Review of “Big Boys Don’t Cry”: Edifying

From Amazon:

BIG BOYS DON’T CRY is a novella by LTC Tom Kratman, U.S. Army (Ret.), known for A DESERT CALLED PEACE and his Carrera series. The story follows the life cycle of a Ratha, a sentient future supertank that dutifully fights Man’s battles on dozens of alien worlds. But will the massive creature still be grateful to its creators when it discovers it has a conscience? And how long will an intelligent war machine with enough firepower to flatten a city be content to remain Man’s obedient slave? DRM-free.

Kratman’s schtick is to speak hard truths about military reality coated with a thin veneer of science fiction to help it go down easy. This novella isn’t about AI-driven tanks, it’s about training soldiers. Specifically, it’s a paean to the soldiers who’ve been left behind in VA hospitals scattered across America. Spoilers follow.

Kratman begins with a mechanical description of the perfect soldier’s psychology as expressed by a futuristic machine intelligence, mixing hot passions with cold, calculating brutality. As they said in Full Metal Jacket, the military doesn’t want to train people into robots because robots are stupid. They lack initiative and judgment. The military wants to train people into killers. There is and will always be high demand for killers who can think their way around a problem and empathize with their allies and enemies.

The second, larger part of the novella follows the sentient tank’s thought processes while it’s laid up in traction, after being irreparably damaged. Analogous to the countless veterans being abused every day by low-cost immigrant workers and an apolitically heartless VA apparatus, the AI begins to wonder Why, why did I do it? Why was I ever loyal to these people. It traces these thoughts back to its “psychological” training by a behavioral scientist, who had put the tank’s mind through a series of ancient military simulations. Using applications of overwhelming pain and pleasure, the tank learned which attitudes and behaviors were “good” and which were “bad”. Ever so coincidentally, the “good” behaviors were those which developed the military virtues of bravery, intelligence, savagery, initiative, loyalty, and so on.

This is really the key point about training, because the only thing you absolutely must train is a soldier’s psychology. People (and tanks) who have the necessary attitude to get a job done will go out of their way to find the tools they need and develop the skills they need to get it done.

The remainder of the novella deals with the real cost of abandoning your own troops. Leaving all empathy aside, there’s a cold, hard pragmatism in the motto “No man left behind”. The fact is, the only difference between the Americans and the Iraqis in the Gulf War was a culture of trust and loyalty. The Iraqis knew that they would be left behind, so they didn’t even bother fighting (why risk it?). The Americans knew they’d get the best medical care in the world. Economies are also built on trust and loyalty, which meant the Americans had the best training, systems, and equipment, because so many people back home could go to work and do their jobs without fear of predation. The other side is that the real cost of abusing veterans is opportunity cost: somewhere in the area 100 other soldiers and potential recruits who see the abuse and hear about it, and make different decisions in self interest than they would have.

The ideas in the previous paragraph are very abstract, like an argument you’d present to an actuary, which doesn’t work in fiction. So Kratman ends the book with the tank slaughtering its users and abusers in a last gasp of righteous fury. You might be surprised how much you empathize with the tank’s perspective by that point.

Five stars.

Posted in Uncategorized | 1 Comment


The transition into the new job is going well so far. It helps to have the systems I’ve built for thinking about such things. Particularly, I’ve made excellent use of free energy from transitions. I’ll possibly get around to describing this pedestrian stuff in more detail later. However, a weird thing happened that feels more important.

Yesterday was my first day at my new job. That is probably important in terms of emotional timing. My dreams last night were about traveling, which is almost always an indicator of letting go of something, emotionally. This morning, I woke up thinking about Aeris from Final Fantasy 7. The music associated with the scene where Aeris is buried was stuck in my head (along with a couple of other important Aeris scenes and associated music) and the feeling was a mix of nostalgic, bittersweet, and lighthearted. It was like I was letting go of something I hadn’t even realized was bothering me.

Now, it’s important to note that before now, I never understood the fascination with this character. Maybe it’s because I was 12 at the time and puberty was a ways off yet, but I felt no attachment to Aeris whatsoever. She wasn’t very strong and I never put her in my party, so when she died I wasn’t really upset. I mean, the permadeath of a playable character is always a little jarring for a 12-year-old, but I was more just pissed that Cloud buried her before I could unequip her Lightning 3 materia. This morning, that changed. I finally understood how people get so attached to that character, and I felt that little tightness in my intestines that I associate with loss and loneliness.

The way I figure is this: I was building a new neurological association with an emotional category related to either pair-bonding or loneliness or both. I think this is my brain reacting to the shift from resource dependence (on parents) to resource independence. This would be because my mind is shifting from resource scarcity to…not quite resource abundance so much as the knowledge that having sex is okay now because having a baby wouldn’t be a complete disaster.

Tangent: I may not be the trooest thal in existence, but I’m definitely not wired to be a loser AND a parent at the same time. Parasitism might be where I’m at, realistically speaking, but I can’t say it works for me. On the other hand, the preferred prole method of reproduction appears to be to have kids by accident, then either take off (black males) or just never interact with the kids because they’re desperately working 80 hours per week (Mexicans, low-class whites). I expect this is why NW European altruism doesn’t allow us to have three or four generations living in the same home.

Posted in Uncategorized | 16 Comments

Naive success prediction index

Here are the factors that appear to have the strongest impacts on whether a person can get the things they desire. (NB: Not the same as happiness, which is getting what you need.)

  • Health
    • Lack of chronic illness
    • Genetics
    • Nutrition, sleep and exercise
    • Religiosity
  • Intelligence
    • General factor (IQ)
    • Liberal education (good cognitive tools for navigating problems)
  • Adaptive attitude
    • Robust optimism
    • Pragmatism
    • Lack of fear of embarrassment
  • Ambition (internal motivation)
    • Strong ego
    • Psychoticism/associative horizon/creative energy
    • High biological energy
  • Conscientiousness
    • People focus (external motivation)
    • Task focus
  • Opportunity
    • Starting position in adulthood (how much X you have)
    • Social mobility trends (index of opportunities to use X to get Y)
  • Practical skills (taken from Scott Adams, assumes “good but not great” skill level)
    • Public speaking
    • Psychology
    • Business writing
    • Accounting
    • Design (the basics)
    • Conversation
    • Overcoming shyness
    • Second language
    • Golf
    • Proper grammar
    • Persuasion
    • Technology (hobby level)
    • Proper voice technique

Here are the top-level ones you can’t change very much.

  • Intelligence
  • Ambition (internal motivation)
  • Opportunity

Here are the ones you can change a medium amount.

  • Health
  • Adaptive attitude
  • Conscientiousness

Here’s the one you can change a lot.

  • Practical skills

Rate each top-level trait from 1-5:

1: Much worse than normal
2: Lower than normal but could be worse
3: Normal
4: Higher than normal but could be better
5: Much higher than normal, with diminishing returns

Plug into this unweighted formula:

Success prediction index = Product of trait ratings (highest possible score is 5^7 = 78125)

For example, I’d rate my 18-year-old self like so:

  • Intelligence: 4
  • Ambition (internal motivation): 3
  • Opportunity: 5
  • Health: 5
  • Adaptive attitude: 3
  • Conscientiousness: 4
  • Practical skills: 1

Plugging this in I get: Success prediction index = 4*3*5*5*3*4*1 = 3600

Rating myself now, at 29, I’d get:

  • Intelligence: 5
  • Ambition (internal motivation): 4
  • Opportunity: 2
  • Health: 4
  • Adaptive attitude: 2
  • Conscientiousness: 2
  • Practical skills: 3

The new index is 1920, indicating that my future is significantly less promising than when I graduated high school, but not hopeless. However, I should readjust my expectations for how many of my desires I can actually fulfill.

Posted in Uncategorized | 22 Comments

Mental energy = bioavailable glucose

This isn’t anything new (probably), so I’m just making the connection for myself. If bioavailable glucose is what we colloquially refer to as “mental energy” then this can explain a lot.

-Why smarter people have more PET activity while watching TV (bigger supply leading to higher baseline expenditure, like with money)
-Why the neural efficiency hypothesis fails to predict the very high mental activity in males who ace the math SAT, but not in females (high engagement in males)


-Why caffeine cures ADHD (temporarily raises average engagement level, where ADHD is analogous to financial conservatism)

If we take all of these anecdotal results as representative of factors, then we should predict that mental energy expenditure on a novel task will have five dependent variables: total bioavailable glucose remaining, conservatism (personality tending to conserve energy), interest level (strength of desire to complete the task), fluid intelligence (novel problems solved per energy spent or just “efficiency”), and difficulty of the task. Here’s a naive math statement of this,

Change in Energy (negative) ~= Interest * Thrift(Total energy, conservatism) * (g-loading/fluid IQ)

A couple of further thoughts on this.

-I expect that Interest is a function dependent on people-focus versus object-focus. For example, females (who tend to prefer people-oriented jobs) would be expected to have more interest in a math lecture than a math book, and vice versa for males (and particularly those on the autism spectrum). This appears to be largely related to testosterone and introversion.
–This may also explain Paul Cooijmans’ observation that high-ability male spergs often suffer from impaired visuospatial abilities, whereas high-ability female spergs don’t. Speaking anecdotally (from introspection), visuospatial tasks require relaxation and unfocus, rather than high focus and effort.
-The thrift function likely takes the same form as personal expenditures. If anyone knows an empirically based microeconomic formula for that, please let me know.
-The typical skill curve in an activity appears (to me) to be an exponential decay curve approaching a person’s highest possible skill level. This probably means that “mastery” is just the absence of novel problems, which is to say crystallized intelligence. This could be measured objectively in units of candela (photons emitted in PET), so that crystallized intelligence would be inversely proportional to energy spent while playing Tetris at a mastery level of skill with diminishing returns in improvement.

The important thing about the latter point is that it’s a ratio scale with physical units and a true zero, with zero indicating the highest possible crystallized intelligence.

Also anecdotally, I’ve noticed that increasing oxygen availability (with cardio exercise) increases readily available mental energy (because of the ease of burning glucose) and that increasing blood flow to the brain (during cardio or hot showers) causes mental energy to be spent more indiscriminately.

Posted in Uncategorized | 5 Comments

The polymorphic preselection rule (of thumb?) for positive manifolds

That’s “the PP rule”, in order to give credit. This is a speculative inference from an observation by Pumpkin Person.

I’m going to generalize this as follows: A polymorphic trait that contributes to a more general capability, but is not specifically useful for that task, will be most competitive at +1 SD. In effect, the most successful white American men in any enterprise requiring genetic fitness but not height, particularly, will form a bell curve around 6’2″ (+1 SD from the white American male average).

Insofar as a task selects for more than +1 SD of a trait in the highest performers, we should expect that the trait is particularly adaptive for the task. It stands to reason that the best height for a man is 6’2″, and the best IQ is 115, unless there’s selection bias to the contrary. Furthermore, we expect that the smartest or most successful white American men will form a bell curve around 6’2″, and vice versa. Taking a quick look at the heights of US presidents, this pattern matches. Recent presidents cluster around 6’2″ (+1 SD) whereas a century ago they clustered around 5’11” (+1 SD at the time).

This may not be correct, but it seems pretty good offhand.

I’m not sure why this would be the case, but here’s an idea: a polymorphic trait is like a statistical sample of the overall genome. Why +1 SD, I have no idea, it’s just a number I picked because 6’2″ is the most sexually attractive male height, as I understand it, and 115 IQ is the beginning of the IQ “sweet spot” between 115 and 130. If this explanation of statistical sampling is correct, then it should hold for environmental variables as well as genetic, in predicting the average phenotypes that are the most successful at some task.

A fun little implication of this is to note that if the college application process produces an average university student IQ of 115, it could be replaced by a system that selects students based on height without losing any validity. Because it is increasingly the case that the average student IQ is approaching 100, we must conclude there is no selection effort whatsoever because it isn’t operating better than a flipped coin.

Posted in Uncategorized | 1 Comment

How I categorize performance characteristics of acquaintances

I’m going to model the behavior of people from sort of an operations engineering angle, using Paul Cooijmans’ trifold breakdown of problem-solving traits: intelligence, conscientiousness, and associative horizon.

High IQ: Capable of completing difficult tasks, and completes easy tasks more quickly.
High conscientiousness: Does what they’re told, spends no time on other things.
High AH: Does what they want. Spends some time on what they’re told, but also time on what they aren’t told. (E.g. Google’s 20% time policy acknowledges this trait.)

All three traits increase the quality of a finished product or project, but that’s basically impossible to measure or plug into a Gantt chart so I won’t go into that yet.

Figure every human process has a threshold IQ, after which the time required is divided by IQ. So some pseudocode for modeling this would be:

if employee.getIQ < difficulty:
print("This employee was not smart enough to complete the task you assigned.")
print("The employee completed the task in " + time_spent(difficulty, employee.getIQ) + " hours.")

I figure that calculating the time_spent function ought to be done the way mechanics are paid: by measuring the time spent by average performers and then estimating based on the folksy observation that two standard deviations in IQ corresponds to performing mental tasks problems about five times faster. Tasks would have to be given a cerebral index for how g-loaded they are, versus how mechanical (writing a report might have an cerebral index of 3 and changing oil might have an index of 1.2).

The interaction between conscientiousness and associative horizon is like time-sharing in a processor. Think about it this way: each person has an external list of prioritized tasks given by authority figures, and an internal list of prioritized tasks representing personal interest. Time is apportioned in order of which tasks are personally weighted as the most important, and each task receives an interval of time representing a person's level of focus. Conscientiousness refers to A) focus (longer time intervals) and B) assigning greater weight to pleasing others. Associative horizon refers to assigning greater weight to gratifying one's own interests. On a macro scale it's better if everyone has high associative horizon because more of the things people care about actually get done (maximizes global problem-solving and happiness). On a micro scale, it's better if everyone has high conscientiousness so that long projects and deferred gratification are possible.

When I'm analyzing acquaintances in broad strokes, I like to use the qualitative categories "low", "normal", and "high". I use "normal" in the vernacular sense meaning "without considerable defect but not special either", so a "normal" IQ would be something like 115, and "low" would be 100 and below. A person of normal intelligence can understand and follow simple instructions, but not much else. A person of normal conscientiousness can adhere to a list of tasks and a calendar most of the time, almost always shows up to work appointments on time, and shows up to personal appointments a little bit late most of the time. A person of normal associative horizon has a reasonably objective sense of himself and the external world and a small deficit in common sense.

Hopefully that gets the gist across, it's mostly not formalized at this point.

Posted in Uncategorized | 2 Comments

Estimating neanderthal and EMH visuospatial IQ

First, let’s presume I’m correct and the correlation between visuospatial IQ and brain size is approximately linear in humans, and that late neanderthals and early modern humans were similar enough to modern humans to be comparable. White people have an average brain size of 1350 cc and a standard deviation of approximately 130 cc, and a visuospatial IQ of 100 by definition. Neanderthals and EMH had an average brain size of 1600 cc. This is +1.9 SD above the modern human mean, predicting a nonverbal IQ of 128.5.

If we correct for encephalization by multiplying by the ratio of neanderthal EQ to modern human EQ (4.8/5.4 = 0.889), then 1600 cc times 0.889 is 1422 cc. This is +0.55 SD above the white mean, predicting an average visuospatial IQ of 108.3.

The same analysis for Amud 1 (1740 cc) yields an expected nonverbal IQ of +3 SD. If I understand correctly, you have to multiply that by the strength of the correlation when estimating individual IQs. This would be a conservative estimate because Amud was an ectomorphic neanderthal.

In addition, Pumpkin Person reports (reluctantly?):

Based on the fact that they left behind no drawings, I now estimate Neanderthals, like Homo erectus, had an artistic IQ of about 26. However in the documentary Apocalypse Neanderthal, a scientist mentions that it took him a year and a half to learn to make the stone tools Neanderthals made all the time. Since scientists probably average about 125 IQ, that might suggest Neanderthals had a spatial IQ as high as 125!

Neanderthal IQ

If something like the Levallois technique is going to survive tens of thousands of years by cultural exchange, the average individual in the culture has to be able to learn it.

Let me know if I’m fucking anything up here.

Posted in Uncategorized | 7 Comments