Westworld is not smart

The premise of the show is that there is a themepark, "westworld", in America in the near future. It’s a simulation of the imagined American wild west. Rich tourists go there. They either play white hats, good guys, or black hats, bad guys. The bad guys can rape and kill and torture the inhabitants of the park. How is this legal? The natives aren’t human actors, they’re robots who look human and, after the themepark resets, loose their memories of the past cycles events. The show does a lot of oooh-aaah about wheather the robots are alive and wheather torturing them is wrong etc..

Why is the show stupid? Because

  • It ignore moral uncertainty. The robots often look, feel and act like real people. As the narrative develops, they increasingly exhibit autonomous behaviours, self-awareness and other traits we associate with people. Even if we can’t be sure they they qualify as moral agents, it still seems to be highly likely or at least possible. Hence killing/torturing maiming them is wrong. There’s an (risk of them being people)% chance that you’re doing horrific things to actual people. A good comparison here is you see a person in a coma. You can choose to shoot them, You are 97% certain they’re braindead. Obviously you shouldn’t shoot them. The 3% risk that you run of commiting murder is too high.
  • It treats moral personhood as binary. The presumption in the show is that if the robots aren’t full people, it’s okay to torture/kill/rape/murder them. Why do they have to be 100% equivalent to humans for it to be bad to torture them? Animals are not people. We still don’t allow their torture. The robots certainly seem more aware and intelligent than animals, The point is there are levels of personhood in between "Rock" and "Sentient human being", those levels still have certain rights/preferences we care about and there’s no way in hell the robots appear to have less personhood than a pig or cow.
  • The world is really, really dumb. There exist super smart robots who have a great deal of decision making power, are autonomous, can be programmed to do many things (including shoot?). All they’re used for is theme parks. Bullshit. This kind of tech would change the face of the world in a decade. All manual or low-creativity jobs could be automated. Wars could be fought mostly without human intervention. Authoritarian regimes would have even less need of a happy populace, knowing their soldiers are always loyal and willing regardless of what the drones think. Hell, it’s literally a singularity level event which could entirely reshape/destroy human society. And that’s assuming the robots can just follow basic scripts and don’t have agency. If they do and if their intelligence can be adjusted to be super human, as the shows says it can, ooooohhhh boy.

Trade-off denialism

Any working ethical system that actually works for human morality will entail trade-off’s. There are cases where different moral values conflict. You will encounter them often. You will have to make choices. What’s interesting is that when talking to most normal people who I’ve worked with, they’re extremely reluctant to discuss or recognize the need to make these kinds of trade-off’s. This may be a more specific case of the general aversion to moral awareness and the need for judgment both of your self and of others that it entails but I think it’s something else. Even when discussing specifically moral subjects on which people are happy to stake out positions, still no one recognizes trade-off’s have to be made.

I wonder why? A few hypothesises:

  • Trade-off’s are socially costly/risky to discuss vs applause lights statements like “we should do good thing X we all agree on”
  • Trade-off’s require a greater degree of ethical thought/development than simple black/white rules
  • Most people either have no coherent moral framework or in some cases are developed enough to have a simple singular conception of the good. Very few get past that stage to the point where they have multiple competing values which they trade-off against each other.

Against Loyalty

Normal people value loyalty, practically and morally. They’re usually wrong to do so. My loyalty hierarchy at work goes like this:

  1. Loyalty to my ethics
  2. Loyalty to my clients
  3. Loyalty to my firm

Where a course of action contradicts my ethics, I will not do it consequences be dammed. Where I can tell a client what they want to hear, helping my firm get contracts but harming the client, I won’t. What’s interesting about this kind of ethical thinking is that it works just as well if you remove the word “loyalty”.

What loyalty means is acting differently towards someone because of your relationship or history with them. This is a good strategy for rent-seeking, but wrong morally and bad for the collective. You should treat people equally and fairly. This means telling people you’re “loyal” to that they’re wrong and standing against them in meetings when necessary, even if they’re your boss. This means firing or demoting low-performers, even if they’re your friends. Doing otherwise may be better for your career in most places (most firms are dysfunctional), but it’s not right.

I don’t know enough about Damore’s case to make a judgement

What I read about Damore disturbed me. He was punished for his memo because he suggested there are biological differences between men and women make women less likely to be good at IT. I personally wouldn’t do that, but I think firms should be free to set their own speech standards. As long as those standards are symmetric[^1], it’s something I dislike but do not think is morally impermissible. But the response does not seem to have been symmetric. People who wrote to Darmore internally and threatened him or insulted him were not punished or fired. He was. This seems wrong.

Then again, I don’t have clear evidence on the matter. Digging into it more, I find that some employees who posted insulting things about him were punished by HR. Reality isn’t as black and white as it first appears.

It’s scary how easy it is to fall into stories or narratives which vindicate existing beliefs. I feel like i want to believe Damore was punished unjustly and silenced, that google was being unfair, That feeling is wrong. I don’t have the evidence to make that kind of judgement.

I continue to believe that anti-white discrimination at large tech firms is bad, but I don’t know if their treatment of Damore was unfair.

[^1]: Symmetric bans are those which affect both sides of the discussion equally. So banning discussion of an election is okay. Banning discussion of one party in said election but allowing other parties is not okay.

Against Bryan Caplan on Raising Children

In "Selfish Reasons to Have More Kids", Bryan Caplan argues that raising children is easier than we think. That twin and adoption studies show that parenting, unless is spectacularly bad, doesn’t have that much impact. That our children are less like clay that we mould and more like plastic that snaps back into it’s natural shape when we relax our grip. He goes from this claim to the conclusion that most of us should probably have more kids. The conclusion I agree with, at least for the demographic his book targets, The premise I’m skeptical of.

1: Metrics for a good life

If you look at how he measures life quality, it’s the standard indicators. Class. Wealth. Profession. Education. Health. That’s not the measure of a life. The measure of a life is character and ethics.

If you disagree with his metric of child success, his argument is less persuasive. He presents no evidence that parenting can’t improve character or ethics.

The problem with Bryans book is that the metrics he uses reflect the people he seeks to convince. Most people are deeply amoral in their thinking and so by extension is his conception of "good" parenting.

2: This may not apply to outlier parenting

Most parents parent in certain ways. Twin studies will reflect this. It seems plausible that outlier parenting methods could be effective at drastically altering outcomes. If that was the case, the data probably wouldn’t show it due to how rare those methods are.

On the low end this seems obviously true. Serious abuse and neglect can and do drastically impact life outcomes. (Then again, this may be genetic)

On the high end it could be true. We don’t know. Maybe there are certain high impact parenting methods which can have lasting effects.

3: I should actually read the book

The wishes of the dead

Our society limits people’s scope of action to their lives. We ban post-death trusts, or at least do not give them the protection of the legal system. We do not allow the creation of corporations/organizations without a terminal human owner. We limit the ability of the past to control the present and future. No hands from beyond the grave.

Are we right to do so? I don’t know. Social questions are difficult. Every decisions spirals into a fractal of consequences and interactions, the tiniest change in initial circumstances creating a different fracture, different patterns and colours. (Not all effects are chaotic. Some things have predictable effects.)

Morally, I do not see why those who exist should have more rights than those who do not. Why the will of the living should matter more than that of the dead or unborn. Existence is morally arbitrary. The reasons we privilege it are practical, not pure.

On fat acceptance

Some people are uncritically in favour of the fat acceptance movement/narrative. Some are against. I think both are wrong. The fat acceptance movement has two independent messages. The first is that people who are fat should not feel ashamed, be discriminated against or be seen as less than thin people. This is correct. The second is that being fat is not bad for your health. This is incorrect.