My claim is that this approach would be superior to that of trying to develop “Friendliness theory” in advance of having any working AGI, because it would allow experiment- rather than theory-based development.
In IT, “superior” more often means getting to market faster than competitors. Approaches that deliberately slow progress down by keeping humans in the loop are not “superior” in a lot of important ways. In particular they make your project less competitive—so you are more likely to completely lose out to the efforts of some other team—in which case the safety of your project becomes irrelevant.
In IT, “superior” more often means getting to market faster than competitors. Approaches that deliberately slow progress down by keeping humans in the loop are not “superior” in a lot of important ways. In particular they make your project less competitive—so you are more likely to completely lose out to the efforts of some other team—in which case the safety of your project becomes irrelevant.