[MG] Axiom 6: People are not competent to choose their proxies

[Scott Raney]

One of the things Kling said has kind of stuck in my craw:

"We discussed this and came to the conclusion that there are
other factors than representativeness which play a role when delegating
a vote, i.e. sympathy and popularity."

I'm sensing basically the same view from Carlo and Sergey, although
they might claim other factors, such as "expertise", as being
important. I'm going to call my Axiom 5, which states that people
should be represented by direct voters who will vote like they would
if they had the time/motivation/experience/etc. to properly
investigate the issue before voting, the "representationist" position.
I'll call the opposite position "elitism" because it implies that
there is something special about those people you propose to want to
delegate to (if you have an alternative proposal for what to call it,
by all means make it).

I frequently point to demagogues as an example of a fatal flaw in the
elitist position, but what exactly is the problem with demagogues? Is
it just that they're racist, nationalistic, and authoritarian? Not at
all! If that were the real problem, then what we should be doing is
designing social engineering countermeasures to those things, which we
could then apply not only to the demagogues but to anyone else who
might vote that way, regardless of whether they were being led by a
demagogue or not.

No, the problem with demogogues is their ability to *amplify* their
position on any and all issues by convincing others, who may or not
agree with the demagogue on any *specific* issue, to effectively grant
a blanket proxy to the demagogue. But what about using "expertise" or
"popularity" or "familiarity" or some other criteria to select a
proxy? That should be OK, right? No, because it has exactly the same
fatal flaw: By granting a proxy to someone based on any single
characteristic and exclusive of knowing their actual personality type,
it is virtually guaranteed that the proxy will distort the voting on a
wide range of issues because of their ability to amplify their votes,
which will over-represent their personality type in the final decision
and which is essentially taking advantage of those who have granted
them a proxy unless the elite just coincidentally happens to cast a
true representationist vote.

All of course is *in addition to* the other problems I've pointed out
with a system that allows granting proxies to specific individuals:
1) The possibility of dead-ending and no vote being cast.
2) The possibility of corruption because the elite then becomes a
target for manipulation.
3) The possibility of corruption due to the elite casting a self-serving vote.

At this point I have to recommend that we equate the elitist position
with those who think they need to have the ability to disable
stability control on their cars, a position that even the experts in
that field consider to be just plain stupid, even on a race track, let
alone out in traffic:
http://www.roadandtrack.com/car-culture/a28714/turning-off-stability-control-is-one-of-the-dumbest-things-you-can-do-in-a-car/

Bottom line, I'm walking back on my proposal to even *calculate* a
vote from those one follows, let alone allowing automatically casting
that vote. This still allows people to make the calculation manually
by actually looking at the votes of those they're following and then
manually voting, but requiring even this extra effort should
discourage almost everyone from doing so, which of course is in their
best interest if what we're trying to build is a representationist
system. I still think it's a good idea to allow blocking of specific
individuals from the AutoMatch, but my current thinking is to have
that be a separate system from the "follow" feature.

Anyone care to try to defend the elitist position?
Regards,
Scott

[Patrick Millerd]

I might consider implementing a warning system too against anyone trying to
game the system by misrepresenting themselves in the personality test. If a
certain number of users switch their vote from the one the proxy chooses,
the system alerts the other auto-matched users that this proxy might not be
representing them.

Post by Scott Raney
"We discussed this and came to the conclusion that there are
other factors than representativeness which play a role when delegating
a vote, i.e. sympathy and popularity."
I'm sensing basically the same view from Carlo and Sergey, although
they might claim other factors, such as "expertise", as being
important. I'm going to call my Axiom 5, which states that people
should be represented by direct voters who will vote like they would
if they had the time/motivation/experience/etc. to properly
investigate the issue before voting, the "representationist" position.
I'll call the opposite position "elitism" because it implies that
there is something special about those people you propose to want to
delegate to (if you have an alternative proposal for what to call it,
by all means make it).
I frequently point to demagogues as an example of a fatal flaw in the
elitist position, but what exactly is the problem with demagogues? Is
it just that they're racist, nationalistic, and authoritarian? Not at
all! If that were the real problem, then what we should be doing is
designing social engineering countermeasures to those things, which we
could then apply not only to the demagogues but to anyone else who
might vote that way, regardless of whether they were being led by a
demagogue or not.
No, the problem with demogogues is their ability to *amplify* their
position on any and all issues by convincing others, who may or not
agree with the demagogue on any *specific* issue, to effectively grant
a blanket proxy to the demagogue. But what about using "expertise" or
"popularity" or "familiarity" or some other criteria to select a
proxy? That should be OK, right? No, because it has exactly the same
fatal flaw: By granting a proxy to someone based on any single
characteristic and exclusive of knowing their actual personality type,
it is virtually guaranteed that the proxy will distort the voting on a
wide range of issues because of their ability to amplify their votes,
which will over-represent their personality type in the final decision
and which is essentially taking advantage of those who have granted
them a proxy unless the elite just coincidentally happens to cast a
true representationist vote.
All of course is *in addition to* the other problems I've pointed out
1) The possibility of dead-ending and no vote being cast.
2) The possibility of corruption because the elite then becomes a
target for manipulation.
3) The possibility of corruption due to the elite casting a self-serving vote.
At this point I have to recommend that we equate the elitist position
with those who think they need to have the ability to disable
stability control on their cars, a position that even the experts in
that field consider to be just plain stupid, even on a race track, let
http://www.roadandtrack.com/car-culture/a28714/turning-
off-stability-control-is-one-of-the-dumbest-things-you-can-do-in-a-car/
Bottom line, I'm walking back on my proposal to even *calculate* a
vote from those one follows, let alone allowing automatically casting
that vote. This still allows people to make the calculation manually
by actually looking at the votes of those they're following and then
manually voting, but requiring even this extra effort should
discourage almost everyone from doing so, which of course is in their
best interest if what we're trying to build is a representationist
system. I still think it's a good idea to allow blocking of specific
individuals from the AutoMatch, but my current thinking is to have
that be a separate system from the "follow" feature.
Anyone care to try to defend the elitist position?
Regards,
Scott
_______________________________________________
Start : a mailing list of the Metagovernment project
http://www.metagovernment.org/
Manage subscription: http://metagovernment.org/mailman/listinfo/start_
metagovernment.org

[Paul Sabin]

Hi all,

Let me present an alternative view of how representation/delegation should
work. This is my own personal take on the matter, but I believe it has a
lot in common with the liquid feedback approach.

1) You get to choose who to delegate to. It's not decided by any kind of
system or algorithm by default. Ideally, you'll feel competent enough to
participate directly. If not, you can delegate. You and only you are
competent enough to choose a delegate. Our decision-making capabilities are
far from perfect, but they're far better than any human-invented algorithm.

2) You delegate in specific areas - no blanket delegations.

3) For every specific area, you can choose to participate directly
yourself, delegate, or do nothing at all. You can change this choice at any
time.

4) You can base your choice of who to delegate to on any criteria you like.
Maybe it's because you think that person is an expert in the field. Maybe
it's because they have a PhD or a lifetime of professional experience.
Maybe not. Maybe you know them personally. Maybe you're confident you share
the same interests. Maybe you've noticed that you tend to agree with their
proposals. It's up to you.

5) Delegations are transitive, so when you delegate to somebody, you're
also trusting their judgement to potentially pass it on to someone else.
If/when this happens, you receive a notification.

6) You can override or revoke delegations at any time. So if there are any
people who have gained a large degree of influence through delegation, they
are always dependent on retaining the trust placed in them by all those who
have chosen to delegate. So there are two major ways that an oligarchy is
prevented from cementing its power at the top: a) division into specific
subject areas, b) potential to revoke delegations.

7) Delegations are open. Everybody can see who's delegating to whom. In the
case of transitive delegations, you can trace the path of delegations as
they are passed from one person to the next.

Now let me make two quick points about the argument as you've laid it out
above, Scott, because I don't believe the distinction you've drawn between
"representationist" and "elitist" holds any water.

On the one hand, you argue that to admit some people understand a certain
issue better than others is elitist. On the other hand, you want people to
be represented by others who have "the time/motivation/experience/etc. to
properly investigate the issue". Yet spending time investigating a subject
and gaining experience in a field - this is precisely how you acquire what
most people call expertise. I'm aware that we live in an age where people
have little faith in "experts" - and often for good reason. But that's not
to say that expertise doesn't exist. What we need to ensure is that experts
- or other people with influence - are acting in good faith. We do this by
designing the system so that they are always reliant on maintaining their
base of support.

Another fundamental problem with this same point:

"people should be represented by direct voters who will vote like they
would if they had the time/motivation/experience/etc. to properly
investigate the issue before voting"

How will we ever know how people would have voted if they *did* have this
motivation, if they *did* spend a lot of time looking into it and if the
*did* have relevant experience? The reason to want to delegate your vote -
on a particular subject, not a blanket delegation - is precisely because
you *don't* have that level of interest/time etc. to devote to that
particular field. You cannot even know yourself how you would vote if you
did. So your definition of "representation" is something unknowable.

Regards,
Paul

Post by Patrick Millerd
I might consider implementing a warning system too against anyone trying
to game the system by misrepresenting themselves in the personality test.
If a certain number of users switch their vote from the one the proxy
chooses, the system alerts the other auto-matched users that this proxy
might not be representing them.

Post by Scott Raney
"We discussed this and came to the conclusion that there are
other factors than representativeness which play a role when delegating
a vote, i.e. sympathy and popularity."
I'm sensing basically the same view from Carlo and Sergey, although
they might claim other factors, such as "expertise", as being
important. I'm going to call my Axiom 5, which states that people
should be represented by direct voters who will vote like they would
if they had the time/motivation/experience/etc. to properly
investigate the issue before voting, the "representationist" position.
I'll call the opposite position "elitism" because it implies that
there is something special about those people you propose to want to
delegate to (if you have an alternative proposal for what to call it,
by all means make it).
I frequently point to demagogues as an example of a fatal flaw in the
elitist position, but what exactly is the problem with demagogues? Is
it just that they're racist, nationalistic, and authoritarian? Not at
all! If that were the real problem, then what we should be doing is
designing social engineering countermeasures to those things, which we
could then apply not only to the demagogues but to anyone else who
might vote that way, regardless of whether they were being led by a
demagogue or not.
No, the problem with demogogues is their ability to *amplify* their
position on any and all issues by convincing others, who may or not
agree with the demagogue on any *specific* issue, to effectively grant
a blanket proxy to the demagogue. But what about using "expertise" or
"popularity" or "familiarity" or some other criteria to select a
proxy? That should be OK, right? No, because it has exactly the same
fatal flaw: By granting a proxy to someone based on any single
characteristic and exclusive of knowing their actual personality type,
it is virtually guaranteed that the proxy will distort the voting on a
wide range of issues because of their ability to amplify their votes,
which will over-represent their personality type in the final decision
and which is essentially taking advantage of those who have granted
them a proxy unless the elite just coincidentally happens to cast a
true representationist vote.
All of course is *in addition to* the other problems I've pointed out
1) The possibility of dead-ending and no vote being cast.
2) The possibility of corruption because the elite then becomes a
target for manipulation.
3) The possibility of corruption due to the elite casting a self-serving vote.
At this point I have to recommend that we equate the elitist position
with those who think they need to have the ability to disable
stability control on their cars, a position that even the experts in
that field consider to be just plain stupid, even on a race track, let
http://www.roadandtrack.com/car-culture/a28714/turning-off-
stability-control-is-one-of-the-dumbest-things-you-can-do-in-a-car/
Bottom line, I'm walking back on my proposal to even *calculate* a
vote from those one follows, let alone allowing automatically casting
that vote. This still allows people to make the calculation manually
by actually looking at the votes of those they're following and then
manually voting, but requiring even this extra effort should
discourage almost everyone from doing so, which of course is in their
best interest if what we're trying to build is a representationist
system. I still think it's a good idea to allow blocking of specific
individuals from the AutoMatch, but my current thinking is to have
that be a separate system from the "follow" feature.
Anyone care to try to defend the elitist position?
Regards,
Scott
_______________________________________________
Start : a mailing list of the Metagovernment project
http://www.metagovernment.org/
Manage subscription: http://metagovernment.org/mail
man/listinfo/start_metagovernment.org

_______________________________________________
Start : a mailing list of the Metagovernment project
http://www.metagovernment.org/
Manage subscription: http://metagovernment.org/mailman/listinfo/start_
metagovernment.org

[Scott Raney]

Post by Paul Sabin
1) You get to choose who to delegate to. It's not decided by any kind of
system or algorithm by default. Ideally, you'll feel competent enough to
participate directly. If not, you can delegate. You and only you are
competent enough to choose a delegate. Our decision-making capabilities are
far from perfect, but they're far better than any human-invented algorithm.

What evidence do you have that people are competent to chose their
delegates? Certainly there is no evidence for this from the PP LQFB
run, nor indeed from the operation of any existing or historical
republic.

I also do find your faith in the competence in everyman being able to
make decisions over the ability of computers (or other elites) to be
somewhat at odds with the rest of your argument...

Post by Paul Sabin
2) You delegate in specific areas - no blanket delegations.

Making the problem of delegation orders of magnitude more difficult.
And probably even *more* vulnerable to exploitation by demagogues and
corrupt (or at least corruptible) individuals. Domain expertise IMHO
is almost synonymous with biased voting. Who's the world's leading on
expert on abortion? The Pope! Does this mean that we should all be
pressured to delegate to him on the issue of reproductive rights, or
moral issues in general?

Post by Paul Sabin
3) For every specific area, you can choose to participate directly yourself,
delegate, or do nothing at all. You can change this choice at any time.

According to the LQFB PP run, "do nothing at all" wins 6 to 1. All
this proposal does is effectively disenfranchise the vast majority of
the population. I know I asked for someone to defend the elitist
position, but I'm somewhat surprised that promoting oligarchy turned
out to be the major component of the argument...

Post by Paul Sabin
4) You can base your choice of who to delegate to on any criteria you like.
Maybe it's because you think that person is an expert in the field. Maybe
it's because they have a PhD or a lifetime of professional experience. Maybe
not. Maybe you know them personally. Maybe you're confident you share the
same interests. Maybe you've noticed that you tend to agree with their
proposals. It's up to you.

And if I want to be manipulated by a demagogue or someone who will
profit from making decisions for me, or who, by simply being honest
about their own genetic programming, will vote against what mine tells
me to do, by god that is my right!

Post by Paul Sabin
5) Delegations are transitive, so when you delegate to somebody, you're also
trusting their judgement to potentially pass it on to someone else. If/when
this happens, you receive a notification.

We haven't gotten the results of this analysis for LQFB yet, but my
contention is that you're just multiplying the level of incompetence
by allowing this (my prediction is that the voting correlation will
drop from a measly 0.6 down to the truly random 0.5 by the second
jump).

Post by Paul Sabin
6) You can override or revoke delegations at any time. So if there are any
people who have gained a large degree of influence through delegation, they
are always dependent on retaining the trust placed in them by all those who
have chosen to delegate. So there are two major ways that an oligarchy is
prevented from cementing its power at the top: a) division into specific
subject areas, b) potential to revoke delegations.

See my "death by a thousand cuts" argument. Even if people are only
manipulated into making bad choices in twos or threes, it's still a
non-viable system because it doesn't *represent* them.

Post by Paul Sabin
On the one hand, you argue that to admit some people understand a certain
issue better than others is elitist.

Not at all: I fully agree with that statement. I'm only saying that
"voting based on level of understanding an issue" is not even
correlated with "voting like I would if I had the time/motivation to
gain that same level of understanding". Why is that so hard to
understand?

Post by Paul Sabin
On the other hand, you want people to
be represented by others who have "the time/motivation/experience/etc. to
properly investigate the issue". Yet spending time investigating a subject
and gaining experience in a field - this is precisely how you acquire what
most people call expertise. I'm aware that we live in an age where people
have little faith in "experts" - and often for good reason. But that's not
to say that expertise doesn't exist. What we need to ensure is that experts
- or other people with influence - are acting in good faith. We do this by
designing the system so that they are always reliant on maintaining their
base of support.

Again, I'm still having trouble understanding why people are having
trouble understanding the problem here. Is it possible that both a
conservative and a progressive will have the same level of
"understanding" or expertise on an issue? If so, how likely is it that
they'll *still* vote opposite each other? The point is that expertise
isn't the issue, personality is.

Post by Paul Sabin
"people should be represented by direct voters who will vote like they would
if they had the time/motivation/experience/etc. to properly investigate the
issue before voting"
How will we ever know how people would have voted if they *did* have this
motivation, if they *did* spend a lot of time looking into it and if the
*did* have relevant experience?

Because, duh, this can be predicted from their personality profile!

Post by Paul Sabin
The reason to want to delegate your vote -
on a particular subject, not a blanket delegation - is precisely because you
*don't* have that level of interest/time etc. to devote to that particular
field. You cannot even know yourself how you would vote if you did. So your
definition of "representation" is something unknowable.

Only if you completely discount the possibility that ideology/party
affiliation/personality can be used to predict voting habits. My
proposal is has the same general characteristics as one based on
ideology/party affiliation (like all our existing misrepresentative
democracies are), it's just much more precise and much less subject to
corruption and misrepresentation because it eliminates the
power-aggregating properties of those simplistic systems.
Regards,
Scott

Post by Paul Sabin
Regards,
Paul

[Flash Cards]

What you propose is quite close to LD.
But (2) and (5) are pretty random. Both restrict people's decision-making capacity.
To make it truly democratic, (2) should be removed and (5) made optional.

Regards


From: Paul Sabin <***@gmail.com>
To: Metagovernment Project <***@metagovernment.org>
Sent: Friday, October 14, 2016 5:04 AM
Subject: Re: [MG] Axiom 6: People are not competent to choose their proxies

Hi all,

Let me present an alternative view of how representation/delegation should work. This is my own personal take on the matter, but I believe it has a lot in common with the liquid feedback approach.

1) You get to choose who to delegate to. It's not decided by any kind of system or algorithm by default. Ideally, you'll feel competent enough to participate directly. If not, you can delegate. You and only you are competent enough to choose a delegate. Our decision-making capabilities are far from perfect, but they're far better than any human-invented algorithm.

2) You delegate in specific areas - no blanket delegations.

3) For every specific area, you can choose to participate directly yourself, delegate, or do nothing at all. You can change this choice at any time.

4) You can base your choice of who to delegate to on any criteria you like. Maybe it's because you think that person is an expert in the field. Maybe it's because they have a PhD or a lifetime of professional experience. Maybe not. Maybe you know them personally. Maybe you're confident you share the same interests. Maybe you've noticed that you tend to agree with their proposals. It's up to you.

5) Delegations are transitive, so when you delegate to somebody, you're also trusting their judgement to potentially pass it on to someone else. If/when this happens, you receive a notification.

6) You can override or revoke delegations at any time. So if there are any people who have gained a large degree of influence through delegation, they are always dependent on retaining the trust placed in them by all those who have chosen to delegate. So there are two major ways that an oligarchy is prevented from cementing its power at the top: a) division into specific subject areas, b) potential to revoke delegations.

7) Delegations are open. Everybody can see who's delegating to whom. In the case of transitive delegations, you can trace the path of delegations as they are passed from one person to the next.

Now let me make two quick points about the argument as you've laid it out above, Scott, because I don't believe the distinction you've drawn between "representationist" and "elitist" holds any water.

On the one hand, you argue that to admit some people understand a certain issue better than others is elitist. On the other hand, you want people to be represented by others who have "the time/motivation/experience/etc . to properly investigate the issue". Yet spending time investigating a subject and gaining experience in a field - this is precisely how you acquire what most people call expertise. I'm aware that we live in an age where people have little faith in "experts" - and often for good reason. But that's not to say that expertise doesn't exist. What we need to ensure is that experts - or other people with influence - are acting in good faith. We do this by designing the system so that they are always reliant on maintaining their base of support.

Another fundamental problem with this same point:

"people should be represented by direct voters who will vote like they would if they had the time/motivation/experience/etc . to properly investigate the issue before voting"

How will we ever know how people would have voted if they *did* have this motivation, if they *did* spend a lot of time looking into it and if the *did* have relevant experience? The reason to want to delegate your vote - on a particular subject, not a blanket delegation - is precisely because you *don't* have that level of interest/time etc. to devote to that particular field. You cannot even know yourself how you would vote if you did. So your definition of "representation" is something unknowable.

Regards,
Paul

On 13 October 2016 at 22:49, Patrick Millerd <***@gmail.com> wrote:

I might consider implementing a warning system too against anyone trying to game the system by misrepresenting themselves in the personality test. If a certain number of users switch their vote from the one the proxy chooses, the system alerts the other auto-matched users that this proxy might not be representing them.    
On 12 October 2016 at 21:34, Scott Raney <***@gmail.com> wrote:

One of the things Kling said has kind of stuck in my craw:

"We discussed this and came to the conclusion that there are
other factors than representativeness which play a role when delegating
a vote, i.e. sympathy and popularity."

I'm sensing basically the same view from Carlo and Sergey, although
they might claim other factors, such as "expertise", as being
important. I'm going to call my Axiom 5, which states that people
should be represented by direct voters who will vote like they would
if they had the time/motivation/experience/etc . to properly
investigate the issue before voting, the "representationist" position.
I'll call the opposite position "elitism" because it implies that
there is something special about those people you propose to want to
delegate to (if you have an alternative proposal for what to call it,
by all means make it).

I frequently point to demagogues as an example of a fatal flaw in the
elitist position, but what exactly is the problem with demagogues? Is
it just that they're racist, nationalistic, and authoritarian? Not at
all! If that were the real problem, then what we should be doing is
designing social engineering countermeasures to those things, which we
could then apply not only to the demagogues but to anyone else who
might vote that way, regardless of whether they were being led by a
demagogue or not.

No, the problem with demogogues is their ability to *amplify* their
position on any and all issues by convincing others, who may or not
agree with the demagogue on any *specific* issue, to effectively grant
a blanket proxy to the demagogue. But what about using "expertise" or
"popularity" or "familiarity" or some other criteria to select a
proxy? That should be OK, right? No, because it has exactly the same
fatal flaw: By granting a proxy to someone based on any single
characteristic and exclusive of knowing their actual personality type,
it is virtually guaranteed that the proxy will distort the voting on a
wide range of issues because of their ability to amplify their votes,
which will over-represent their personality type in the final decision
and which is essentially taking advantage of those who have granted
them a proxy unless the elite just coincidentally happens to cast a
true representationist vote.

All of course is *in addition to* the other problems I've pointed out
with a system that allows granting proxies to specific individuals:
1) The possibility of dead-ending and no vote being cast.
2) The possibility of corruption because the elite then becomes a
target for manipulation.
3) The possibility of corruption due to the elite casting a self-serving vote.

At this point I have to recommend that we equate the elitist position
with those who think they need to have the ability to disable
stability control on their cars, a position that even the experts in
that field consider to be just plain stupid, even on a race track, let
alone out in traffic:
http://www.roadandtrack.com/ca r-culture/a28714/turning-off- stability-control-is-one-of- the-dumbest-things-you-can-do- in-a-car/

Bottom line, I'm walking back on my proposal to even *calculate* a
vote from those one follows, let alone allowing automatically casting
that vote. This still allows people to make the calculation manually
by actually looking at the votes of those they're following and then
manually voting, but requiring even this extra effort should
discourage almost everyone from doing so, which of course is in their
best interest if what we're trying to build is a representationist
system. I still think it's a good idea to allow blocking of specific
individuals from the AutoMatch, but my current thinking is to have
that be a separate system from the "follow" feature.

Anyone care to try to defend the elitist position?
  Regards,
    Scott

______________________________ _________________
Start : a mailing list of the Metagovernment project
http://www.metagovernment.org/
Post to the list: ***@metagovernment.org
Manage subscription: http://metagovernment.org/mail man/listinfo/start_metagovernm ent.org



______________________________ _________________
Start : a mailing list of the Metagovernment project
http://www.metagovernment.org/
Post to the list: ***@metagovernment.org
Manage subscription: http://metagovernment.org/ mailman/listinfo/start_ metagovernment.org

[Flash Cards]

People are not competent to choose their proxies
This still allows people to make the calculation manually
by actually looking at the votes of those they're following and then
manually voting, but requiring even this extra effort should
discourage almost everyone from doing so, which of course is in their
best interest

Maybe you indeed should try to design a benevolent dictatorship system, which is not difficult, according to you. I think you might be psychologically predisposed for it.

Regards

From: Scott Raney <***@gmail.com>
To: Metagovernment Project <***@metagovernment.org>
Sent: Wednesday, October 12, 2016 6:34 PM
Subject: [MG] Axiom 6: People are not competent to choose their proxies

One of the things Kling said has kind of stuck in my craw:

"We discussed this and came to the conclusion that there are
other factors than representativeness which play a role when delegating
a vote, i.e. sympathy and popularity."

I'm sensing basically the same view from Carlo and Sergey, although
they might claim other factors, such as "expertise", as being
important. I'm going to call my Axiom 5, which states that people
should be represented by direct voters who will vote like they would
if they had the time/motivation/experience/etc. to properly
investigate the issue before voting, the "representationist" position.
I'll call the opposite position "elitism" because it implies that
there is something special about those people you propose to want to
delegate to (if you have an alternative proposal for what to call it,
by all means make it).

I frequently point to demagogues as an example of a fatal flaw in the
elitist position, but what exactly is the problem with demagogues? Is
it just that they're racist, nationalistic, and authoritarian? Not at
all! If that were the real problem, then what we should be doing is
designing social engineering countermeasures to those things, which we
could then apply not only to the demagogues but to anyone else who
might vote that way, regardless of whether they were being led by a
demagogue or not.

No, the problem with demogogues is their ability to *amplify* their
position on any and all issues by convincing others, who may or not
agree with the demagogue on any *specific* issue, to effectively grant
a blanket proxy to the demagogue. But what about using "expertise" or
"popularity" or "familiarity" or some other criteria to select a
proxy? That should be OK, right? No, because it has exactly the same
fatal flaw: By granting a proxy to someone based on any single
characteristic and exclusive of knowing their actual personality type,
it is virtually guaranteed that the proxy will distort the voting on a
wide range of issues because of their ability to amplify their votes,
which will over-represent their personality type in the final decision
and which is essentially taking advantage of those who have granted
them a proxy unless the elite just coincidentally happens to cast a
true representationist vote.

All of course is *in addition to* the other problems I've pointed out
with a system that allows granting proxies to specific individuals:
1) The possibility of dead-ending and no vote being cast.
2) The possibility of corruption because the elite then becomes a
target for manipulation.
3) The possibility of corruption due to the elite casting a self-serving vote.

At this point I have to recommend that we equate the elitist position
with those who think they need to have the ability to disable
stability control on their cars, a position that even the experts in
that field consider to be just plain stupid, even on a race track, let
alone out in traffic:
http://www.roadandtrack.com/car-culture/a28714/turning-off-stability-control-is-one-of-the-dumbest-things-you-can-do-in-a-car/

Bottom line, I'm walking back on my proposal to even *calculate* a
vote from those one follows, let alone allowing automatically casting
that vote. This still allows people to make the calculation manually
by actually looking at the votes of those they're following and then
manually voting, but requiring even this extra effort should
discourage almost everyone from doing so, which of course is in their
best interest if what we're trying to build is a representationist
system. I still think it's a good idea to allow blocking of specific
individuals from the AutoMatch, but my current thinking is to have
that be a separate system from the "follow" feature.

Anyone care to try to defend the elitist position?
  Regards,
    Scott

[Scott Raney]

Post by Flash Cards

People are not competent to choose their proxies
This still allows people to make the calculation manually
by actually looking at the votes of those they're following and then
manually voting, but requiring even this extra effort should
discourage almost everyone from doing so, which of course is in their
best interest

Maybe you indeed should try to design a benevolent dictatorship system,
which is not difficult, according to you. I think you might be
psychologically predisposed for it.

Snark aside, I think I've come as close to this as we're able to at
this point in the development of our technology. The trick is to limit
the power of the "dictator" to executing the laws, rather than making
them. Eventually some AI will be capable of actually making the laws
based on its ability to figure out what makes us tick, but IMHO that
is is far enough off (decades, maybe centuries) that it's too risky to
our species to plan to wait that long before we start using something
better.

Mean time, you're right that it doesn't bother me at all to restrict
your freedom to do dumb things. For example I'd vote to remove the
"disable stability control" button from *all* cars because that's
exactly the kind of thing that people who value their "freedom" think
they need and are entitled to have, nevermind that they're wrong both
on a moral and a technical level. And when self-driving cars start
becoming common I'll be voting to subsidize them to accelerate the
adoption of them and then when they're universally available, vote to
criminalize the practice of driving your own car on the public roads
entirely. Your "freedom" doesn't extend to the right to put other
people's lives at risk, especially when all it gets you is a sop to
your delusion that the Dunning-Kruger effect doesn't apply to you.
Your insistence on the "freedom" to pick your own delegates/proxies I
consider to be a textbook example of that.
Regards,
Scott

Post by Flash Cards
Regards