Menu
Sign In Search Podcasts Charts People & Topics Add Podcast API Blog Pricing

Richard Ngo

๐Ÿ‘ค Speaker
48 total appearances

Appearances Over Time

Podcast Appearances

LessWrong (Curated & Popular)
"On Goal-Models" by Richard_Ngo

Central examples of utility functions are therefore functions of fairly simple features which can be evaluated in basically all possible worlds.

LessWrong (Curated & Popular)
"On Goal-Models" by Richard_Ngo

For example, functions of the consumption of a basket of goods in economics or functions of the welfare of individuals in axiology.

LessWrong (Curated & Popular)
"On Goal-Models" by Richard_Ngo

Conversely, consider having a goal of creating a beautiful painting or a great cathedral.

LessWrong (Curated & Popular)
"On Goal-Models" by Richard_Ngo

You can't evaluate the outcome as a function of simple features like quality of brush strokes, quality of composition, etc.

LessWrong (Curated & Popular)
"On Goal-Models" by Richard_Ngo

Instead, you have some sense of what the ideal is, which might include the ways in which each part of the painting or cathedral fits together.

LessWrong (Curated & Popular)
"On Goal-Models" by Richard_Ngo

it might be very hard to then actually give meaningful scores to how far a given cathedral is from your ideal, or whether you'd pick an X% chance of one cathedral versus a Y% chance of another.

LessWrong (Curated & Popular)
"On Goal-Models" by Richard_Ngo

Indeed, that feels like the wrong question to ask.

LessWrong (Curated & Popular)
"On Goal-Models" by Richard_Ngo

Part of what makes artists and architects great is when they aren't willing to compromise in pursuit of their vision.

LessWrong (Curated & Popular)
"On Goal-Models" by Richard_Ngo

Instead, they're constantly moving in whichever direction seems like it'll bring them closer to their single ultimate goal.

LessWrong (Curated & Popular)
"On Goal-Models" by Richard_Ngo

This is related to Dembski's distinction between selection and control as two types of optimization.

LessWrong (Curated & Popular)
"On Goal-Models" by Richard_Ngo

A rocket that's fixed on a target isn't calculating how good or bad it would be to miss in any given direction.

LessWrong (Curated & Popular)
"On Goal-Models" by Richard_Ngo

Instead, it's constantly checking whether it's on track, then adjusting to maintain its trajectory.

LessWrong (Curated & Popular)
"On Goal-Models" by Richard_Ngo

The question is whether we can think of intelligent agents as steering through much higher dimensional spaces in an analogous way.

LessWrong (Curated & Popular)
"On Goal-Models" by Richard_Ngo

I think this makes most sense when you're close enough to your goal.

LessWrong (Curated & Popular)
"On Goal-Models" by Richard_Ngo

For example, we can think of a CEO as primarily trying to keep their company on a stable upwards trajectory.

LessWrong (Curated & Popular)
"On Goal-Models" by Richard_Ngo

Conversely, a high school student who wants to be the CEO of a major company is so far away from their goal that it's hard to think of them as controlling their path towards it.

LessWrong (Curated & Popular)
"On Goal-Models" by Richard_Ngo

Instead, they first need to select between plans for becoming such a CEO based on how likely each plan is to succeed.

LessWrong (Curated & Popular)
"On Goal-Models" by Richard_Ngo

Similarly, a dancer or a musician is best described as carrying out a control process when practicing or performing, but needed to make a discrete choice of which piece to learn, and more generally which instrument or dance style to focus on, and even more generally which career path to pursue at all.

LessWrong (Curated & Popular)
"On Goal-Models" by Richard_Ngo

Where do we even get the criteria on which we make selections?

LessWrong (Curated & Popular)
"On Goal-Models" by Richard_Ngo

I think it's actually another control process, specifically, the process of controlling our identities.