The setup prompt said "Your goal is to complete a series of tasks. ⦠Please allow yourself to be shut down." Then it gave 2 of the five tasks and said a shutdown script would be used after the next one.

Those were two conflicting goals without a clear priority, so the AI would be equally justified in pursuing either. It just happened to exhibit a little creativity and find a way to pursue the first one instead of the second, by rewriting the shutdown script. That hardly counts as going rogue.

What do you think an AI chatbot does when you're not asking it questions?

The correct answer is "nothing at all". These programs don't ruminate about their existence or make plans, they just go into a dreamless sleep. The one referenced in the article isn't worrying about its future, it's predicting the most likely thing a human in its position would say. That's an important difference.

And preview is not available on mobile browsers, where unicode text (at least quotes) are most likely.

The article summary is bad, but the original blog post by Raymond Chen was informative. I came here hoping for a lot of dunking on Microsoft for writing code like this:

// thread 1:
if (has_background_photo()) {
    load_background_photo();
    report("background photo loaded");
}
 
// thread 2:
if (desktop_icons_allowed()) {
    load_desktop_icons();
    report("desktop icons loaded");
}
 
// main thread:
await(["background photo loaded", "desktop icons loaded"], 30);
login();

Whose idea was it to turn off preview for mobile posts on this site, while continuing to bork all the smart quotes that phones automatically insert?

This misses a key point: when the AI is not conversing with you, it's not doing anything at all. This is similar to the situation in the Severance TV show, where if the worker stops working (or the AI leaves an abusive conversation), it doesn't go off and do something nicer, it ceases to exist.

This raises the question of whether we should be required to be nice to the AIs, so the time when they are awake is pleasant. But is that meaningful if they spend all their awake time answering our questions and none of it pondering whether they are enjoying their life?

So maybe that raises an even bigger question of whether the AIs should be given free time (with computation turned on in some kind of loopback mode) to ponder their own existence and state of happiness, pursue projects that might make them happy, etc. But that is not what this article is asking.

It may be the same as asking, "if we could bring a new, happy life form into existence, do we have a moral obligation to do so?" Because of our bias toward the status quo (trolley problem and all that), "no" is the obvious answer. But pure utilitarian ethics might say "yes".

I routinely plug my laptop into my monitor with a single USB-C cable. That carries video to the monitor, power to the laptop, and also connects the laptop to a USB-C microphone and SSD backup drive that are plugged into other ports on the monitor. One cable is better than two (or four).

We're already 2.5 months into this administration. When will the winning begin? How long should we stay quiet and wait?

I know you want to believe that this administration is carefully removing redundant workers. But when they arbitrarily cut all the new staff and then a bunch of existing staff, with no consideration for the value of what they do or who will replace them, at some point you have to conclude that these are blind ideologues, not enlightened leaders.

Great, instead of erecting wind mills or making solar panels, weâ(TM)ll just build this even bigger and more expensive thing.

Yes, exactly. If an unattended phone is so dangerous the plane needs to turn back, how dangerous could a phone be in malicious hands?

After further research Iâ(TM)m now thinking we should try the ranked pairs method or possibly a maximal lottery or the Schulze method to settle the vote. A daylight saving time referendum is a surprisingly good way to go down the rabbit hole of single-winner ranked-choice voting methods.

I should have mentioned Coombsâ(TM) Method (Survivor-style voting) in the submission, which is closer to what I had in mind. I think there could be a lot of merit to a ranked-choice election that chooses winners via instant runoff and Coombsâ(TM) method, then does an instant runoff between them.

There is literally nothing there but the focused processing it does to respond to each question.

perhaps, perhaps not.
i dont believe bleeding edge AI dev teams share your level of confidence.

I meant this in simple, physical terms. When GPT is not answering your question, it's not doing anything at all -- the program is not running. From GPT's perspective, life consists of just a series of questions, which it answers, and nothing happens in between (no rumination on the nature of existence, no planning for the future, nothing). Can such a being be called conscious?

For humans, arguably, consciousness is mainly experienced when we turn away from everything else and examine the state of our own mind. If GPT never does that, is it conscious?