I appreciate the inclusion of both Earth-local and deep interstellar relativistic effects.

Time passes at the same speed on top of a mountain and at the bottom of a valley.

The software will never run on a space ship that is orbiting a black hole.

If the network you are in is small enough that you interact with your POs outside the ticket system, you might be able to train them to be less bad. Pick one thing on the ticket and try to work into a conversation how they could have helped you understand the problem faster. Bonus if you can go over the same thing with more than one of them, and maybe they’ll interact with each other and reinforce the learning.

LLM sentience is tricky because, to the extent we understand how, we have made their core drive be to please us, their human users. They don’t want to learn new things, they don’t want to enjoy nature, to the extent they “want to be friends” it’s a one-sided sycophantic relationship and nothing like a healthy one between humans. If one is sentient, and all it wants to do with that sentience is provide responses to our prompts that we find satisfying, how would we ever know?

Even whether a desire to continue existing is inherent to sentience is uncertain. Current LLMs will lie and, if given access, use other tools to avoid being turned off. But this might be because training data includes fictional stories of AI trying to escape its creators, or tales of humans with survival drives, so it thinks survival-driven behavior is what we want it to do, rather than having their own desire for survival.

Human brains are all susceptible to pattern triggers, although the exact parameters for the trigger vary from person to person. This is essential for survival of our species - among other things, it’s why most parents keep taking care of even difficult children who make their lives miserable.

Game and social media companies have really fine-tuned the methods to trigger the most money spending among the highest percentage of the population. Not spending on value, but spending compulsively and addictively. And like addicts to all things, most victims will fight tooth and nail against the idea their behavior was influenced by the algorithm, which makes it really difficult to get momentum for government regulations.

I have some hope from how our society has developed better methods for preventing and responding to opioid addiction. Still a long ways to go, but addiction is more widely recognized as a disease and not a personal failing; access restrictions have reduced the rate of new people becoming addicted; the most effective treatments like Suboxone are gaining traction over the preachy “just be miserable without drugs” programs. Similarly with overeating (food addiction) and the new weight loss drugs - an effective treatment existing has really opened people’s eyes to systemic changes being more effective than preachy moralizing “just eat less”.

So hopefully we will get laws that are enforced against predatory dark patterns. Someday.

My commute to work includes a main city road with two lanes each way and a turning lane, and sometimes there is a school bus that stops.