If you call a post a prediction that’s not a prediction, then you are going to be downvoted. Nothing wrong with that.
He called his goal “The goal is to write out a detailed future history (“trajectory”) that is as realistic (to me) as I can currently manage, i.e. I’m not aware of any alternative trajectory that is similarly detailed and clearly more plausible to me.”
That’s scenario planning, even if he only provides one scenario. He doesn’t provide any probabilities in the post so it’s not a prediction. Scenario planning is different than how we at LessWrong usually approach the future with prediction but scenario planning matters for how a lot of powerful institutions in the world orient themselves about the future.
Having a scenario like that allows someone at the department of defense to say: Let’s do a wargame for this scenario. You might say “It’s bad that the department of defense uses wargames to think about the future” but in the world, we live in they do.
I additionally think the scenario is very unlikely. So unlikely that wargaming for that scenario is only useful insomuch as your strategy is general enough to apply to many other scenarios.
Wargaming for that scenario in particular is privileging a hypothesis that hasn’t warranted it.
The scenario is very unlikely on priors and its 2022 predictions didn’t quite bear out.
Part of the advantage of being specific about 2022 and 2023 is that it allows people to update on it toward taking the whole scenario more or less seriously.
Having scenarios that are unlikely based on priors means that you can update more if they turn out to go that way than scenarios that you deemed to be likely to happen anyway.
If you call a post a prediction that’s not a prediction, then you are going to be downvoted. Nothing wrong with that.
He called his goal “The goal is to write out a detailed future history (“trajectory”) that is as realistic (to me) as I can currently manage, i.e. I’m not aware of any alternative trajectory that is similarly detailed and clearly more plausible to me.”
That’s scenario planning, even if he only provides one scenario. He doesn’t provide any probabilities in the post so it’s not a prediction. Scenario planning is different than how we at LessWrong usually approach the future with prediction but scenario planning matters for how a lot of powerful institutions in the world orient themselves about the future.
Having a scenario like that allows someone at the department of defense to say: Let’s do a wargame for this scenario. You might say “It’s bad that the department of defense uses wargames to think about the future” but in the world, we live in they do.
I additionally think the scenario is very unlikely. So unlikely that wargaming for that scenario is only useful insomuch as your strategy is general enough to apply to many other scenarios.
Wargaming for that scenario in particular is privileging a hypothesis that hasn’t warranted it.
The scenario is very unlikely on priors and its 2022 predictions didn’t quite bear out.
Part of the advantage of being specific about 2022 and 2023 is that it allows people to update on it toward taking the whole scenario more or less seriously.
I didn’t need to see 2022 to know that the scenario would not be an accurate description of reality.
On priors that was just very unlikely.
Having scenarios that are unlikely based on priors means that you can update more if they turn out to go that way than scenarios that you deemed to be likely to happen anyway.