Less Wrong is a community blog devoted to refining the art of human rationality. Please visit our About page for more information.

Comment author: Daniel_Burfoot 23 May 2017 04:39:56PM 1 point [-]

Can someone give me an example problem where this particular approach to AI and reasoning hits the ball out of the park? In my mind, it's difficult to justify a big investment in learning a new subfield without a clear use case where the approach is dramatically superior to other methods.

To be clear, I'm not looking for an example of where the Bayesian approach in general works, I'm looking for an example that justifies the particular strategy of scaling up Bayesian computation, past the point where most analysts would give up, by using MCMC-style inference.

(As an example, deep learning advocates can point to the success of DL on the ImageNet challenge to motivate interest in their approach).

Comment author: jsalvatier 24 May 2017 02:42:28AM 2 points [-]

There's not that many that I know of. I do think its much more intuitive and lets you build more nuanced models that are useful for social sciences. You can fit the exact model that you want instead of needing to fit your case in a preexisting box. However, I don't know of too many examples where this is hugely practically important.

The lack of obviously valuable use cases is part of why I stopped being that interested in MCMC, even though I invested a lot in it.

There is one important industrial application of MCMC: hyperparameter sampling in Bayesian optimization (Gaussian Processes + priors for hyper parameters). And the hyperparameter sampling does substantially improve things.

Comment author: jsalvatier 23 May 2017 03:45:27AM 7 points [-]

Funny enough, as a direct result of reading the sequences, I got super obsessed with Bayesian stats and that eventually resulted in writing PyMC3 (which is the software used in the book).

Comment author: Raemon 14 May 2017 10:33:04AM 3 points [-]

I currently have almost zero knowledge of Naruto and I'm interested in hearing more things about the perception/action skills thing as it applies to Naruto Classic (and/or rationalist!naruto)

Comment author: jsalvatier 14 May 2017 10:57:50PM 0 points [-]


Comment author: jsalvatier 14 May 2017 08:35:41AM 1 point [-]

If you want to see a billion examples of details mattering, watch anything about shipbuilding by this guy: https://www.youtube.com/watch?v=jM6R81SiKgA

Comment author: RomeoStevens 14 May 2017 08:12:24AM 5 points [-]

Ontology lock in. If you have nice stuff built on top of something you'll demand proof commensurate with the value of those things when someone questions the base layer even if those things built on top could be supported by alternative base layers. S1 is cautious about this, which is reasonable. Our environment is much safer for experimentation than it used to be.

Comment author: jsalvatier 14 May 2017 08:32:37AM *  1 point [-]

Great description. Yes, I think that's exactly why people are reluctant to see other people's points.

Comment author: Raemon 13 May 2017 09:01:39PM 3 points [-]

This felt important but I'm not quite sure what my next action is supposed to be.

Comment author: jsalvatier 14 May 2017 06:41:59AM 1 point [-]

Yeah, I wasn't too specific on that. I do endorse the piece that jb55 quotes below, but I'm still figuring out what to tell people to do. I'll hopefully have more to say in the coming months.

Comment author: jsalvatier 13 May 2017 08:31:15PM 1 point [-]

John Maxwell posted this quote:

The mystery is how a conception of the utility of outcomes that is vulnerable to such obvious counterexamples survived for so long. I can explain it only by a weakness of the scholarly mind that I have often observed in myself. I call it theory-induced blindness: once you have accepted a theory and used it as a tool in your thinking, it is extraordinarily difficult to notice its flaws. If you come upon an observation that does not seem to fit the model, you assume that there must be a perfectly good explanation that you are somehow missing. You give the theory the benefit of the doubt, trusting the community of experts who have accepted it.

-- Daniel Kahneman

[Link] Reality has a surprising amount of detail

14 jsalvatier 13 May 2017 08:02PM
Comment author: SquirrelInHell 29 March 2017 10:47:02PM 0 points [-]

This probably means that I don't understand your motivations. I honestly have trouble empathizing with them, or even imagining what they would be like, even after reading your post and thinking about it some. My previous question was an attempt to probe in a direction that I thought might be a superstimulus for the desire you were talking about - a super strong power structure, masculine-oriented, extraordinary sense of belonging and teamwork, run by trustworthy people for the right reasons. But apparently it's not. So what's different here?

Comment author: jsalvatier 30 March 2017 01:48:25AM 2 points [-]

I want you to come up to me, put your arm around me, ask me how I am and start telling me about the idea you’ve got. Show me you ought to be in charge, because right now I’m a little lost and you’re not.

My desire is not for some permanent power structure, but for other people to sometimes and temporarily take leadership with the expectation that I will probably do so in the future as well. I think one of the most valuable things I do is sit people down and say 'look, there's this problem you have that you don't see, but I think its fixable. You're stuck thinking of things as X, but actually Y.' And I wish people would return the favor more often.

In retrospect, I should have way more clear about this.

Comment author: Benquo 29 March 2017 04:02:59AM *  1 point [-]

I think there are a few things going on here worth teasing apart:

Some people are more comfortable with social touch than others, probably related to overall embodiment.

Some people are more comfortable taking responsibility for things that they haven't been explicitly tasked with and given affordances for, including taking responsibility for things affecting others.

Because people cowed by authority are likely to think they're not allowed to do anything by default, and being cowed by authority is a sort of submission, dominance is correlated with taking responsibility for tasks. (There are exceptions, like service submissives, or people who just don't see helpfulness as related to their dominance.)

Because things that cause social ineptness also cause discomfort or unfamiliarity with social touch, comfort with and skill at social touch is correlated with high social status.

Comment author: jsalvatier 29 March 2017 06:14:01PM 1 point [-]

Yes, I was trying mostly to talk about #2. I like the dominance frame because I think this kind fluid dominance roles is the something like the Proper Use of Dominance. Dominance as enabling swift changes status to track changes in legitimate authority.

Seems like that wasn't really very clear though.

I think I want to additionally emphasize, people being comfortable temporarily taking responsibility for other people. Sometimes I want someone to come in and tell me I have a problem I don't see and how to solve it. I try to do this for others because I think its one of the most valuable services I can provide for people. Letting them see outside themselves.

View more: Next