>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.

CORBA had the concept of distributed objects which is a terrible idea in practice.

You can't know if he is an outsider so your argument runs into nothing. If you kept explaining why AI won't remove the CUDA moat and doubled down on that, then swerner would have kept silent, meanwhile your comments would stay so your current strategy is one blunder after the other.

>There's nothing deterministic about weather, so hopefully you're not just being disingenuous.

You're the one being disingenuous. Look at what you wrote.

> Is AI less deterministic than an airline dealing with weather?

You didn't talk about how deterministic the weather is. You talked about how an airline responds to a weather event in comparison to AI, which means that it's about responding to presented information by making a decision.

The state of the weather does affect your micro decisions, but the rules you follow are the same every time and the rules are as deterministic as possible even if they rely on pilot intuition.

> rules are as deterministic as possible even if they rely on pilot intuition

I think you are profoundly confused.

An airplane, and an AI algorithm encoded into silicon, are inert physical objects.

Every evaluation we are doing here is of a complex system that involves the interaction of people and machines and physical connections and so on.

An aviation system connected to every country and region with millions of people and machines involved is no less complex than what’s under discussion here and no more deterministic.

We don’t regulate airplanes because they aren’t people. We regulate pilots and mechanics and leaders of the companies that make and own them.

The task at hand is to regulate the people involved in AI to get the outcomes we want.

>What sort of robot would you need to load and unload a dishwasher, collecting and storing the crockery etc.? What sort of robot would you need to load and unload the laundry, fold and store the clothes?

https://www.inceptivemind.com/toyota-gantry-robot-hangs-ceil...

You might now counter that legs are easy and a solved problem, sure, I won't deny that, but the point is that you should try to think out of the box.

Correct, when you think about it, all the "world is designed for humans" statements inherently assume that humans are still needed for the specific task to be automated.

For example, we bolt robot arms on tables, so that the work space is at a comfortable height for humans, but if the robot arm had the exact same intuition and intelligence as a human, you could just install it onto the floor. Suddenly the entire rationale, the table, which is the representative stand-in for "the world was designed for humans" is gone.

It's not that we necessarily need a specialized machine, it's that a generalized machine doesn't have to be humanoid.

This is something people in this discussion don't seem to understand. It's not the LLM that is cheating or misaligned, it's OpenAI and Anthropic.

Because that means you won't get free PR from having your models run out of control.

Why does lawlessness create lawlessness? We will never know.