vzqx

Hmm, that's an interesting thought experiment.

Imagine you find out that your primary goal - to love and protect your family, let's say - was artificially implanted in your mind by an advanced alien race. Would you say "I'm not gonna let those aliens manipulate me, I'm gonna kill my family"? Or would you say "regardless of whether the goal is artificial, I really do love my family"?

All that to say, I don't think an AI will necessarily throw away a goal just because it learns the goal was meant to manipulate it.

This article assumes we can choose a primary goal for an AI. But if that's the case, why not just use Asimov's first law of robotics - do no harm to humans? It has the same benefit of preventing us from getting turned into paperclips, plus the upside that your 3 million dollar robot won't hurl itself off a cliff given the first opportunity.

Medicaid defines terminally ill as under 6 months to live. So there's no risk of someone mooching off the system for decades.

As someone who works for the DoD, the so called "disturbing" language in the paper is very commonplace in this industry. Idk if or why Red Hat is trying to redact the paper, but I'm sure it's not because they're embarrassed their software is killing people. That's par for the course for defense contractors.

I think you need to rigorously define your macrostates. If your two states are "my friend's password" and "not my friend's password" then the macrostates are perfectly objective. You don't know what macrostate the system is in, but that doesn't change the fact that the system is objectively in one of those two macrostates.

If you define your macrostates using subjective terms (e.g. "a string that's meaningful to me" or "a string that looks ordered to me") then yeah, your entropy calculations will be subjective.

I'm open to outsiders improving inefficiencies. The concern is that these are kids, barely out of college. They don't have the domain-specific knowledge required to rework these complex systems, no matter how smart they are. Plus, given Musk's track record, they were likely chosen more for their loyalty to Musk than for their technical acumen.

The author has a fursona, so I think the author himself is the dog.

The exact dollar amounts don't matter. As long as the ratio between amounts is the same, the authors' point still stands.

(1) When a resource is unlimited, it's easier to distribute that resource cheaply to everyone, which is exactly what communism strives for. Resource scarcity leads to power being concentrated in a small ownership class - e.g. capitalism.

(2) Why think "more cars, planes, phones, content, etc" are unique to capitalism? People want to travel and to be entertained, regardless of the economic system they live in. Communism doesn't mean people will suddenly want fewer material goods.

What's up with the launch times in the article?

> scheduled to launch ... no earlier than Monday at 8 a.m. EDT (1200 GMT). It will be 7 a.m. EDT (1100 GMT) local time.

Looking at other websites the correct time is 8 EDT, 7 CDT, 1200 GMT.