Blog

LW Sequence notes – Mysterious answers to mysterious questions (Part 1)

I’m reading the LessWrong Sequences from the beginning. Most of them are a repeat for me but there may be some sections I skipped originally, and of course plenty of things forgotten.

The sequences provide stepping stones of logic to help people internalize non-trivial concepts about rationality and similar topics.

I’m going to leave notes on the main ideas I take away. Mostly to serve as a reminder to me, but hopefully it encourages other people to read along.

The first sequence is Mysterious answers to mysterious questions

  • A belief is only worth holding if it effects the distribution of anticipated experience. If following a chain of beliefs does not lead to an anticipated experience, or leads to a prediction that is contradicted by reality with sufficient evidence, the belief[s] should be deleted.
  • No two rationalists can agree to disagree. This is a symptom of disagreeing on the facts. Check your definitions.
  • Focus on narrow, small steps of inference. It may seem wise and impressive to generalize but it is much harder to prove, or identify logical errors while trying to do so.
  • “A hypothesis that forbids nothing, permits everything, and thereby fails to constrain anticipation. Your strength as a rationalist is your ability to be more confused by fiction than by reality. If you are equally good at explaining any outcome, you have zero knowledge.” – Yudkowsky
  • Absence of evidence is evidence of absence. A state of the world is less likely if we see no evidence for it.
  • The expectation of the posterior probability (right side), after viewing the evidence, must equal the prior probability (left side). Evidence cannot be one-sided, it cannot just confirm or just deny.
  •  P(H) = P(H|E)*P(E) + P(H|~E)*P(~E). This is important:

If you expect a strong probability of seeing weak evidence in one direction, it must be balanced by a weak expectation of seeing strong evidence in the other direction.  If you’re very confident in your theory, and therefore anticipate seeing an outcome that matches your hypothesis, this can only provide a very small increment to your belief (it is already close to 1); but the unexpected failure of your prediction would (and must) deal your confidence a huge blow.  On average, you must expect to be exactly as confident as when you started out.  Equivalently, the mere expectation of encountering evidence – before you’ve actually seen it – should not shift your prior beliefs.  –Yudkowsky

  • Hindsight bias is hard to deal with. We need to imagine a world before the evidence came in to truly admire the value of the science.
  • An explanation is not an explanation if it does not constrain the probability space. Using scientific words does not add value if the meaning is not understood. Do not merely accept the naming of a scientific phenomena or theory as an explanation; use it’s implications to see if the evidence can truly be explained by the specific theory or system (and not just anything, “magic”)

Self-awareness as a level of sentience

Can we define physically independent entities that experience self-awareness as separate highly sentient beings *worth caring about*? Humans possess a dense cluster of nerve cell in their brain with the comparative minority spread out throughout the body and gut. Throughout our life almost all of our cells die and are replaced, the exception being those nerve cells. Nevertheless, it is fair to treat neurons as substrate upon which information is encoded rather than something that is a timeless definition of “who we are.”

The only similarity between the same person at an early age and a late age is their DNA (encoding with too much compression to define an independent sentient creature, so meaningless for this exercise), much of their nervous system (many cells don’t regrow — although technology is changing that — and can live until the rest of the body dies), and some of the information encoded in their cortical connections. Without a concept of self; particularly an ability to generate narrative and a timeline which connects the past self to the future self, I propose the being has no concept of individuality or on-going existence!

This raises some interesting conclusions. People with surgically split brains can be considered two independent creatures (they are not completely independent since they are physically confined within the same system — their body — but this doesn’t change the fact that they will be able to function independently on a neurological level which is all that matters). More importantly, a lack of self-awareness may imply that a creature is unable to distinguish between it’s selves over time, let alone between itself and others. If this can be described as a lack in ability to narrate one’s life, it seems to prevent the experience of high level emotions such as suffering, joy, etc. This would also imply that it is not morally wrong to kill such creatures. Interestingly, it seems that humans fall into this category when they are at an early enough age, although the future of said human involves an ascension towards self-awareness. We can say that it is morally justifiable to kill cows since they cannot achieve self-awareness, but not morally justifiable to kill baby humans since they will achieve self-awareness in a future state (although it is maybe more morally justifiable than killing an already self-aware human?).

Transaction costs

Noah seems to argue that taking property rights to extremes and privatizing everything is the libertarian dream, will increase efficiency, but will also introduce onerous monetary and psychological transaction costs and lead to a feeling of less freedom, rather than the desired increase. This seems like a straw man to me. No economist will argue for such extreme levels of privatization precisely because it does not increase efficiency. We know that in practice, public goods and pseudo-public goods (public trash cans and benches, parks, etc.) tend to be more efficiently provisioned by government. Largely because of the difficulty of making capital allocation decisions at such a high frequency and for such small amounts as would be required if these were provided for privately (assuming current technology).

However, we can imagine a world where our usage of semi-public goods is metered unobtrusively. In exchange for paying less in income taxes, we would be taxed proportional to our usage of roads (already in testing in Netherlands), parks and sidewalks just as we now pay for water and electricity (yes these are provided by pseudo-private for-profit institutions, but the structure and regulation of the industries results in a market that mimics government-provided utilities). If we can have a sufficiently transparent and trustworthy infrastructure to meter our usage of such things, we’d have no problem paying for them one time at the end of the year. To eliminate the remaining potential for psychological costs when determining whether to allocate capital to such small privileges as walking on the sidewalk, we can simply make these taxes/costs steeply progressive (since the end of year bill will come with knowledge of the user’s wealth). This would eliminate most of the problems cited by Noah and I believe is coming soon to first world nations.

What do tigers dream of

http://www.nytimes.com/2011/07/31/books/review/perplexities-of-consciousness-by-eric-schwitzgebel-book-review.html?_r=1&partner=rss&emc=rss&pagewanted=all

He begins with the curious case of color in dreams. When people today are asked whether they regularly dream in color, most say they do. But it was not always so. Back in the 1950s most said they dreamed in black and white. Presumably it can hardly be true that our grandparents had different brains that systematically left out the color we put in today. So this must be a matter of interpretation. Yet why such freedom about assigning color? Well, try this for an answer. Suppose that, not knowing quite what dreams are like, we tend to assume they must be like photographs or movies — pictures in the head. Then, when asked whether we dream in color we reach for the most readily available pictorial analogy. Understandably, 60 years ago this might have been black-and-white movies, while for most of us today it is the color version. But, here’s the thing: Neither analogy is necessarily the “right” one. Dreams don’t have to be pictures of any kind at all. They could be simply thoughts — and thoughts, even thoughts about color, are neither colored nor non-colored in themselves.

Econofail

Modern economic practice appears to be doomed due to our artificially high discount rates on the future. Simply put, people that aren’t alive yet can’t participate in pricing at the present. This primarily effects depletion rates of things that take a long time to naturally regenerate, such as the environment, natural resources, sovereign credit, etc. Individuals fail to price in the preferences of great-great-grandchildren, let alone the soon-to-be massive developed middle class of China and India. In theory, an artificial intelligence could solve our problem, assuming friendliness is possible and is achieved. Somewhat paradoxically, it seems our pricing would improve and thereby reduce the chances of market failure and extinction (due to running out of resources before we can leave the planet) by extending lifespans of living humans.

It is important to remember that the dwindling supply of a resource does not cause a problem as long as a market has time and ability to function freely and prices adjust to approximate value. The market will never run out of supply if there exists a demand and pricing is able to incorporate all market participants. See: whale oil

Causes of War

This kind of stuff makes me think the next (last?) world war will surely be triggered by a tragedy of the commons scenario — most likely environmental, or possibly resource scarcity. People thought another “great war” was unthinkable after WW1, just like our current view of major international conflict post-WW2.

Population, technology and growth

I have been spending a fair amount of time thinking about the costs and benefits of population changes and Tyler Cowen’s examination into the effect of population on technology is an interesting one. I like his immediate focus on the impact on growth rather than technology and ideas  just for their own sake (largely useless in the grand scheme of things).

One variable that seems oddly left out is the effect of availability and efficient allocation of capital. The data and theory behind investment and growth are pretty solid. And what kinds of things make investment and efficient allocation of capital possible? Population is largely a prerequisite, since it allows for specialization of labor and increases efficiency to the point where we can stop chasing animals around all day and instead plant crops (that was the first step). But it’s not sufficient. We also need a sufficiently liberal governing body to protect the physical and intellectual interests of the people without excessive interference into the allocation of capital. This seems to explain relatively cleanly why some states experience more growth than others.

Long term impact of technology

Bookstaber is on point as usual in his post about the impact of technology on our lives. His conclusion is obvious to most technologists and futurists, I would bet, and its refreshing to see a policy maker having this in mind. He is far from the first person to note the massive collapse in lifestyle difference between the extremely rich, merely well off and even the middle class. We spend most of our time with the same cheap gadgets and electronic toys (well, most people do). Wealth doesn’t buy you the enormous conveniences and life improvements that it did 100 or even 40 years ago.

As simulation and brain stimulation technology improves exponentially over the next 20-40 years, we will have the ability to cheaply simulate any stimulus we like directly into our brains; the sci-fi dream machine. This is wireheading in a pretty pure form, and most people will eat it up. Just as Africa jumped from lack of electricity directly to cell phones, the emerging middle classes will likely be able to skip Playstations and cocaine directly to total immersion pleasure machines.

The real thing will take a long time to flesh out and come down in price enough for the masses, but here’s an example of where we’re headed in the interim. Not much different in practice.

Self

I definitely agree with Justin on sense of self. There is no real self. We think and act as a function of our past experiences and current real-time stimulus. We are constantly in a state of change, and present a different side of ourselves in different environments (with friends, acquaintances, strangers).

A life strategy of self-improvement (which is essentially just changing our ‘self’ to correspond to something we aspire to) appears to be highly correlated with happiness and even health. While the ideal situation seems to be to become content with what we currently have and who we are (to disassociate from a sense of self altogether), the human brain is very easily pleased — and perhaps thats the problem, it’s easy but a short term fix — with incremental improvements and successes. So, if you’re not willing to invest time in meditation and aren’t willing to hone your abilities in self control, you should find parts of yourself that you want to change, enter a strict regimen of incremental improvement, and enjoy the thrills of the small successes.