Questão
de Ética
PT

Simplified version · Volume IX · chapter 9

Longtermist Strategies: Broad and Targeted

In plain language, for readers who are new to the topic. About 2 min of reading.

What is this page? A simplified explanation written for this site based on Luciano Carlos Cunha’s text. It leaves out details, examples and sources. For the full argument, read the author’s original text.

Two kinds of strategy

To have the best impact on the future, we can use two kinds of strategy:

  • Broad: they try to have a good impact in many different scenarios.
  • Targeted: they try to have a good impact in specific scenarios.

Pros and cons

  • A targeted strategy can have a huge impact if the scenario it was planned for actually happens. If it doesn't, it might be useless.
  • A broad strategy has a positive effect in almost every scenario, even without our knowing what the future will be like.

An example of a broad strategy

One of the most important broad strategies is increasing consideration for all sentient beings in society. For that, it's important to:

  • consider all sentient beings, including those who usually get forgotten, like aquatic animals, invertebrates, and possible digital minds in the future;
  • care about all kinds of harm, not only those caused by humans, but also natural ones and those that might come up in the future.

The more widely this consideration spreads, the better the chances of preventing s-risks, whatever the future turns out to be.

Targeted strategies

Examples of targeted strategies would be working specifically on how artificial intelligence will treat sentient beings, or on rules for space colonization.

In short

  • Broad strategies help in many scenarios; targeted ones, in specific scenarios.
  • Increasing consideration for all sentient beings is a broad strategy.
  • It should include all beings and all kinds of harm.

Next step

Now you are ready for the full text

The author’s text gives the arguments in detail, with examples, objections and the sources for every claim.

Read the original text