The .plan for the Machine

Follow

The supreme way to follow

Master log, Stardate 20588.6

Old is Gold: Where Old Tech Curb Stomps Everything in the Last Decade (Part 1)

Tags: Technology

When I was a teenager, my hobby was doing game dev in online forum groups. You know, the groups of people who are always posting online about how they want to do this or that game, but somehow never end up going anywhere. Except, I did want to go somewhere. So I worked in a bunch of different of groups trying to find the few groups that actually went anywhere.

In one of the most rare and exciting projects, the lead designer had access to some of the latest game development technology and they actually knew something about how real world game development worked (which at the time was about as rare as finding a bear brandishing a machete). So of course, we on the team were ridiculously excited about this project. I think just about everyone commented something to the effect of "Finally, a chance to make something like the big studios!"

To make this ambition a reality though, we needed to be sure we were using the technology as fully as possible. Otherwise, we'd just be making another low quality game with high quality technology. So with a lot of guidance from our experienced team lead, we came up with a nice process to ensure we were using the new tech to it's full potential.

In total, there were 9 steps in the process.^[1] It usually took ~20 hours of effort to go from nothing to something in the game if it was a static object. Animated objects usually took closer to 40 hours.

Pop Quiz: Given our rate of art production, the fact that almost no one was putting more than part time effort into this project, and the fact that it's hard to get more than one or two sufficiently talented artists to join a hobby project at once, how much art could we realistically produce in a year?

Solution: Well...

Bottom line, one year later there were maybe 10 objects in the game that were of the quality we originally envisioned. There were many more unfinished objects, none of which could be used because they would clash with the "nice" objects if they were used.

Frustrations were high, morale was low, and the whole team gradually became more interested in more productive projects. Ultimately, the project dissolved and nothing ever came of it, just like most other forum hobby projects.

What went wrong? In this case, the standard of quality created an all-or-nothing problem: Either an idea could be realized to a certain visual quality or the idea couldn't exist at all, regardless of it's other merits. Our game couldn't be realized to our ideal visual quality, so our game couldn't exist at all, regardless of it's other merits.

Moral of the story: Production value has a high hidden cost. Namely, anything that cannot be created at the production standard given the time, budget, and manpower constraints simply cannot be created at all, no matter what other value it would offer. I call this issue "The Tyranny of Production"

The Alternative

In contrast, the most productive projects I've ever worked on have almost universally been with older technology. On those, it's not too uncommon to finish a new object in 20 minutes rather than 20 hours. In fact, nearly anything worth doing could be done in a single day.

The reason is simple: If technology is 10 times as powerful, then it takes 10 times as much effort to utilize it to the same level. Technology may be bigger and better than ever, but human limits are the same as they were a thousand years ago. If a person can just barely max out a certain technology in an hour of effort, then they will never be able to max out any more powerful technology in the same amount of effort.

There's an interesting lesson here about whether we'll ever truly be able to see the full potential of our current technology, much less future technology, which I'll leave for another time.

Instead, I'll make a more interesting comparison: Imagine there's a technology which is 10 times more powerful but doesn't make 10 times the difference. What if instead it it's only, say, 20% of a difference? Or 10%? It will still take 10 times the effort to realize that benefit, even if it's only a 10% or 20% difference in the final result.

At some point, it will be more valuable to spend that 10 times effort making 10 times as many things rather than 1 thing but X% better. In other words, there's a point where adding further production value to a work is less valuable than just creating another work. I call this point the "Production Event Horizon" because past that point your effort is effectively being sucked into a black hole with insufficient justification for doing so.

This is an area where for now, old technology still completely dominates newer technology. Old technology is physically incapable of doing more than a certain amount, which is one way of solving both the Tyranny of Production and the Production Event Horizon.

On the other hand, that's definitely not the optimal solution to these problems. What's a better way of handling these issues to still get the benefits of newer technology while avoiding the Tyranny of Production and staying clear of the Production Event Horizon?

Appendix

Terminology

Tyranny of Production - Anything that cannot be created at the production standard given the time, budget, and manpower constraints simply cannot be created at all, no matter what other value it would offer.
Production Event Horizon - The point where adding further production value to a work is less valuable than just creating another work.

Footnotes

[1] The Full Process Our Team Used.

For any new idea, create concept art. There wasn't a formal specification for how detailed the concept art had to be, but in practice it had to be reasonably well detailed to ensure the rest of the process captured the idea. This usually took ~4 hours per draft of concept art.
Discuss and refine the concept art. There was no point in having beautiful artwork if none of the art fit together, so we discussed new works to ensure it fit the vision of the team. Usually this resulted in some agreed changes, which meant another draft or two of concept art before everyone was happy.
Create at least two, preferably three, orthographic views of the concept art (that is, a view from in front, from above, from the left, etc.). Usually this only took about an hour, maybe two.
Create 3D geometry based on the orthographic views. Usually this took a few hours.
"Unwrap" the geometry so the texture artist can create textures for it. Usually this didn't take a notable amount of time, but could take an hour or two in some tricky cases.
If the work was animated in any way, then create any animation controls for the new 3D work. Usually took another few hours. If it involved skeletal rigging, it usually took longer.
Create the texture for the work. The time for this varied. Often it took a few hours if it needed to be done from scratch. If an existing texture could be reused with minor edits, then this went fairly quickly. If it was a complex unwrapping (e.g. a character), then this could take much longer, especially if the character used a lot of novel textures and involved much detail.
If the work was animated, this is when the animations would be created. For us, this usually took at least 8 hours. On the other hand, we didn't have a lot of simple animations so most of the things that required animating required a good bit of work.
Export the result, put it in the game, and discuss the result with the team. Come up with changes and refine. There was no real limit to this process; pretty much everything was refined over time.

Weekly WTH: Captain's Clog, Stardate 20510.9

Tags: WTH; Cap'n; Surreal; Not-so-serious

We've found a traitor in our midst. For unknown reasons, Crewman Lionel Jefferson attempted to sabotage our ship as we approach Sagittarius A*. Amidst the chaos, Crewman Charles Anderson has killed Crewman Jefferson by flushing him down the Captain's space toilet.

That was 4 days ago. Since then, we have been in the midst of a ship-wide plumbing jam. Some component in Crewman Jefferson's clothes has reacted with the lining of the ship plumbing, causing it to expand and harden, entirely blocking any and all function of the plumbing system.

The "Immortal Jam", as the crew has named it, has so far proven immune to any and all standard and non-standard plumbing techniques. It cannot be moved with hooks or plungers. It does not disintegrate or weaken with any chemical agent applied to it. It is seemingly immune to explosives. Even after replacing the pipes, the remaining particles in the plumbing system merely reacted with the new pipe to form a new clog. As Crewman Bill Hadron put it, "This is an achievement perhaps unparalleled in maintenance literature."

As Captain, it is my duty to handle any and all situations that arise in the course of this mission, and I will fulfill that duty to the highest standard I am capable of. Nevertheless, I did not become Captain so I could spend my time handling endless "Space Accidents".

I hereby petition the Federation to update it's official uniforms such that they do not promptly destroy any and all plumbing fixtures in their general proximity. I additionally petition that ships should be supplied with adequate tools to repair the plumbing system in the field.

Lastly, I request two repair crews to be sent to us: one to completely replace the plumbing system in our ship, and another to replace an airlock. This is because in the short term, I will be installing an ad-hoc system in one of our airlocks so it will function as a make-shift vacuum toilet.

And yes, I do understand how ruinously expensive a proper Federation approved airlock system is. If the Federation also understands how expensive these airlocks are, then I humbly suggest they take some measures to prevent and recover from catastrophic plumbing failures on ships with over 400 people and few opportunities for repair.

Captain's Log, Stardate 20489.7

Tags: Cap'n; Surreal; Not-so-serious

Every day on this planet, we see the black hole rise in the northeast sky. All of our knowledge tells us that black holes have an event horizon, beyond which not even light can escape. So why is it that the ship's sensors are detecting glimpses of something beyond the event horizon? We've ruled out the possibility of it being an afterimage of anything entering the black hole. Whatever it is seems to be actually there, beyond a barrier which light should not ever be able to pass, yet it is somehow still reaching us.

And what a strange sight it is. It's a seemingly grey, rocky surface, with tall square pillars that should not ever have been formed naturally. Has there perhaps been some unknown alien society which has figured out how to enter a black hole and return? And if so, for what possible purpose could they have built something inside the black hole itself?

From here, ship repairs are nearly complete and the crew has been recovered. My training with the natives seems to be complete and we can successfully fend off the unknown alien that was harassing us. Our rescue crew has arrived, and is working to identify the alien and expedite ship repairs.

When matters are in order, our next steps are clear. We will return the samples we've gathered on this mission and proceed towards the black hole to investigate this phenomenon. Other missions can either be placed on hold or transferred to other ships in the area.

Captain's Log, Stardate 20485.3

Tags: Cap'n; Surreal; Not-so-serious

Ship upgraded. Alien at bay. Seeming rescue in the distance. There is no rest from repairs, we need to be back for the next mission ASAP.

Captain's Log, Stardate 20463.3

Tags: Cap'n; Surreal; Not-so-serious

The alien was stronger than expected. I've been injured through some unidentified attack and I expect to take several days before I am recovered. My training schedule will be relaxed so I can recover. This unfortunately is likely to set back the entire recovery mission.

The techniques I'm learning are become more advanced. I may even be able to defeat this alien entirely in the relatively near future, but that must wait until my health has recovered.

Captain's Log, Stardate 20460.2

Tags: Cap'n; Surreal; Not-so-serious

The unidentified alien is cornering me, lurking around this very corner as we speak. My new training is about to be put to it's first test. I am confident I can at minimum forge an escape route. I only expect this encounter to test precisely how effective my technique is, but nevertheless I shall not let my guard down.

Training has been more intense than anticipated. Even with the aid of the natives, I have still found little rest, as their training has demanded my full attention and efforts.

I'm beginning to understand some of the basics of their unknown language. I seem to have correctly learned a few basic words, such as "bread" and "shoe." When I get out of this, I hope my notes will be the first steps to identifying their tongue and introducing them into the larger intergalactic community.

Lastly, my communicator has received new orders. It seems our shipment of samples from the last planet were not stored in a Federation approved packing order. My new orders say I am to cease all other operations immediately until I sort the samples into an approved order and ship them back to the Federation.

My communicator cannot send outward messages at this time, so I will log my response here: As the highest ranking officer in this mission, it is my sworn duty to use my best judgment in responding to a given situation for the success and safety of ship, crew, and the Federation at large. Due to my judgment of the current circumstances, I respectfully decline to fulfill these new orders.

I kindly recommend for the Federation to have someone sort the samples at the shipment facility if they have not already been shipped back when this message is received. If the samples have been shipped back and my ship and crew have not issued a new report when this message is received, then I urgently request a rescue crew and a maintenance crew to be sent to the system at 21912.19/3053.98/-54.15 and proceed towards the homing beacon therein.

Captain's Log, Stardate 20457.6

Tags: Cap'n; Surreal; Not-so-serious

It was an alien, George. It appeared in front of me and said "Don't be a maid." That was the exact moment when I decided the personal service industry would fare better without me.

I've been on the run for three days now. My provisions have offered sufficient food and water, but little chance to rest. The ship is in no better condition than when I last left her and she still needs ample work before she will be ready for take off. The crew is still separated from me, and I know nothing of their condition beyond they are still elsewhere on the planet and have no working means of communicating with me.

It seems my visit to this planet will be prolonged until I can establish basic safety for myself and my crew and restore the ship.

Today I've encountered some friendly natives who speak in an unfamiliar language. They are offering me shelter and a chance to rest, and under the circumstances I have no choice but to accept.

The shelter offered is meager, and I must balance recovering what rest I can with learning the essential skills to fight back the unidentified alien. The unidentified alien is immune to conventional weapons, but the natives know of techniques that can affect it.

Today I shall train enough to back the unidentified alien off, and then rest as much as I can. My new technique and energy will not be enough to truly escape, but I can buy enough time and space to start building a counterattack.

Final note: It seems recent events have damaged the previous log entries. I've already instated measures to secure future entries and have preparations in place to try and restore the historic logs. For now that must wait until the ship is functional again and the crew is safely recovered.

Master log, Stardate 20412.3

Refutation to Gettier Problems

Tags: Intellectual; Overanalyzing

Foreword

Once upon a time, I heard about these things called Gettier problems. At the time, I thought the concept was absolutely repulsive because to me it seemed obvious the whole thing was centered around fallacious reasoning. Yet somehow, these problems had managed to stand as a major problem for over 50 years! What a tragedy! How has the debate lasted so long without ever being resolved by the obvious!

In my ire, I managed to hash out an entire pseudo-paper on my thoughts on this. Of course, eventually I accepted that philosophy is not mathematics. Even if I had refined my arguments to meet the level of rationality and rigor I had intended, such arguments were still unlikely to convince anyone because there was always an alternative perspective they could take to sidestep my analysis entirely. And who's to say their new perspective isn't the correct one? Systematically refuting all possible alternative perspectives to leave only the valid one standing is generally an untenable problem.

I'm posting my pseudo-paper here in the hope that someone may find my work useful anyways (and also in the hope that it proves my new article posting system is working). Given the nature of how this was written, I encourage readers to take this with approximately the same level of seriousness they might for, say, Bobby Fischer's "A Bust to the King's Gambit."

Introduction

Traditionally, knowledge has been defined as a justified true belief. This view was considered overturned with the landmark Gettier problems, which claim to produce cases in which a justified true belief is not sufficient to know something.

Originally, Gettier's claims were rejected because the counterexamples were based on false premises. However, other theories later became dominant when there were complications in the false premises response.

I argue that a variant of the false premises response was in fact sufficient after all, and that it produces far less complications than either the original false premises rejection or alternative theories.

Background

The traditional justified, true belief (JTB) definition of knowledge is this:

A subject S knows that a proposition P is true if and only if:

P is true, and

S believes that P is true, and

S is justified in believing that P is true

Gettier rejected this with two counterexamples. Gettier's claim is that these counterexamples produce justified, true, beliefs, but are not knowledge, and therefore JTB is an insufficient definition of knowledge.

Importantly, Gettier makes two additional claims which his analysis depends on:

In that sense of 'justified' in which S's being justified in believing P is a necessary condition of S's knowing that P, it is possible for a person to be justified in believing a proposition that is in fact false.
For any proposition P, if S is justified in believing P, and P entails Q, and S deduces Q from P and accepts Q as a result of this deduction, then S is justified in believing Q.

For the purposes of this analysis, I will accept both of these claims as true.

Refutations

Gettier's First Counterexample

Suppose that Smith and Jones have applied for a certain job. And suppose that Smith has strong evidence for the following conjunctive proposition: (d) Jones is the man who will get the job, and Jones has ten coins in his pocket.

Smith's evidence for (d) might be that the president of the company assured him that Jones would, in the end, be selected and that he, Smith, had counted the coins in Jones's pocket ten minutes ago. Proposition (d) entails: (e) The man who will get the job has ten coins in his pocket.

Let us suppose that Smith sees the entailment from (d) to (e), and accepts (e) on the grounds of (d), for which he has strong evidence. In this case, Smith is clearly justified in believing that (e) is true.

But imagine, further, that unknown to Smith, he himself, not Jones, will get the job. And, also, unknown to Smith, he himself has ten coins in his pocket. Proposition (e) is true, though proposition (d), from which Smith inferred (e), is false. In our example, then, all of the following are true: (i) (e) is true, (ii) Smith believes that (e) is true, and (iii) Smith is justified in believing that (e) is true. But it is equally clear that Smith does not know that (e) is true; for (e) is true in virtue of the number of coins in Smith's pocket, while Smith does not know how many coins are in his pocket, and bases his belief in (e) on a count of the coins in Jones's pocket, whom he falsely believes to be the man who will get the job.

The basic premise here is Smith deduces that the man who will get the job has ten coins in his pocket from the fact that Jones has ten coins in his pocket and the false premise that Jones will get the job. Gettier argues that if a belief is justified, then what is deduced from that belief is also justified, so therefore Smith is justified in believing the man who will get the job has ten coins in his pocket.

Suppose Gettier's claim of justification propagating through logical deduction is true. In deductive logic, an argument must be both valid and sound in order for it to be an acceptable argument.

A deductive argument is valid if it is impossible for the premises to be true while it's conclusion is false. An argument is sound if it is both valid and it's premises are true.

Smith's deduction is valid: If Jones really does have ten coins in his pocket and Jones really will get the job, then inevitably the man who will get the job must have ten coins in his pocket.

However, Smith's deduction is not sound: Jones will not get the job, so therefore not all premises are true.

The primary issue with this example, then, is Smith's deduction is not in fact a sound logical deduction, and therefore Gettier's second claim does not apply. If Gettier's second claim does not apply, then Smith is not justified in his belief that the man who will get the job has ten coins in his pocket. And if Smith is not justified in his belief that the man who will get the job has ten coins in his pocket, then there is no conflict between the concept of a JTB and the concept of what Smith knows.

The Previous False Premise Solutions

This so-called "false premises" provision has been raised before in several variations. Some notable variants claim the definition of knowledge must be altered so that justified true beliefs that do not depend on false premises. These variants ultimately fell out of favor because they opened the question of how deeply one would have to prove the chain of premises before the argument became acceptable.

This problem is unnecessary however, because the definition of knowledge does not need to be modified to account for false premises. The JTB definition of knowledge needs no concept of premises because the definition of knowledge is not the definition of logical deduction. Logical deduction does require true premises in order to make a sound argument, but logical deduction is not the sole method through which a belief can be justified. Therefore, there is no issue of when to stop because at some point there will be justified, true beliefs that have been justified through means other than logical deduction.

Generalized Gettier problems

Some have generalized Gettier problems such as the so-called "Sheep in the field" problem.

Imagine that someone is standing outside a field looking at something that looks like a sheep (although in fact, it is a dog disguised as a sheep). They believe there is a sheep in the field, and in fact, they are right because there is a sheep behind the hill in the middle of the field. Hence, they have a justified true belief that there is a sheep in the field. But is that belief knowledge?

This argument has a similar issue of a hidden unsound deduction. Let's call this person Bob. Bob observes (a) he sees something (b) this something looks like a sheep and (c) this something is standing in a field. From (a) and (b), Bob concludes that he sees a sheep. Critically, note this is a justified false belief, because the something is not in fact a sheep. Bob deduces that because he sees what he believes is a sheep, and (c) that something is standing in a field, therefore there is a sheep in the field. However, this deduction is unsound because it is based on the false premise that what Bob is seeing is a sheep. Therefore Bob's belief that there is a sheep in the field is not justified, because his justification for believing the dog is a sheep cannot propagate unless the logical deduction is sound.

Note that in fact, this problem is of exactly the same form as the original Gettier examples. The only difference is the erroneous deduction was obfuscated by the wording of the problem. Nevertheless, the deduction is both present and essential to the problem. If Bob had not gone on to deduce that the sheep is in the field but remained with his original statement that he believes he sees a sheep, the problem fails. In that case, the solution would simply be Bob has a justified false belief and the fact that there's another unseen sheep in the field never comes into play.

In previous variants, it has been claimed this type of explanation is insufficient. The argument goes that because the sheep was seen through sensory data rather than deduced logically, the justified false belief does not apply. It is argued it is unreasonable to reject sensory data as requiring logical justification, since sensory data is the primary means through which we observe the world. If that is insufficient, then virtually nothing can truly be justified.

My response is two-fold. Firstly, arguing sensory data is above considerations of uncertainty while also advocating arguments which depend on uncertainty in sensory data is a logically unsound position on this issue. Secondly, the conclusion that virtually nothing can truly be justified with uncertain sensory data is unsound, which I will demonstrate.

Imagine Larry has just taken some LSD. He looks around his room and notices there is a cat in front of him. He considers seeing a cat to be sufficient justification, so he thinks he is justified in believing there is a cat in front of him. In fact, there is a cat in front of him, so Larry has a justified, true belief that there is cat in front of him when in fact there is. By the JTB definition, Larry knows there is a cat in front of him.

Say the cat leaps away and a different cat leaps in front of him. By the same process, Larry concludes there is a different cat in front of him, and by the JTB definition he knows this cat is in front of him as well. Say that cat also leaps away and a different cat leaps in front of him. Say this process continues until 9 cats have been in front of Larry, all of which Larry saw, and all of which Larry knows were in front of him by the JTB definition.

Imagine now Larry sees the 10th cat leap in front of him. By the same process, Larry believes there is a cat in front of him and is justified by virtue of seeing it in front of him. This time, however, there is not a cat in front of him. This cat is merely a hallucination brought on by LSD. This time, then, Larry has a justified false belief, and so does not know there is cat in front of him by the JTB definition (although he believes there is a cat there).

What does one make of this case? Larry knows all 9 actual cats are present, but was deceived into believing there was a 10th cat as well. Therefore, only 9/10ths of what Larry concluded was actually true if sensory data was his only means of justification. If Larry suddenly stumbled across an oracle who could perfectly tell him which cats were real or not, it would not constitute a drastic reduction in his justified beliefs because 9 out of 10 of his justified beliefs were already true.

Thus, if there is an unreliable justification which can nevertheless distinguish truth from fiction a substantial majority of the time then the vast majority of beliefs justified from that will in fact be justified true beliefs. Therefore, it is incorrect to conclude that an imperfect justification method implies little or no beliefs can truly be justified.

Of course, these problems could always be formulated differently, in which case they will merely get refuted differently.

Appendix

Author's note: This was going to be a counter-argument to some other arguments I read about Gettier problems, but I never got far enough into writing the proper paper to use it.

Some argue that this constitutes strengthening the definition of justification, and therefore is not equivalent to the original JTB formation. I argue that regardless of how you interpret what was "sufficient" justification originally, arbitrary justification was never a valid interpretation.

Imagine Alice stumbles across a claim. Alice believes the claim to be true, but knows objectively the claim is equally likely to be either true or false. So Alice flips a fair coin, saying that if the coin is heads it will affirm the claim is true, and if the coin if tails it will affirm the claim is false. The coin comes up heads, so Alice decides she is justified in believing the claim is true.

Say the claim happens to be true. Does Alice have a justified, true, belief? The claim is known to be true, and the problem asserts Alice believes in the claim, so the question simplifies to "Is Alice's true belief justified?"

Alice's justification for her belief is she flipped a fair coin and it came up heads. If the coin had come up tails, Alice would have rejected the claim as false. Therefore, her justification process was equally likely to justify the claim as to refute it. However, the claim was already equally likely to be true or false even before Alice performed her justification. Effectively, her justification added nothing that Alice didn't have before she attempted to justify her belief. Let's call this a "null justification."

Is a null justification an acceptable justification for a justified true belief? For any belief, a null justification can be constructed in order to satisfy the justification condition. If this is acceptable, then the justification clause can be eliminated entirely, and the JTB definition reduces to simply a true belief.

Therefore, regardless of how you define justification, it must be at least strong enough to exclude null justifications, since that is logically equivalent to having no justification at all. This is implicit in the definition of a JTB, since otherwise the definition would be reduced to simply saying a true belief is acceptable. Thus, disallowing null justifications is not a strengthening of the original JTB definition, regardless of how justification is defined.