= 27d567bc2d48d9f40e9ccc20ea370b15)
BaconServ comments on Looking for opinions of people like Nick Bostrom or Anders Sandberg on current cryo techniques - Less Wrong
You are viewing a comment permalink. View the original post to see all comments and the full post content.
= 783df68a0f980790206b9ea87794c5b6)
Loading…= b715d02e85f7b3b4ae65ca3d3e450868)
Subscribe to RSS Feed
Powered by Reddit
= f037147d6e6c911a85753b9abdedda8d)
Comments (184)
Naive measurements are still measurements. Personally, I would like to see more research into prediction as a science. It is difficult because you want to jump ahead decades, and five-year-out predictions have only so much value, but I (naively) predict that we'll get better at prediction the more we practice concretely doing so. I would (naively) expect that the data used to calibrate predictions would come from interviews with researchers paired with advanced psychology.
You could try reading up on what is already known. Silver's Signal and the Noise is not a bad start if you know nothing at all; if you already have some expertise in the area, the anthology Principles of Forecasting edited by Armstrong (available in Google & Libgen IIRC) covers a lot of topics.
If you want non-domain-specific, that's called statistics, specifically statistical modeling.
I specifically want it to be more specific.
Do you mean more confidence in the prediction? Narrower error bands?
More domain-specific, invention specific, plotting events with ever-narrower error bands. If you could isolate the specific month that news media gets in a buzz about uploading three years in advance, that would be a significant increase in prediction usefulness.
That doesn't sound realistic to me. It sounds impossible.
It may be impossible even with significant effort, but it's certainly impossible without having first tried. I want to know, if nothing else, the limits of reasonable prediction. That itself would be useful information. Even if we can never get past 50% accuracy for predictions, knowing that all such predictions can only be 50% likely is actionable in the presence of any such prediction.
These are very much domain-specific plus are the function of available technology.
You seem to want psychohistory -- unfortunately it's entirely fiction.
I want exactly as I've stated: Research into a method.
So, um... go for it?
Maybe, but until then we still have to carry on with what we have got.