Thursday, November 19, 2020

Student debt forgiveness. The interesting argument for why it "feels right"

Recently, many are talking about forgiving student debt. 

Countering are policy analysts and economists raising various policy and fairness arguments against it (see Justin Wolfers at Freakonomics, among many others). 

So, is the case for student debt forgiveness senseless?

I will argue that it has a deep logic to it. Even if one does not ultimately agree to the proposed policy. 

The education system, argues Bryan Caplan, is mostly about signaling. 

You prove to employers that you are smart, hard-working, and docile.

To get those papers, a student spends years and a fortune. And might leave with $100,000 in debt.

Going to the job market, one isn't intuitively seeing that his education has neither helped his current job - it mostly does not at all. Not is it visible how it improves one's job prospects. 

That CVs of those without a degree are not looked at is true and known. But this is not that visible to be felt and experienced as worthwhile enough. 


The full feeling of unfairness is a summary of the multiple factors:
1. The education itself is very wasteful. Not adding much value. 
2. The cost is extreme. And it is made expensive by various idiotic and irrational aspects of "the education system"
3. The effects on employment are random. Not all are benefitting. Especially the drop outs. 
4. Most effects are not visible enough to feel justified.  


Thus, the policy analysts might be right. But those two groups are talking at cross purposes. 

Students straddled with debt, feeling to have been dealt an utter injustice, have a point to feel that forgiving this debt is justice done
. 
The policy analysts are doing a completely different analysis. About fairness in total in society, and about the consequences of forgiveness, economically and for future students.

I hope I helped both camps understand each other slightly better.

Saturday, October 10, 2020

The idea market deficiencies.

1. Mixing up certifications (degrees that people get to help getting a job) and research / education. One might even split research and education. But certification/signaling is the big problem for me. 


2. There is no market nor collaboration system for small ideas, or partial ideas. 


As an academic, you either publish a paper or not. So small but very good ideas of good value do not have a good vehicle. Both to get accumulated, found by others and recombined. And to get credit / value from (incentive)


3. No incentive and no funding for infrastructure / quality / connections of data. 

The lack of linking between papers to their refutations is partly because nobody has done it, because no institution invests in infrastructure. 


This complaint is also that showing that a paper is flawed isn't getting rewarded, or not even getting published!. (@jamesheathers

 and @sTeamTraen have elaborated on this a few times)


Skeptics finding faults in the literature should be richly rewarded. 


But no one pays for those things for whatever reason.... 

PS. Thought triggered by this MarginalRevolution post 

Tuesday, July 21, 2020

Why can't we optimize our life like iPhone is optimized?

The intuitive wonderment of "why can't we optimize life" can be demonstrated by comparing own life to an iPhone.

Look at the unbelievable optimization seen in your recent smartphone. It is optimization upon optimization. All laid to utter perception almost.

Your life? maybe you know whether 1 vs 2 spoons of sugar is best (fun fact. nobody does 1.3 spoons, even though this will be taste optimal for some certainly!)

The answers to this wonder are a good start for considering the difficulties of life optimization.

1) its machinery. Humans a complicated, unstable, and hard to pin down. There is no "0.46% of boron-oxide" type consistency and simplicity.

2) Economics of scale.
You can invest $10 billion in the machinery of producing CPUs. But this one system will produce millions of units.

In humans, each one might need its own system. No economics of scale.

3) The agency problem.

In order for someone to create an optimization for you, he needs to earn money from it.
In phones its simple. Apple creates them, and gets paid if they are good enough.

Is anyone going to get paid $10,000 to make you healthier?

Of course, one can make pennies by selling a book, or ads on her diet site.
But the money earned is minuscule in relation to to value created.

This means that nobody is going to invest $2,000 to reliably make you healthier.
Even if the value of making you healthier is $100,000, nobody is going to get paid serious money to make it happen.

Apple makes billions to make phones.
Get you to be healthy?

maybe one can get some google ads cents. 

Tuesday, June 30, 2020

Optimization and its dicontents

Optimization here means colloquially.

"I want to optimize my life" etc.

Why is it so difficult? (either to optimize life or any other real life targets)

First, lets note the implied definition of this elusive yet tempting optimization

Its is always beyond the obvious. Nobody goes "I have not slept, let me optimize my sleep and go to bed". This is technically an optimization. But it is too obvious, and a given to be included.

We will normally optimize our lives, cooking, etc without resorting to explicit optimization talk.

Optimization talks about the less obvious, usually ignored, yet still plausibly achievable plans.

Thus, optimization will normally be somewhat marginal. Not marginal really, but veering towards the margin in terms of added value.


Difficulty and not being obvious is another feature of so-called optimization. The reason is as above. Adding more salt to a dish might be marginal in value, but is so easy to do that it sin;t included.


Those definitions in themselves explain some mysteries about optimization. Why it is not always done. They also lay the ground to deeper study optimization, how it might work, and why so often it either is not initiated, and why it sometimes actually fail. 

Tuesday, June 12, 2018

Stanford prison study - how moralizing destroys science

A recent inquiry argues that the famous Stanford prison study was faulty and scientifically invalid.

It in an example of how missing moralizing in science destroys scientific inquiry.

The prison study is unique. It created a huge backlash on ethics etc.

The unintended consequence of the ethics distraction was two fold:
1) deter from any kind of replication
2) create the impression that the science is solid, if not for the annoying / out of context ethics "issues"
3) there was an implied "mission" to reform prisons. Which - again - messed up proper scientific inquiry.

Great example on how moralizing leads to bad scientific progress.

PS.  I'm not in position to decide if the study was indeed faulty. I honestly don't know. I would need very serious study of the details and the various parties version of events to have a proper opinion!

Monday, June 4, 2018

#MeToo: the skewed distribution of repeated offenders.

An old Jewish tale.

Two merchants have a dispute.
They go the the rabbi to arbitrate.
Come merchant 1 and details his side of events, the rabbi listens and nods: "You are right"
His opposite side come in with his perspective, and the rabbi goes "you are right"

The rabbi's wife is sitting there listening to the whole thing and she can't stand it:
What idiocy is this? he is right, and he is right too! this makes no sense.

The rabbi - calmly - "You are absolutely right"

This joke is quite descriptive of the #MeToo controversy.

Can both be right?
The women listing their various grievances and experiences AND the men saying that this is wierd extremism?

Yes, both are right.

How could that be?

We have many women that report on copious amounts of harassments / rape etc.
We have numerous men that listen to it dubiously. "This can't be true" they say with utter conviction.

The solution is statistical and very simple.

Repeated offenders.

Let's assume that we have 10 out of 100 men that are prolific sexual offenders.
He will harass a lot, rape if an opportunity arises, and generally engage in various unsavory behaviors towards women.

The other 90, are the average Joe. They might wink at a woman at the coffee shop, or even try to talk to her if they think she is looking back. They sometimes go to dates, and occasionally things aren't super clear. So they might on rate occasion displease a woman, but usually without malice of ill intent.


The women will suffer a lot. Few harassers are enough to terrorise many women.
A noxious men can inconvenience 20 women a day.
Average Joe will start talking to a woman once in awhile, but not in industrial numbers.
The average Joe looking at a woman to see if she is looking back, might not even be noticed by her. And if he is polite enough, she might not feel discomfort.
But the calouss one will be well remembered.

The number of times a woman is inappropriately approached might be very high. Even if most men will not harass.


Why would the average men believe the #MeToo stories? If he and his close friends never behave this way, how could he believe the stories?

Here comes the mix up of harassment and awkward encounters.

The average Joe has had his numbers of misunderstandings with women. Not malicious. usually not damaging either. But discomfiting nevertheless.

Average Joe will assume that #MeToo is about those subtle misunderstanding. And gesture wrongly comprehended. The mutually drunk encounter. The bad date. You name it.

Extreme feminists will help in this error.
They will drag real women suffering into the gray dating mishaps. They will say
     "YES! you see this traumatised woman? She suffered from people like you that have not filled in the 200 page consent form before touching the lady finger"
     "It is all insensitive gross misogynists like you all. Yes all men and you are not exempt"

Causing derision and disbelief. Making rape victims harder to believe. And putting off men from listening.

But the suffering is common and real. It just comes from a noxious minority.

PS. The implications for policy and public discourse are many. The reader is invited to think it out.


Reference:
   "Mate" by Geoffrey Miller and Tucker Max discusses the asymmetry of sexual misconduct. Where a minority of repeated offenders cause innumerable numbers of offences.





Monday, May 28, 2018

Criticism threshold: Middling criticism masquerading as devastating takedown

Levels of arguments have their own meaning and implications.

Saying "argument 7b in this book is doubtful" differs from "this guy is a murderer"

But those get mixed up.

Jordan Peterson awakens this point yet again.

The thrust of anti Peterson views is prohibitive for those hating him..
his derision of Social Justice Warriors. His refusal to kowtow the PC rules and many other sins.

But those arguments fail for people outside those ideological lines.
And attempts to find more serious sins in his texts and videos seems to have failed.

Now we got a new genre of anti Peterson articles.
The middling critics.

Here is something he said that is wrong.
This aspect of his personality is - ehh - not to my taste.
His moods do not look good to me.

Which are all find and dandy points to make in reviewing public personas.

But, deception.

Those middling points are masquerading as being devastating arguments.
because someone does not like various things peterson is dangerous.

It is a subtle form of deception. But deceptive nonetheless.