Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Harry Potter and the Methods of Rationality finished today

fanfiction.net

Post reply on HN
Print view View original

141–150 of 173 posts

Re: Harry Potter and the Methods of Rationality finished today

#141

Post by davidgerard »

judk wrote: post #98
andolanra wrote: post #55

Earlier quoted context omitted.

This is the kind of thing which could very easily turn into a flame war, and is much further off-topic. I'm not interested in that kind of argument right now, so all I will say in this thread is this: My personal experience with Less Wrong-style rationalism, to simplify the situation aggressively, is that it has a core of good, useful tools (Bayesian reasoning, strict positivism, utilitarianism) that I have no proble…

And yet no one has really made an convincing argument of why the conclusions of utilitarianism contradict the axioms of utilitarianism. And no one seems to offer the rather simple solution that utility is non-linear (so dust in ones eye is really a quite small fraction as bad as a century of torture), and people who mock Eliezer's utilitarianism hypocritically do so on sweatshops built&recycled computers, clothes, an…

>And yet no one has really made an convincing argument of why the conclusions of utilitarianism contradict the axioms of utilitarianism.

No, they've said "this is a reductio ad absurdum, perhaps reality is a bit more complicated than that." You don't get to assert your assumption - that simplistic utilitarianism works as advertised - as evidence.

Consider, for instance, minmaxing as an alternative. (Like real-life AI work tends to.) What answer does that give?

Re: Harry Potter and the Methods of Rationality finished today

#142

Post by MichaelGG »

davidgerard wrote: post #134
MichaelGG wrote: post #107

Earlier quoted context omitted.

>.g. "it's better for one person to be tortured for fifty years than for an extremely large number of people to each get a speck of dust in their eye"), The number in question was 3^^^3 using Knuth's Up Arrow Notation. As he explains, that's a lot of people: (3^(3^(3^(... 7625597484987 times ...)))). If we decide that one person being tortured for 50 years is worth a quick blink by all those other people, then a stil…

In real-world AI, they tend to go for minmaxing, which avoids such absurdities. Minmaxing says "don't torture people, your dust speck has literally been defined to be insignificant."

Ok so now instead of one person being tortured for 50 years vs 3^^^3 people with dust specks, let's pose a new question:

7 trillion people tortured for 50 years vs 3^^^3 people with 7 trillion dust specks.

Does that make the flaw more obvious? You can't round before multiplying. What am I misunderstanding?

Re: Harry Potter and the Methods of Rationality finished today

#143

Post by davidgerard »

MichaelGG wrote: post #142
davidgerard wrote: post #134

Earlier quoted context omitted.

In real-world AI, they tend to go for minmaxing, which avoids such absurdities. Minmaxing says "don't torture people, your dust speck has literally been defined to be insignificant."

Ok so now instead of one person being tortured for 50 years vs 3^^^3 people with dust specks, let's pose a new question: 7 trillion people tortured for 50 years vs 3^^^3 people with 7 trillion dust specks. Does that make the flaw more obvious? You can't round before multiplying. What am I misunderstanding?

That in real-life AI, they don't use simplistic utilitarianism.

You're dodging the core of the problem, as stated above: "However, when pushed too far, those tools tend to break down—but the rationalist answer to that breakdown is all too often to embrace the model and discount the reality." https://news.ycombinator.com/item?id=9204442

Just because you love your model doesn't make its conclusions at extremes true.

Re: Harry Potter and the Methods of Rationality finished today

#144

Post by mirithekiwi »

eru wrote: post #126
Eliezer wrote: post #25

Earlier quoted context omitted.

Mileage varies! I'd like to affirm and congratulate your wise decision to stop reading at that point instead of continuing to torture yourself, and advise against anyone recommending parent to try again or go back. HPMOR is not going to be everyone's cup of tea!

I liked HPMOR, but mostly as a guilty pleasure. For comparison, `Harry Potter and the Natural 20' is a better piece, because the author just wants us to have a good time, and is not constantly trying to prove how smart he is. (HPMOR reminded me a bit about Freakonomics in that respect. But Freakonomics is way worse.) HPMOR could do with some radical editing, perhaps?

I got that same impression of HPMOR. However, I liked Freakonomics at least because I thought it was better edited or had better explanations.

Re: Harry Potter and the Methods of Rationality finished today

#145

Post by Dylan16807 »

TeMPOraL wrote: post #81
zyxley wrote: post #78

Earlier quoted context omitted.

We have no way to know if homulilly got drunk and puked on the floor in public, but we do know that Yudkowsky and assorted Lesswrong visitors had a reaction to the Roko's basilisk concept that most relatively "normal" people would judge as embarrassing or humorous.

That was one situation in the long history of a community with high standards of reasoning and discussion, that produces tons of interesting and educative content. Are you really going to judge the whole community by a single overreaction they AFAIR later admitted was an overreaction?

As far as I know Yudkowsky still treats it, or at least something 'near it in idea-space' very seriously.

It's not in the past.

Re: Harry Potter and the Methods of Rationality finished today

#146

Post by cbd1984 »

siglesias wrote: post #129
cbd1984 wrote: post #96

Earlier quoted context omitted.

> I think this is a very naive refutation of what the point was about Kuhn. I don't deny that it's naive. > First off, the notion that engineering "validates" science. What do you mean by validate? Do you mean successful engineering informed by a set of scientific principles, somehow, in a scientifically rigorous way, renders those principles true? As in, end of story true? Nothing is end-of-story true except in math…

My point about Mercury was important because it shows that there is an appreciable level of "give" that a theory has before the scientific community agrees that there is something wrong with it. That level of give is socially determined. It does matter if the discoverer of the anomaly is a Cambridge phd or a crackpot with no credentials. The measurement instruments matter, and the fallability of those instruments pla…

> We (and by we I mean Popperians) want to believe that science is a series of universally positive logical assertions that can be cut down by a single negative observation, as logic would dictate. But we don't always know what the criteria are for successful negative observations. The criteria are less rigid and well defined than we would be willing to admit.

Bayesian reasoning helps here, I think, because people are wrong, and different people are wrong with different probabilities. For example, overturning mass-energy conservation because someone said they saw a professor turn into a small cat or a strange spacecraft appear and disappear is not reasonable: The probability of one person being wrong or insane is a lot higher than the probability of something really well-verified being completely incorrect.

Is it political at times? Yes. Can it be improved? Sure. But it is flawed, not completely broken, and I think Kuhn makes too much of the flawed-ness which encourages people to imagine that it's completely broken and therefore the next paradigm shift will validate homeopathy.

Re: Harry Potter and the Methods of Rationality finished today

#147

Post by raincom »

saraid216 wrote: post #119
raincom wrote: post #115

Earlier quoted context omitted.

Aren't we past Kuhn these days? For instance, Imre Lakatos take it further than Kuhn in his "Research Programs". And Larry Laudan takes it further than these two: research traditions and problem-soloving. Well, we have all -isms (positivism, empiricism, constructive empricism) pragmatism, realism), which are to explain the progress of knowledge (natural sciences, that is).

You have to remember that, in this era where there are more books in existence than is possible for a single person to read within a lifetime, most people haven't heard of everything. I've certainly never heard of the two names you dropped.

In a way, you are right. However, they (Kuhn, Lakatos, Laudan) are common names for those who have studied philosophy of sciences.

Re: Harry Potter and the Methods of Rationality finished today

#148

Post by siglesias »

cbd1984 wrote: post #146
siglesias wrote: post #129

Earlier quoted context omitted.

My point about Mercury was important because it shows that there is an appreciable level of "give" that a theory has before the scientific community agrees that there is something wrong with it. That level of give is socially determined. It does matter if the discoverer of the anomaly is a Cambridge phd or a crackpot with no credentials. The measurement instruments matter, and the fallability of those instruments pla…

> We (and by we I mean Popperians) want to believe that science is a series of universally positive logical assertions that can be cut down by a single negative observation, as logic would dictate. But we don't always know what the criteria are for successful negative observations. The criteria are less rigid and well defined than we would be willing to admit. Bayesian reasoning helps here, I think, because people ar…

I don't think anybody is claiming that science is "completely broken," but there are some who want nice, clean, logical delineations, who think that science is filling out some invisible, giant truth table. And that by each assertion in that truth table, there's a straightforward "this is how to falsify me" entry that scientists can look up and enact.

Based on how science actually works, this notion is fanciful. No such table exists. If you were to have the luxury of asking the top physicists, say, to create such a table for you, they'd very likely all look different.

Also, your comment regarding homeopathy is something of a strawman. Paradigms are incommensurate. If we do incur a paradigm shift in our lifetimes, it's likely that our current ways of speaking about science will be unable to capture it.

Re: Harry Potter and the Methods of Rationality finished today

#149

Post by eru »

mirithekiwi wrote: post #144
eru wrote: post #126

Earlier quoted context omitted.

I liked HPMOR, but mostly as a guilty pleasure. For comparison, `Harry Potter and the Natural 20' is a better piece, because the author just wants us to have a good time, and is not constantly trying to prove how smart he is. (HPMOR reminded me a bit about Freakonomics in that respect. But Freakonomics is way worse.) HPMOR could do with some radical editing, perhaps?

I got that same impression of HPMOR. However, I liked Freakonomics at least because I thought it was better edited or had better explanations.

If you liked Freakonomics' topics, but abhorred the self-serving style, you will love Tom Schelling's The Strategy of Conflict.

Re: Harry Potter and the Methods of Rationality finished today

#150

Post by kissickas »

davidgerard wrote: post #133
kissickas wrote: post #83

Earlier quoted context omitted.

I haven't read all of the "Sequences," nor am I familiar with most details of the Less-Wrong-mindset's idiosyncrasies. However, I just re-read the whole book and I'm not sure either of those concepts have any bearing on this work. There is a bit of assertion that "timeless physics" is representative of true reality, but my reaction was to look it up and find that it's an interesting fringe theory. For all I know, EY…

He does assert it, then assume its truth in later pieces. http://wiki.lesswrong.com/wiki/The_Quantum_Physics_Sequence#...

Interesting. I still have no problem with it, of course, buttressed by my quick glance at that page to see, at the very top:

    Warning:  The central idea in today's post is taken seriously by serious physicists; but it is not experimentally proven and is not taught as standard physics.

    Today's post draws heavily on the work of the physicist Julian Barbour, and contains diagrams stolen and/or modified from his book "The End of Time".  However, some of the arguments here are of my own devising, and Barbour might(?) not agree with them.
Post reply on HN