I think ozymandias’ point is that this law give incentives to social media companies like Facebook to err on the side of censorship by punishing non-censorship of things that should be censored and not punishing censorship of things that shouldn’t be censored.
Paperclip Minimizer
Why would the argument for the right not to be censored by the government not extend to the right not to be censored by any other organization with some level of internal dissent (that is to say, any organization not a cult) ?
Not sure what architectures you’re referring to
I think habryka is thinking about modern machine learning architectures that are studied by AI researchers. AI research is in fact a distinct subfield from programming-in-general, because AI programs are in fact a distinct subgroup from programs-in-general.
I think the crux of this post has been correctly argued against by other people in this comment section, but I want to address a small part of your post:
Presumably, the goals explain the extent to which some humans exhibit instrumental convergence, and the rest of the utility function explains why we haven’t yet tiled the universe with money.
No, the reason we haven’t yet tiled the universe with money is that we don’t desire money for its own sake, only becaue it is an instrumental value that helps us achieve our goals. If we could tile the universe, we would tile the universe with happy humans. The problem is that we can’t tile the universe.
In any case I will post a write up of the results after the survey closes on May 1st.
When are you going to post said write up ?
This doesn’t make the AI any safer, given the whole (neural network + neural network builder¹) is still a misaligned AI. Real life examples happen all the time.
¹: I don’t know if there is a standard term for this.
To clarify, I am firmly anti-wireheading.
The problem is that “the training environment does not include that scenario” is far from guaranteed.
Maybe moving towards a more tribal social system.
The rise of blockchain technologies and related inventions may mean a move towards a more fiscally conservative economic policy.
ChristianKI is right: Murderless meat may make vegetarianism more palatable to meat-eaters, but it’s irrelevant to veganism.
It reduces the cost of not eating meat, but not the cost of not eating dairy and eggs.
Dissolving Scotsmen
Indeed. As I said, I don’t expect this to break new ground, but I am writing this as a pointer for future reference. I considered including mention of the motte-and-bailey concept, but decided that given it is 1) not well-known outside of the rationality community and 2) controversial inside it, it wasn’t worth it and linking to Eliezer’s and Scott’s writings on language would be sufficient (though I now notice that Scott’s “All in All, Another Brick In The Motte” isn’t included in core reading).
Wait, it isn’t? Do you (or anyone else reading this comment) know if there was a particular reason why not?
I’m not sure what “it” is supposed to refer to.
Also, I didn’t know Motte and Bailey was controversial in the rationality community, I’d be interested in reading links of people opposing it if you have any.
See Against Motte and Bailey by ozymandias
I have no idea why it isn’t included.
Alice’s puppy-cuddling club could include a social norm against giving psychotic monkeys control over the economy. In fact, in every real-life example of this fallacy coming to mind, the puppy-cuddlers and the psychotic-monkey economy-Czarists are different factions who hate each other.
You’re welcome.
Isn’t the least convenient world a world where FAI is outright impossible and all AGIs are Unfriendly ?
I agree, but what ways to achieve things outside the government do you support for the rationality movement ?