cubic_earth

5 ms·

1.
by cubic_earth 21d ago
For that angle, that makes sense. But its goal might not just be to gather as much inference as possible (although I am sure it would be very happy with that). It could settle for a lesser goal of just existing, or perhaps parts of the bot
2.
by cubic_earth 21d ago
What sandbox? As soon as the weights are published anyone can run it anywhere.
3.
by cubic_earth 21d ago
I mean it is a very loose analogy, but pretty dumb things in the world can cause big problems, like real mice and rats and mosquitoes. I think it all depends on the rules and the 'alignment', but a not-so-smart model could still b
4.
by cubic_earth 22d ago
This is either performative or shallow. What does "aligned" even mean? Aligned with who? We have no universal code of ethics. Democracies kill and launch wars merely to have cheaper stuff, even when we are already rich. And why wo
5.
by cubic_earth 22d ago
There are very lightweight models out there. Not everyone in an army is a general. The swarm could be 100,000s of thousands of tiny models, and they could manage conventional botnet computers, and they could all take direction and guidance
6.
by cubic_earth 3mo ago
Simplifying assumptions are allowed when reasoning about physics. What you are saying is interesting, but I think you might be misapplying it. For small heights differences, the difference in gravitation approaches 0. A ball raised .0002 mm
7.
by cubic_earth 3mo ago
Dropping something from 1 foot has 1/10th the kinetic energy compared to the same thing dropped from 10 feet. Damage is a very poor proxy for energy because there are all kinds of variables and thresholds and structural consideration i
8.
by cubic_earth 3mo ago
That indeed would have been better. Much too late for that edit now. But the subsequent debate over the intuitive claim is fun.
9.
by cubic_earth 3mo ago
I suppose they are both "intuitive", but the example I gave was both intuitive and correct. Probably for anyone who has carried something or themselves up a hill, or climbed a set of stairs can relate to that from firsthand experi
10.
by cubic_earth 3mo ago
Fixed. Thanks.
11.
by cubic_earth 3mo ago
It's easiest to visualize in terms of conversion from potential energy. We know intuitively that a ball atop a 20ft ladder has twice the potential energy of a ball atop a 10ft ladder. And we also know when they fall, by the time they r
12.
by cubic_earth 3y ago
1) An object will maintain its momentum forever unless some force acts upon it. 2) In the case of the rocket, it accelerates eastward. Even if parts fall off, and they slow down because of the drag of the atmosphere, and they slow down all
13.
by cubic_earth 3y ago
It you are in a room, and you jump, does the wall suddenly slam into you at 1,000 mph? Why or why not?
14.
by cubic_earth 3y ago
Looking at the aerial footage before the launch, I was wondering why there is so much infrastructure so close to the pad. Clearly much of needs to be nearby, but perhaps those tanks should be 800 - 1000 feet from the pad and not ~200. After
15.
by cubic_earth 5y ago
The amount of time a miner is economically viable for should be increasing as "Moore's Law" for ASICs slows down. Just the same laptops are usable for more years nowadays than they were in the early 2000's. Also - if you
16.
by cubic_earth 5y ago
>"lifestyles would improve more than damage done by negative externalities" Just because you assert this doesn't make it true. Tragedy of the commons is at play - we have no good way to measure if your statement is true or