alkexr

Karma: 269

alkexr Jun 8, 2021, 9:42 PM
1 point
in reply to: Stuart_Armstrong’s comment on: The reverse Goodhart problem
You have a true goal, $V$ . Then you take the set of all potential proxies that have an observed correlation with $V$ , let’s call this $S (V)$ . By Goodhart’s law, this set has the property that any $U \in S (V)$ will with probability 1 be uncorrelated with $V$ outside the observed domain.
Then you can take the set $S (2 U - V)$ . This set will have the property that any $U^{'} \in S (2 U - V)$ will with probability 1 be uncorrelated with $2 U - V$ outside the observed domain. This is Goodhart’s law, and it still applies.
Your claim is that there is one element, $U \in S (2 U - V)$ in particular, which will be (positively) correlated with $2 U - V$ . But such proxies still have probability 0. So how is that anti-Goodhart?
Pairing up $V$ and $2 U - V$ to show equivalence of cardinality seems to be irrelevant, and it’s also weird. $2 U - V$ is an element of $2 S (V) - V$ , and this depends on $V$ .

alkexr Jun 8, 2021, 5:30 PM
10 points
on: The reverse Goodhart problem
Your $V^{'}$ is correlated with $U$ , and that’s cheating for all practical purposes. The premise of Goodhart’s law is that you can’t measure your true goal well. That’s why you need a proxy in the first place.
If you select a proxy at random with the only condition that it’s correlated with your true goal in the domain of your past experiences, Goodhart’s law claims that it will almost certainly not be correlated near the optimum. Emphasis on “only condition”. If you specify further conditions, like, say, that your proxy is your true goal, then, well, you will get a different probability distribution.

alkexr Jun 3, 2021, 6:42 PM
1 point
on: Networks of Trust vs Markets
Closely related: Leaky Delegation: You are not a Commodity (by Darmani).

alkexr May 31, 2021, 8:18 PM
8 points
on: Why don’t long running conversations happen on LessWrong?
Some frames worth considering:
- Strong Prune, weak Babble among LessWrongers
- Conversation failing to evolve past the low-hanging fruit
- People being reluctant to express thoughts that might make their account look stupid in a way that’s visible to the entire internet
- Everyone can participate, and as the number of people involved in a conversation increases it becomes more and more difficult to track all positions
- Even lurkers like me can attempt to participate, and it’s costly in terms of conversational effort to figure out what background knowledge someone is missing
- Most topics that appear on LessWrong are suited for mental masturbation only, they offer no obvious course of action through which people can decide to care about said topics
- There is way, way too much content (heck, I’ve only skimmed through the comments under this post)
- Long-running conversations don’t tend to happen; therefore there is little incentive in delving deep into one topic, so people (well, me at least) end up engaging with more topics, but in a shallow manner; which in turn creates conditions where long-running conversations are less likely
- Due to the way the platform is designed, the only real way to maintain a long-running conversation between persons A and B is if the pattern of response is ABABABAB..., so anyone not being confident at any point is a single point of failure
I also have a suggestion. After the discussion here inevitably fades, you could write another post in which you summarize the main takeaways and steelman a position or two that look valuable at that time. That might generate further discussion. Repeat. This way you could attempt to keep the flame of the conversation alive. But if you end up doing this, make sure to give the process a name, so that people realize that this is a Thing and that they are able to participate in this-Thing-specific ways.

alkexr May 29, 2021, 12:20 AM
1 point
on: What’s your visual experience when reading technical material?
The first layer of internal visual experience I have when reading is a degree of synesthesia (letters have colors). Most of the time I’m not aware that this is happening. It does make recalling writing easier (I sometimes deduce missing letters, words or numbers from the color).
Then there is the “internal blackboard”, which I use for equations or formulas. I use conscious effort to make the equation appear as a visual experience (in its written form). I can then manipulate this image as if the individual symbols or symbol groups were physical objects that can move and react with each other. This apparently allows me to solve more complex equations in my head than most mathematicians. (I believe this is a learnable skill.)
Finally, there are the visual experiences that I use to understand concepts. I’m not sure how to describe these, because these certainly aren’t actual images that are actually possible. More like structures of shapes, spatial relations and other “sub-visual” experiences. It’s not like I can actually visualize an n-dimensional subspace, but it isn’t simply a lower-dimensional analogue either. It looks thin, but with a vast inside, in a way that would be contradictory in “normal” visual experience.
Whenever I read about a concept that seems interesting (e.g. Moloch), I pause. Then I take the verbal experience of what I’ve read, and use it as a guide for some internal thought process to follow. The nature of this process is the creation and manipulation of impossible visual experiences of this kind.
These days my visualization is a lot fainter than it used to be, so faint in fact that sometimes I barely see anything at all, in spite of knowing what I’m (not) seeing. This includes my dreams, and maybe even waking experience (how would I tell?), and I believe this is unnatural. This only seems to have a negative effect on the “internal blackboard”, but not on any of the other mechanisms I mentioned.

alkexr May 15, 2021, 2:24 PM
8 points
on: Zvi’s Law of No Evidence
Absence of evidence of X is evidence of absence of X.
A claim about the absence of evidence of X is evidence of:
- the speaker’s belief of the listeners’ belief in X,
- absence of evidence of NOT X,
- the speaker’s intention of changing the listeners’ belief in X.
No paradox to resolve here.

alkexr Apr 7, 2021, 7:24 PM
2 points
in reply to: Ozyrus’s comment on: Don’t Sell Your Soul
Non sequitur. Buying isn’t the inverse operation of selling. Both cost positive amounts of time and both have risks you may not have thought of. But it probably is a good idea to go back in time and unsell your soul. Except that going back in time is probably a bad idea too. Never mind. It’s probably a good investment to turn your attention to somewhere other than the soul market.

alkexr Mar 31, 2021, 11:32 PM
3 points
on: Bureaucracy is a world of magic
These rituals are inefficient in cases where there is mutual trust between all participants. But sticking to formality is a great Schelling fence against those trying to gain an advantage by exploiting unwitting bureaucrats.

alkexr Jan 21, 2021, 8:11 PM
2 points
in reply to: neotoky01’s comment on: What is going on in the world?
The basis of the original post isn’t existential threats, but narratives—ways of organizing the exponential complexity of all the events in the world into a comparatively simple story-like structure.
Here’s a list of alternative high level narratives about what is importantly going on in the world—the central plot, as it were—for the purpose of thinking about what role in a plot to take
Memetic tribes are only tangentially relevant here. I didn’t really intend to present any argument, just a set of narratives present in some other communities you probably haven’t encountered.

alkexr Jan 18, 2021, 1:02 PM
43 points
on: What is going on in the world?
The above narratives seem to be extremely focused into a tiny part of narrative-space, and it’s actually a fairly good representation of what makes LessWrong a memetic tribe. I will try to give some examples of narratives that are… fundamentally different, from the outside view; or weird and stupid, from the inside view. (I’ll also try to do some translation between conceptual frameworks.) Some of these narratives you already know—just look around the political spectrum, and notice what narratives people live in. There are aslo some narratives I find better than useless:
1. Karma. Terrible parents will likely have children who can’t reach their full potential and can’t help the world, and who will themselves go on becoming terrible parents. Those who were abused by the powerful will go on abusing their power wherever and whenever they have any. Etc. Your role is to “neutralize the karma”, to break the part of the cycle that operates through you: don’t become a terrible parent yourself, don’t abuse your power, etc. even though you were on the recieving end.
2. The world is on the verge of collapse because the power of humanity through technology has risen faster than our wisdom to handle it. You have to seek wisdom, not more power.
3. The world is run by institutions that are run by unconscious people (i.e. people who aren’t fully aware of how their contribution as a cog to a complex machine affects the world). Most problems in the world are caused by the ignorant operation of these institutions. You have to elevate people’s consciousness to solve this problem.
4. Humans and humanity is evolving through stages of development (according to something like integral theory). Your role is to reach the higher stages of development in your life, and help your environment to do likewise.
5. History is just life unfolding. Your job isn’t to plan the whole process, just as the job of a single neuron isn’t to do the whole computation. The best thing you can do is just to live in alignment with your true self, and let life unfold as it has to, whatever the consequences (just as a neuron doing anything other than firing according to its “programming” is simply adding noise to the system).
6. Profit (Moloch) has overtaken culture (i.e. the way people’s minds are programmed). The purpose of profit (i.e. the utility function of Moloch that can be reconstructed from its actions) isn’t human well-being or survival of civilization, so the actions of people (which is a manifestation of the culture) won’t move the world toward these goals. Your role is to raise awareness, and to help reclaim culture from the hands of profit, and put a human in the driver’s seat again (i.e. realign the optimization process by which culture is generated so that the resulting culture is going to be aligned with human values).
7. Western civilization is at the end of its lifecycle. This civilization has to collapse, to make way for a new one that relates to this civilization in the same way the western civilization relates to the fallen Rome. Your role isn’t to prevent the collapse, but to start creating the first building blocks which will form the basis for the new civilization.
8. The world is on the brink of a context switch (i.e. the world will move to a formerly inaccessible region of phase space—or has already done so). Your models of the world are optimized to the current context, and therefore they are going to be useless in the new context (no training data in that region of the phase space). So you can’t comprehend the future by trying to think in terms of models, instead you have to reconnect with the process that generated those models. Your role is to be prepared for the context switch to try to mess things up as little as possible, though some of it is inevitable.
9. Reality (i.e. the “linear mapping” you use to project the world’s phase space to a lower dimensional conceptual space through your perception and sensemaking) is an illusion (i.e. has in its Kernel everything that actually matters). Your role is to realize that (and after that your role will be clear to you).
10. The world is too complex for any individual to understand. Your role is to be part of a collective sensemaking through openness and dialog that has the potential to collectively understand the world and provide actionable information. (In other words, there is no narrative simple enough for you to understand but complex enough to tackle the world’s challenges.)
11. The grand narrative you have to live your life by changes depending on your circumstances, just like it depends on where you are whether you have to turn left or right. Your role is to learn to understand and evaluate the utility of narratives, and to increase your capacity to do so.
This list is by no means comprehensive, but this is taking way too much time, so I’ll stop now, lest it should become a babble challenge.

alkexr Jan 16, 2021, 1:57 PM
3 points
on: Covid 1/14: To Launch a Thousand Shipments
You’d also have to consider the long-term effects on the incentive landscape of e.g. establishing the precedent of companies getting $4B deals in case of a pandemic regardless of whether their vaccine works or not. In general, doing things the reasonable way has the downside of incentivizing bad actors to extract any free energy you put into the system by being reasonable until you’re potentially no better off than the way Delenda Est Club is handling the situation right now. In any case, I don’t see any long-term systemic effects even being considered here, so I’d be surprised if the suggestions didn’t have some significant fallout further down the line.

alkexr Oct 14, 2020, 10:58 PM
4 points
on: Have the lockdowns been worth it?
Lockdown incentivized politicians to establish positions on a lockdown, which has led to people having strong opinions about it. Even assuming no damage from further polarization, we have a roughly 50% chance of having an anti-lockdown government when the next pandemic hits, with a 10% chance of this new incentive being the deciding factor in not enacting a lockdown (or failing to implement it). Even if we assume that only 10% of the effects of this polarization is the result of the lockdown actually happening, with a 1% yearly chance of a pandemic more dangerous in expectation than the current one (~100M dead), we have ~1M QALYs lost, extrapolated worldwide over the next 10 years (while this effect is most pronounced).
Note: This is just a quick check to see that the effect is at least plausibly on an order of magnitude worth taking into consideration. I’m only somewhat confident that the effect isn’t in the opposite direction. I’m only commenting (as opposed to answering) because primarily I expect weak points in my general process of speculation pointed out, not because I believe this to be well-informed enough to be useful.

alkexr May 7, 2020, 10:03 PM
6 points
on: Why do you use a pseudonym on LessWrong?
I live in a social environment where expressing opinions or otherwise giving information about myself could have negative consequences, ranging from mild inconvenience to serious discrimination. I have no intention to hide my real identity from those who know the account, but I do want to hide my account from those who know my real identity (and aren’t close friends). I use this name for most online activity.

alkexr Apr 23, 2020, 10:00 PM
3 points
in reply to: Raemon’s comment on: My experience with the “rationalist uncanny valley”
I’ve been aware for a while now that having enough awareness to notice being trapped is not enough to step outside the pattern, but I can’t step outside this pattern. I also believe that admitting that there is no substitute for practice isn’t going to be causally linked to me actually practicing (due to a special case of the same trap), so I’ll just go on staying trapped for now I guess.

alkexr Apr 8, 2020, 1:29 AM
2 points
in reply to: CellBioGuy’s comment on: What will happen to supply chains in the era of COVID-19?
Being self-sufficient and robust as a national economy is accepting a competitive disadvantage relative to a global just-in-time supply chain in times of prosperity in exchange for a competitive advantage during a crisis. Selection pressures will push economies accepting this tradeoff towards being actively interested in a world with more crises.

alkexr Oct 15, 2018, 10:02 AM
3 points
on: Outline of Metarationality, or much less than you wanted to know about postrationality
Question: how does postrationality and instrumental rationality relate to each other? To me it appears that you are simply arguing for instrumental rationality over epistemic rationality, or am I missing something?

alkexr Oct 15, 2018, 10:00 AM
3 points
in reply to: nshepperd’s comment on: Outline of Metarationality, or much less than you wanted to know about postrationality
However, if this is really what ‘postrationality’ is about, then I think it remains safe to say that it is a poisonous and harmful philosophy that has no place on LW or in the rationality project.
It feels like calling someone’s philosophy poisonous and harmful doesn’t advance the conversation, regardless of its truth value, and this proves the point of the main post well.

alkexr Sep 19, 2018, 11:07 PM
3 points
in reply to: Richard_Ngo’s comment on: Tradition is Smarter Than You Are
Being able to speak is probably more important than being as smart as a human. Cultural / memetic evolution is orders of magnitude faster than biological, but its ability to function is dependent on having a memory better than mortal minds. Speech gives some limited non-mortal memory, as does writing the printing press, or the internet. These inventions enable more efficient evolution. AI will ramp up evolution to even higher speeds, since external memory will be replaced with internal 1) lossless and 2) intelligent memory. As such I am unconvinced that this would mean slower takeoff speeds. (You just explained that the most important factor in doing well as humans is something humans are not overly good at, instead of the special magic that only humans posess.)

alkexr Jun 10, 2018, 5:02 PM
1 point
in reply to: Dacyn’s comment on: Physics has laws, the Universe might not
People not being able to come up with any idea but that diseases are a curse of the gods is strong evidence not for diseases being a curse of gods but for the ignorance of those people. The most likely answer to that question is either something no one will think of for centuries to come or simply that the model of separating objects into “sorts of things” is not useful for deciphering the misteries of the universe despite being an evolutionary advantage on the ancestral savanna.

alkexr May 26, 2018, 12:45 AM
5 points
on: Of Two Minds
You might have gone too far with speculation—your theory can be tested. If your model was true, I would expect a correlation between, say, the ability to learn ball sports and the ability to solve mathematical problems. It is not immediately obvious how to run such an experiment, though.