What wage would they have to pay you to pick berries? I suspect it's cheaper to automate these jobs (developing the tech to do so as necessary) rather than raising wages so US citizens take the job. These really aren't desirable jobs by US standards, and automation is already underway.

Edit: I acknowledge that you're just explaining a viewpoint and you don't necessarily hold it.

Current labor costs are too low preventing the necessary investment in automation.

Anthropic/OpenAI aren't planning to have their superintelligence take over the world, but they're still afraid that someone else will do it.

AI psychosis is literal psychosis in humans associated with AI use. I don't think the author is psychotic? We have all these useful new terms for AI phenomena, if we can keep them straight...

Fair critique, voiced elsewhere in this thread as well, I am regretting my use of the term.

I absolutely understand that, because that response puts me in one after a couple times.

> You’re absolutely right! Let me fix that.

"AI style" is an artifact of certain writing styles being overrepresented in training data. I expect in the long run it will be impossible to distinguish AI writing reliably. https://marcusolang.substack.com/p/im-kenyan-i-dont-write-li...

I did predict as much, I figured it would occur as a side effect of improved information density. As the models got smarter they would have more useful pints to make. On one hand it feels validating, on the other hand I am a bit worried that my written work will look like AI. While I would consider it no longer slop I do worry about a loss of relative advantage.

I think LLM companies should standardize censorship of some totally innocuous obscure topic, like Furbies. That way, we can attempt to jailbreak AIs by asking about Furbies without any risk of getting banned.

I think there's a precedent here:

https://www.qwantz.com/index.php?comic=879

You don't need to have a right answer to question someone else's answer. "You are too certain" is a valid point in itself.

I think this is meant to detect insider knowledge of imminent risk of a nuclear attack. It's not meant to detect if nukes are already inbound.

Okay but that wouldn't produce the signal "all of the planes have suddenly taken off at the same time." No one who knows Manhattan's going to be nuked on Tuesday is going to be in Manhattan on Tuesday counting on their plane to get them out.

Unless the prez misses the nurse button and hits nuke

Great land of confusion reference!

That, or any other potentially apocalyptic event

Assuming that apocalyptic event is something which would impact the elites, is reliably predictable by the elites, escapable by flying, and has a window of opportunity where everyone will be trying to escape at the same time.

In reality, look at things like the 9/11 attacks, the 2004 tsunami, the Covid lockdowns, the Maduro raid: the elites didn't fly out all at once shortly ahead of time before the plebs knew what was going to happen. No doubt you can think of good reasons why it didn't happen in those circumstances, but trying to come up with a scenario where it actually would happen is quite a bit more challenging.

Also works for the Davos summit

Cool. How do they stop?

I think there might have been some other reasons Trump got elected.

Which do you think works better--protesting/suing the people making the decision? Or being the person making the decision? It's much harder to change the course of law from the outside, without access to power. The problem is that power corrupts.

Either works if enough people care enough. Maybe instead of blaming the messengers for not putting their own careers on the line, you should blame all the people who just didn't care at all?

What about you, do you work for the EFF? If not, I'll give you an out - donate just $100. (I just did.) As a bonus you can even get the book this article is from.