LLMs should have an Agreeableness setting similar to Effort. Sometimes I want it to go along with whatever I sayโฆ Other times I want it to go full Argument Clinic.
Iโve seen people offer this as an explanation for why models tend to express positive views of Christianity as well. It would be interesting to find out what aspects of the training data led to the association. Like, was it the Buddhist and Christian teaching contained in the training data, or was it more what people write about how they perceive Buddhists and Christians?
@simonsarris@powerbottomdad1 IMO Interstellar is a pretty solid adventure movie with interesting themes right up until the end. The plot unravels at that point because it trades away the mystery of the wormhole and who constructed it just for pure emotional impact.
Seems like weโve been through this cycle already with nuclear.
Whatโs interesting is that unlike with nuclear, the people leading early development of AI are often warning the public about catastrophic risks. People accuse them of seeking regulatory capture, but maybe their bigger concern is avoiding more onerous regulation in the long term by getting ahead of the problem.
Itโs hard to have a balanced perspective on this, because the tides of public opinion are always going to push us toward some version of this good guys/bad guys narrative.
Personally, I think itโs better just to gradually build relationships (even if merely parasocial ones) with trustworthy people I find online, and so far Iโve found plenty of decent people across each of the major labs. This doesnโt make the labs 100% trustworthy as a whole, but at least it helps me moderate my POV and avoid taking sides in every case.
@narmourism@waitbutwhy Itโs worth noting that the people who vote blue on this poll are not actually putting their lives on the line, and they plausibly have a social incentive to say they voted blue. So the results might be skewed in the other direction. Hard to say for sure!
Playing it out 100 times is not part of the original scenario, but Iโll grant you this for the sake of argument.
If we assume the average number of children per-individual is equal on both sides, then if >50% push blue the first time, shouldnโt the same outcome occur in every subsequent iteration?
Exactly. People saying everyone lives if 100% of people push red ignore the fact that this is not a realistic scenario (assuming the vote is truly private). You only need >50% to push blue to achieve the same result.
If you permit a condition where *everyone* is made to know beforehand that there will be a red majority, then I think the logic for pushing red makes sense. In that scenario, everyone basically knows that pushing blue is a death sentence, so they can be reasonably expected to bear responsibility for self-preservation.
So basically, the most reasonable choice depends on the extent of information sharing that could be achieved prior to the decision point.
I find this stuff super fun to speculate about as a Christian.
Ofc there was a long period of time on our own planet before Jesus lived, so if we find intelligent alien life and they donโt have their own Space Jesus, this could just mean theyโre at an earlier point on their own spiritual timeline.
The areas of doctrinal alignment/misalignment would still be empirically significant though. If we find they have similar concepts of morality and an ultimate perfect being, this could be indicative of divine influence. Total absence of moral instincts would suggest the opposite.
His use of โI think therefore I amโ seems contrary to the original point of this saying. Descartes was trying to demonstrate a logical ground for belief in oneโs existence as a thinking agent, even from a stance of radical doubt. It had nothing to do with proving consciousness.
The point of anti-functionalism is that a thinking agent/system can exist without being conscious. This agent would be capable of reasoning that it exists based on its ability to think, and it could even tell you that it exists, but it wouldnโt experience its own existence. There would be no โwhat it is likeโ to be that system.
Hmm, it seems like you're arguing that the basic THAT provided by the field would be a fact of reality independent of whether there are any brains to individuate it. If this is true, would you grant the possibility that other structures would produce some kind of what it is like? If so, what level of complexity is required, and how do you justify that threshold?
@KleynMichael@DouthatNYT Basically it would severely undermine functionalist explanations of consciousness, leading to the need for a separate explanation for the functional complexity of the world and all the qualities that arise from it, such as rationality and moral awareness.