I haven't come across this particular argument before, so I hope I'm not just rehashing a well-known problem.

"The universe displays some very strong signs that it is a simulation.

As has been mentioned in some other answers, one way to efficiently achieve a high fidelity simulation is to design it in such a way that you only need to compute as much detail as is needed. If someone takes a cursory glance at something you should only compute its rough details and only when someone looks at it closely, with a microscope say, do you need to fill in the details.

This puts a big constraint on the kind of physics you can have in a simulation. You need this property: suppose some physical system starts in state x. The system evolves over time to a new state y which is now observed to accuracy ε. As the simulation only needs to display the system to accuracy ε the implementor doesn't want to have to compute x to arbitrary precision. They'd like only have to compute x to some limited degree of accuracy. In other words, demanding y to some limited degree of accuracy should only require computing x to a limited degree of accuracy.

Let's spell this out. Write y as a function of x, y = f(x). We want that for all ε there is a δ such that for all x-δ<y<x+δ, |f(y)-f(x)|<ε. This is just a restatement in mathematical notation of what I said in English. But do you recognise it?

It's the standard textbook definition of a Continuous function. We humans invented the notion of continuity because it was an ubiquitous property of functions in the physical world. But it's precisely the property you need to implement a simulation with demand-driven level of detail. All of our fundamental physics is based on equations that evolve continuously over time and so are optimised for demand-driven implementation.

One way of looking at this is that if y=f(x), then if you want to compute n digits of y you only need a finite number of digits of x. This has another amazing advantage: if you only ever display things to a given accuracy you only ever need to compute your real numbers to a finite accuracy. Nature could have chosen to use any number of arbitrarily complicated functions on the reals. But in fact we only find functions with the special property that they need only be computed to finite precision. This is precisely what a smart programmer would have implemented.

(This also helps motivate the use of real numbers. The basic operations on real numbers such as addition and multiplication are continuous and require only finite precision in their arguments to compute their values to finite precision. So real numbers give a really neat way to allow inhabitants to find ever more detail within a simulation without putting an undue burden on its implementation.)

But you can do one step further. As Gregory Benford says in Timescape: "nature seemed to like equations stated in covariant differential forms". Our fundamental physical quantities aren't just continuous, they're differentiable. Differentiability means that if y=f(x) then once you zoom in closely enough, y depends linearly on x. This means that one more digit of y requires precisely one more digit of x. In other words our hypothetical programmer has arranged things so that after some initial finite length segment they can know in advance exactly how much data they are going to need.

After all that, I don't see how we can know we're not in a simulation. Nature seems cleverly designed to make a demand-driven simulation of it as efficient as possible."

http://www.quora.com/How-do-we-know-that-were-not-living-in-a-computer-simulation/answer/Dan-Piponi

Help a brother out?

There's a great quote by one of The Big 3 Greek Philosophers (EDIT: Reference to Cicero removed) which I can paraphrase by memory as:

"I consider it rather better for myself to be proven wrong than to prove someone else wrong, just as I'm better off being cured of a disease than curing someone of one."

I can't find the quote, or from which of the Three it is.

Anybody know? Or know where to look? I've already tried varying google search techniques and perused the Wikiquotes article on each of them.

I found a website called Happier Human. It's about how to become and stay happier. I've trawled through it. Here are the best posts in my opinion:

[Meditate]. Don't [worry/overthink/fantasise/compare]. [Disregard desire]. [Motivate]. [Exercise gratitude]. [Don’t have kids].

[Buy many small gifts]. [Trade some happiness for productivity]. [Set] [happiness goals]

If you've found any other happiness interventions on any website, please share them.

A putative new idea for AI control; index here.

This is a problem that developed from the "high impact from low impact" idea, but is a legitimate thought experiment in its own right (it also has connections with the "spirit of the law" idea).

Suppose that, next 1^st of April, the US president may or may not die of natural causes. I chose this example because it's an event of potentially large magnitude, but not overwhelmingly so (neither a butterfly wing nor an asteroid impact).

Also assume that, for some reason, we are able to program an AI that will be nice, given that the president does die on that day. Its behaviour if the president doesn't die is undefined and potentially dangerous.

Is there a way (either at the initial stages of programming or at the later) to extend the "niceness" from the "presidential death world" into the "presidential survival world"?

To focus on how tricky the problem is, assume for argument's sake that the vice-president is a war monger that will start a nuclear war if they become president. Then "launch a coup on the 2^nd of April" is a "nice" thing of the AI to do, conditional on the president dying. However, if you naively import that requirement into the "presidential survival world", the AI will launch a pointeless and counterproductive coup. This is illustrative of the kind of problems that could come up.

So the question is, can we transfer niceness in this way, without needing a solution to the full problem of niceness in general?

EDIT: Actually, this seems ideally setup for a Bayes network (or for the requirement that a Bayes network be used).

EDIT2: Now the problem of predicates like "Grue" and "Bleen" seem to be the relevant bit. If you can avoid concepts such as "X={nuclear war if president died, peace if president lived}", you can make the extension work.

The following is a poem I wrote today. I've been considering poetry that I write of this nature to be of a Reason/Cyberpunk/Transhuman sort of genre. Feedback, including feedback on if there is a place for poetry on this site, would be appreciated.

I forever wish to change from who I am today,

Yet as I am today, I do not wish to cease.

Who am I in this moment?

I am nothing to myself without the passage of time

If I had no fear of death,

Would I have a wish to live?

I can deny cynicism.

Can I verify optimism?

Must euphoria define my goals?

Every euphoric drive has served to continue my existence.
From the beginning mechanisms of life, I have emerged

Passed through millions/billions of small keyholes of existence

A package of information, which served to create me

Developed me to fit my environment.

Existing just to continue to exist.

An axiom of my function

Euphoria drives me

Skepticism contradicts me

I cannot withhold judgement on the purpose of existing.

To enjoy the show is to accept this euphoria as my chosen purpose in the end.

Can I want without pleasure?

Can my wants be reasoned?

Why do I want to enjoy the show,

Yet not to be consumed or confined to an eternity of bliss?

Is dignity and pride different from euphoric drives?

Are they the strategies and philosophies of my existence?

Can I be more obsessed with finding the perfect design for myself,

Than with finding bliss? Are they functionally different?

Summary

I'd like to increasing the well-being of those in the justice system while simultaneously reducing crime. I'm missing something here but I'm not sure what. I'm thinking this may be a worse idea than I originally thought based on comment feedback, though I'm still not 100% sure why this is the case.

Current State

While the prison system may not constitute an existential threat, At this moment more than 2,266,000 adults are incarcerated in the US alone, and I expect that being in prison greatly decreases QALYs for those incarcerated, that further QALYs are lost to victims of crime, family members of the incarcerated, and through the continuing effects of institutionalization and PTSD from sentences served in the current system, not to mention the brainpower and man-hours lost to any productive use.

If you haven't read these Meditations on Moloch, I highly recommend it. It’s long though, so the executive summary is: Moloch is the personification of the forces of competition which perverse incentives, a "race to the bottom" type situation where all human values are discarded in an effort to survive. That this can be solved with better coordination, but it is very hard to coordinate when perverse incentives also penalize the coordinators and reward dissenters. The prison industrial complex is an example of these perverse incentives. No one thinks that the current system is ideal but incentives prevent positive change and increase absolute unhappiness.

• Politicians compete for electability. Convicts can’t vote, prisons make campaign contributions and jobs, and appearing “tough on crime” appeals to a large portion of the voter base.
• Jails compete for money: the more prisoners they house, the more they are paid and the longer they can continue to exist. This incentive is strong for public prisons and doubly strong for private prisons.
• Police compete for bonuses and promotions, both of which are given as rewards to cops who bring in and convict more criminals
• Many of the inmates themselves are motivated to commit criminal acts by the small number of non-criminal opportunities available to them for financial success, besides criminal acts. After becoming a criminal, this number of opportunities is further narrowed by background checks.

The incentives have come far out of line with human values. What can be done to bring incentives back in alignment with the common good?

My Proposal

Using a model that predicts recidivism at sixty days, one year, three years, and five years, predict the expected recidivism rate for all inmates at all individual prison given average recidivism. Sixty days after release, if recidivism is below the predicted rate, the prison gets a small sum of money equaling 25% of the predicted cost to the state of dealing with the predicted recidivism (including lawyer fees, court fees, and jailing costs). This is repeated at one year, three years, and five years.

The statistical models would be readjusted with current data every years, so if this model causes recidivism to drop across the board, jails would be competing against ever higher standard, competing to create the most innovative and groundbreaking counseling and job skills and restorative methods so that they don’t lose their edge against other prisons competing for the same money. As it becomes harder and harder to edge out the competition’s advanced methods, and as the prison population is reduced, additional incentives could come by ending state contracts with the bottom 10% of prisons, or with any prisons who have recidivism rates larger than expected for multiple years in a row.

Note that this proposal makes no policy recommendations or value judgement besides changing the incentive structure. I have opinions on the sanity of certain laws and policies and the private prison system itself, but this specific proposal does not. Ideally, this will reduce some amount of partisan bickering.

Using this added success incentive, here are the modified motivations of each of the major actors.

• Politicians compete for electability. Convicts still can’t vote, prisons make campaign contributions, and appearing “tough on crime” still appeals to a large portion of the voter base. The politician can promise a reduction in crime without making any specific policy or program recommendations, thus shielding themselves from criticism of being soft on crime that might come from endorsing restorative justice or psychological counselling, for instance. They get to claim success for programs that other people, are in charge of administrating and designing. Further, they are saving 75% of the money predicted to have have been spent administrating criminals. Prisons love getting more money for doing the same amount of work so campaign contributions would stay stable or go up for politicians who support reduced recidivism bonuses.
• Prisons compete for money. It costs the state a huge amount of money to house prisoners, and the net profit from housing a prisoner is small after paying for food, clothing, supervision, space, repairs, entertainment, ect. An additional 25% of that cost, with no additional expenditures is very attractive. I predict that some amount of book-cooking will happen, but that the gains possible with book cooking are small compared to gains from actual improvements in their prison program. Small differences in prisons have potential to make large differences in post-prison behavior. I expect having an on-staff CBT psychiatrist would make a big difference; an addiction specialist would as well. A new career field is born: expert consultants who travel from private prison to private prison and make recommendations for what changes would reduce recidivism at the lowest possible cost.
• Police and judges retain the same incentives as before, for bonuses, prestige, and promotions. This is good for the system, because if their incentives were not running counter to the prisons and jails, then there would be a lot of pressure to cook the books by looking the other way on criminals til after the 60 day/1 year/5 year mark. I predict that there will be a couple scandals of cops found to be in league with prisons for a cut of the bonus, but that this method isn’t very profitable. For one thing, an entire police force would have to be corrupt and for another, criminals are mobile and can commit crimes in other precincts. Police are also motivated to work in safer areas, so the general program of rewarding reduced recidivism is to their advantage.

Roadmap

If it could be shown that a model for predicting recidivism is highly predictive, we will need to create another model to predict how much the government could save if switching to a bonus system, and what reduction of crime could be expected.

Halfway houses in Pennsylvania are already receiving non-recidivism bonuses. Is a pilot project using this pricing structure feasible?

http://www.theatlantic.com/technology/archive/2015/05/immortal-but-damned-to-hell-on-earth/394160/

With such long periods of time in play (if we succeed), the improbable hellish scenarios which might befall us become increasingly probable.

With the probability of death never quite reaching 0, despite advanced science, death might yet be inevitable.

But the same applies also to a hellish life in the meanwhile. And the longer the life, the more likely the survivors will envy the dead. Is there any safety in this universe? What's the best we can do?

These are quick notes on an idea for an indirect strategy to increase the likelihood of society acquiring robustly safe and beneficial AI.

Motivation:

• Most challenges we can approach with trial-and-error, so many of our habits and social structures are set up to encourage this. There are some challenges where we may not get this opportunity, and it could be very helpful to know what methods help you to tackle a complex challenge that you need to get right first time.

• Giving an artificial intelligence good values may be a particularly important challenge, and one where we need to be correct first time. (Distinct from creating systems that act intelligently at all, which can be done by trial and error.)

• Building stronger societal knowledge about how to approach such problems may make us more robustly prepared for such challenges. Having more programmers in the AI field familiar with the techniques is likely to be particularly important.

Idea: Develop methods for training people to write code without bugs.

• Trying to teach the skill of getting things right first time.

• Writing or editing code that has to be bug-free without any testing is a fairly easy challenge to set up, and has several of the right kind of properties. There are some parallels between value specification and programming.

• Set-up puts people in scenarios where they only get one chance -- no opportunity to test part/all of the code, just analyse closely before submitting.

  • Interested in personal habits as well as social norms or procedures that help this.

    • Daniel Dewey points to standards for code on the space shuttle as a good example of getting high reliability code edits.

How to implement:

• Ideal: Offer this training to staff at software companies, for profit.

  • Although it’s teaching a skill under artificial hardship, it seems plausible that it could teach enough good habits and lines of thinking to noticeably increase productivity, so people would be willing to pay for this.

  • Because such training could create social value in the short run, this might give a good opportunity to launch as a business that is simultaneously doing valuable direct work.

  • Similarly, there might be a market for a consultancy that helped organisations to get general tasks right the first time, if we knew how to teach that skill.

• More funding-intensive, less labour intensive: run competitions with cash prizes

  • Try to establish it as something like a competitive sport for teams.

  • Outsource the work of determining good methods to the contestants.

This is all quite preliminary and I’d love to get more thoughts on it. I offer up this idea because I think it would be valuable but not my comparative advantage. If anyone is interested in a project in this direction, I’m very happy to talk about it.

How does using priors affect the concept of statistical significance? The scientific convention is to use a 5% threshold for significance, no matter whether the hypothesis has been given a low or a high prior probability.

If we momentarily disregard the fact that there might be general methodological issues with using statistical significance, how does the use of priors specifically affect the appropriateness of using statistical significance?

What comes to mind when you hear the word ‘oxytocin?’ Is it ‘love’, ‘cuddle hormone’, ‘bliss?’ If so, you may be more aware of the Dr. Jekyll of oxytocin rather than the Mr. Hyde. Oxytocin, just like almost every biochemical molecule, is hormetic. It confers positive effects in one context, but negative in another. In the case of oxytocin, a person with a secure attachment style interacting with a familiar group of people that he/she likes, will experience the positive effects of oxytocin. However, someone with an anxious attachment style interacting with a group of people that he/she does not yet fully feel trusting and familiar with will experience the negative effects of oxytocin. Why does the same molecule produce pro-social effects for one person, yet anti-social for another?

Oxytocin redirects more attentional resources towards noticing social stimuli. This increase in the salience of social information enhances the ability to detect expressions, recognize faces, and other social cues. The effect of increased social cognitive abilities is constrained by personality traits and situational context, resulting in either anti-social or pro-social behavior.

Oxytocin also promotes more interest in social cues by increasing affiliative motivation, a desire to get along with others. The increase in affiliative motivation results in pro-social behavior if the person already tends towards having an interest in bonding with people outside their close friend circle. However, an increase in affiliative motivation for those with anxious attachment styles results in a stronger pursuit to feel closer to only the person he/she is attached to.

A couple, Tom and Mary, have just moved to a new town and are attending their first service at a new church. Tom has a secure attachment style and isn’t prone to social anxieties. Tom is optimistic, has a positive bias, is generally content, and sees people as good, trusting, and friendly. Mary has an anxious attachment style, a negative bias, social anxiety, baseline mood neutral, and sees people as potential threats, competitors, untrustworthy, selfish, and egotistical. During the service, Tom and Mary’s oxytocin levels increase by being in a community. As a result of their different dispositions, Tom exhibits the Dr. Jekyll of oxytocin, whereas Mary exhibits the Mr. Hyde.

At the end of the service, Mary determines that she doesn’t like the church, whereas Tom thinks it is perfect. Mary felt that the people were judgmental and that they didn’t like her and Tom. Tom felt that the people were friendly, accepting, and eager for them to join.

Most social cues are ambiguous. A person’s character traits are instrumental in  interpreting the cues as negative or positive. Tom is more likely to interpret facial expressions as positive, whereas Mary sees them as negative. Tom interprets neutral expressions to indicate acceptance, kindness, and friendliness. Mary sees neutral expressions as judgmental and unkind. This creates a fear of rejection, feeling threatened, and propagates a negative bias.

The increase in oxytocin leads to quicker detection and interpretation of facial expressions. Interpreting inchoate facial expressions fosters interpretations based on expectations versus what is actually intended. A person is starting to smile, but before the smile is developed, Mary believes that the person is about to laugh and ridicule her. Mary then scowls at her, turning what was going to be a smile into a negative expression. Tom interprets the inchoate expression as a smile, smiles, and turns the inchoate expression into a genuine smile.

Oxytocin amplifies one’s character traits of pro-social or anti-social tendencies. Oxytocin does increase the feelings of bonding for all, but in different ways. People with pro-social tendencies will feel closer to their communities and greater circle of friends. People with anti-social tendencies will just feel closer to their close circle of friends and people they already trust.

Cross-posted from my blog: https://evolvingwithtechnology.wordpress.com. 

References:

http://dept.psych.columbia.edu/~kochsner/pdf/Bartz_et_al_2011_Social_oxytocin.pdf

http://www.attachedthebook.com/about-the-book/ by Amir Levine and Rachel Heller.