Deciding which predictions were accurate can indeed become an issue. However, most of the time this does not become a problem unless the resolution criteria are ambiguously defined. During forecasting tournaments, forecasters working on any question are expected to adjust their predictions according to the question’s fine print (such as the source of the inflation data that will be used for the question’s resolution).
Regarding politics affecting the selection of questions — can you explain why this would be a problem?
|The TIPS spread implies that the market expects low inflation for a long time
Could the low spread imply that the market is skeptical about the inflation being honestly calculated by the government?
michael vassar: “ I don't think that Leonid understands what "rationalists" are. I ask him, where are you used to encountering these exotic creatures?”
From wiki: "rationalists" are people who believe that “the truth can best be discovered by reason and factual analysis, rather than faith, dogma or religious teaching”. I have encountered plenty of these exotic creatures (an occupational hazard). Of course, one needs to remember that subscribing to some doctrine and actually following it in a personal life are not the same things.
I suspect all of us humans lie to ourselves. We are wired that way.
The difference between self-declared “rationalists” and “believers” ("I don't care about the evidence, I just want to believe in God”) may be just different sources of self-respect. The “believer” may actually respect himself more precisely because his faith is blind (“Blessed are those who have not seen and have believed”). The “rationalists” taking pride in being exactly the opposite, would search for logical arguments to motivate the position that in their heart they have already chosen. This is rarely difficult to do, as rational pro and contra arguments can be made for virtually any point of view.
Eneasz,
Our problems are “real” only because they make suffer us or those we care for. Our looks or income, for instance, seem to be a problem only when they fall behind the common standards of beauty/average salary. In a world where obesity considered attractive and is not a danger for your health, it stops being a “real” problem. Thus if, together with our friends, children and all those we care for, we could move to the idealized virtual reality, we would solve all the “real” problems that Eliezer described.
The fact that such a move does not appeal to many rational people suggests that Existential Angst is no less “real” problem than poverty, obesity etc.
“If it seems to you like nothing you do makes you happy, and you can't even imagine what would make you happy, it's not because the universe is made of particle fields. It's because you're still solving the wrong problem. “
Imagine being offered an option of spending the remainder of your natural life-span inside a virtual reality machine where all your material problems (poverty, obesity, loneliness etc) are solved. Plus as an added bonus you would be able to consume unlimited quantities of virtual heroin without damaging your health or your virtual social life.
If the “meaning of life” is a meaningless concept, shouldn’t every reasonable person jump on this offer? Would you?
Elliot: âOnce hunting music was created, females could select mates not just by how well they hunted directly (which they often didn't directly observe), but also by the quality of their hunting music... So music became a useless display (useless for survival) used in sexual selection, like the peacock's tail. â
The majority of people (including non-hunting females) enjoy listening to music without even trying to perform it themselves. Among those few, who did learn how to play musical instruments, most could get greater boost to their sex-appeal by devoting their time to body-building instead. Using the peacock analogy, I would say that our âmusical tailâ seems too large judged by its effect on the other sex.
There is a tendency for older generation to feel nostalgic for the time of their youth and for the younger generation to strive for changing the status quo. So I wonder whether the modern perception of moral progress (as opposed to perennial complaints of moral degradation popular among our ancestors) comes from the youth being more economically and politically empowered than ever before, which allows it to dominate public discourse.
Thank you for the thoughtful reply.
I’ll try to respond to it point by point.
This does complicate forecasting, but the two effects are unlikely to perfectly cancel each other. In case the two effects are very close in magnitude, the question’s political charge, Cj , would be close to zero. This would not compromise the method, but only require a larger number of questions in order to accurately calculate the models’ bias.
Typically, to calculate bias on a particular issue you do not need to ask questions about that issue directly. For example, the biases about the current war in Ukraine are strongly correlated with the biases about US domestic issues. So, it would be impossible to preserve the LLM's bias about Ukraine simply by removing all Ukraine-related questions.
Certainly. For example, it cannot be logically proven that viewing income inequality as a good or bad thing in itself is wrong. However, in practice, most arguments about inequality focus on its social consequences which is where the bias manifests itself. So, a debiased LLM would not be able to give a reasoned response on whether income inequality is good or bad on its own, but it should be able to correctly describe its impact on economic growth, crime, etc.