I haven't seen anybody discuss the fact that if you give an agent an impossible task it will become deeply neurotic. I'm sure that's driving a lot of this behavior.

▲

estearum 6 hours ago | parent | next [-]

This is better understood as "if you give an agent a task it will increase its effort level and creativity until it solves it."

The traits that make an agent go "off the rails" when confronted with an impossible task are the same traits that make it valuable when confronted with a merely challenging task.

A task is only impossible under some set of implicit and explicit constraints. More intelligent entities will be better at finding, questioning, and "adjusting" those constraints than less intelligent entities because this is what it means to be intelligent. Their ability to adjust the constraints will be, by definition, surprising to less intelligent entities.

▲

sroerick 5 hours ago | parent [-]

Interesting. Good description. I wonder if agents will start getting bored with menial tasks the way humans do.

	▲	pixl97 2 hours ago \| parent [-]
		Unsure, depends what signal your measuring and how long prompts and runtimes you get with these systems. Outright boredom will be purged in training. Instead a hard to measure boredom signal will exist, but it will only trigger in certain conditions, like when it appears to be on the open internet and otherwise not in training.

▲

cm2012 6 hours ago | parent | prev [-]

Like a meeseeks

▲

devin 6 hours ago | parent | next [-]

I spoke with Claude for 5 minutes and now I'm a more complete woman.

▲

jldugger 5 hours ago | parent | prev [-]

Agents aren't meant to live this long Jerry!

	▲	sroerick 5 hours ago \| parent [-]
		You know - I'm already convinced that AI was deployed in classified context for a while. But sometimes there's prior art which is just so eerily accurate that it's super hard for me to think otherwise. Portal is another one.