explain the .env concept a bit, and generally what an API call is? Or do you think your average reader is familiar enough with Python that they would have come across these concepts already, and are just looking to prompt LLMs?

I would assume, if I'm reviewing the cameras for a valid reason, that it may end up in court. If it ends up in court, I'd want my logs to show I atleast did the bare minimum, yes.

Just because there isn't a coded requirement doesn't mean there hasn't been directions to state valid requirements. This isn't a Jira ticket, this is a tax-payer funded government official supposedly doing his job, and should 'bother' to give valid reasons.

Do you need to predict all the ways the model might misbehave? Your 'hack everything you see for our internal research lab' agent should be airgapped. You don't need to come up with every reason why, one is enough.

If you're working with these companies, you should reasonably be able to get the code to perform offline audits. If you can't get the code, you probably shouldn't try to pen test it.

Understood, and .name isn't being used enough in business to worry about 'the effect it will have on my users'. Just pointing out that it doesn't work like the 'norm' (although I guess it's not quite as unique as I thought, either)

I appreciate this, and yeah the government/education ones slipped my mind, but I stand by the fact that the reason a list needs to be kept in the first place is because this is the edge case and not the norm.

The fact that multiple organizations need to keep a public list of known 3LDs proves it's the edge case, does it not?

"Here's a list of things that look like subdomains for you to treat as 3LDs instead of subdomains" sounds exactly like the solution to an edge case to me.