blog.lettucehead.com

Humans are Software!

During COVID, I had a conversation with Mark Slee, who wrote Chromatik and other stuff, about this Lobsters post by ferd, who wrote Learn You Some Erlang.


From: Charlie Eddy <charlie@lettucehead.com>
To: Mark Slee <xxxx@xxxxxx.xx>
Subject: "complexity"
Date: Fri, 1 May 2020 16:43:00

Any chance you find this interesting?

https://lobste.rs/s/emww5p/complexity_has_live_somewhere

> On Sat, May 2, 2020 at 9:27 AM Mark Slee <xxxx@xxxxxx.xx> wrote:

Yeah it was interesting. I get what this guy is driving at, but think I agree with a lot of the commenters pushing back though. My two main reactions:

(1) Complexity doesn't *have to* exist, at all. At the most pedantic zoomed out level, entropy increases, we're on our way to the heat death of the universe. We are all free to stand by, do nothing. No complexity to it, thermodynamics can slowly do its thing and that's that. (Open question: is the universe in "heat death" state more or less complex than before? I don't know, I could see arguments either way - I think this reveals a flaw in the argument that complexity "must go somewhere" - I'm not even sure all humans would agree about what's "complex" and what's not).

But at a non-pedantic level, I think this kind of observation applies to software design too - complexity is not *required* depending upon what you're trying to do. As a designer of software (or any tool), you get to decide what the goals are, what problems are worth solving, etc. As he points out, if you put some things off the table, some other guy might come along with some other goal of his own, and hack around your thing and make a big mess. IMO this is the other guy's problem, not yours. Skilled design IMO is often as much about deciding what problems not to solve as it is finding clever/elegant/complex-but-contained solutions. The fact that somebody else has different goals than me, which might create complexity - seems more an observation about the fact that we cannot control other people, than about the fundamental requirement of complexity existing somewhere. Complexity is a human creation, a subjective judgement from our POV. The universe itself doesn't intrinsically contain or not contain "complexity," and yet people can create and experience varying amounts of it.

(2) If it must live somewhere, some places are better than others! I guess this is sort of his point, but it feels like he framed it in a weird way and is getting a lot of pushback because he makes it sound like he's arguing for the opposite - as if that there's some "Universal Law of Conservation of Complexity."

Silly example, let's swap "Complexity" for "Garbage." I could pile up all my garbage in my back yard. Or throw it in the river. Or we could have a functioning waste management system.

In the functioning waste management system world, no one is under any illusions that "garbage does not need to go anywhere." Yes, we get that garbage exists, and that it must go somewhere. But we actually don't have to spend much of any time at all thinking about it, acknowledging it, adapting to it, designing around it, educating our users about it (beyond "put your bins out on Tuesday"). When a system is well-designed and well-functioning, I think it *should* actually enable you to behave *as if* the problem doesn't exist (garbage in this example, complexity in software).

Also - of course, we can actually control how much garbage we create. You could say that all this cardboard packaging comes from trees so it was there before, it's got to be somewhere! And yet no one seems to perceive the trees that were used to make the paper and the throw-away cardboard boxes as having been "garbage" to begin with. Or view "trees" as a "well-designed pre-emptive garbage-containing system." The garbage was created by the decision to create and use all these boxes that would then immediately not be needed. Didn't *have to* happen.

Again, maybe this is not actually at odds with what the guy is trying to say. But I find something about the direction he's coming at it strange. One of his points does actually seem to be that you ought to think carefully about where the complexity goes. I of course agree with this. But he seems to simultaneously defeat himself by making this parallel claim about universality that I think feels flawed.

> On Sat, May 2, 2020 at 10:54 AM Charlie Eddy <charlie@lettucehead.com> wrote:

What i love about the discussion, here, is that this great Unixbeard's rather self-defeating & strange, possibly doubled-up, arguably flawed argument about thinking carefully, which could be understood by anybody, results in the following emotional reactions, from an audience who likely considers itself well-equipped for objectivity:

He is accused of being dangerously wrong, justifying shoddy work, of speaking about culture/education and treating tech like social science, of trying to overcome an indestructible "will to ignorantly producing crap," of ignoring time constraints, ignoring free choice, and completely changing arguments, -- of trying to encompass too much, of talking about a problem that doesn't need to exist, or possibly about having "unnecessary" problems imposed by powerful third parties (???), of making a defeatist and responsibility-shirking argument that is either trivially true or empirically false (!!), and arguing against having insights or changing one's point of view, and being counterproductive and factually wrong, -- of nihilism, -- of arguing that software cannot be maintained (and told he "needs treatment" for burnout), of being far from reality and inessential to software development, and of being easily misunderstood, yet, in the same breath, flying in the face of experience (which prompts an imaginary dialogue featuring these amazing lines:

The code is simpler, yes, but at what cost?
Can we not say that the training required to prevent that engineer from overcomplicating a time routine is itself a form of complication?
Can any among us not indeed be considered but spirited complications upon the timeface of software development?

... by a former participant in our garden party), of bordering on unproductivity and relying on red herring examples, of arguing for the lawful conservation of unfulfilled desire, of (again) making dangerous assumptions, of mistakenly conflating complexity with a seemingly made-up concept (though see here) (note: this ends in agreement... so why did it begin?), -- of not using enough words or aspects of words, of not using enough metrics, of not referring to an important article by an important thinker which would make his definitions "REAL" (quote marks indicate original emphasis...) (also note: the commenter recommended reading a 66-page paper, a 21-page paper, and watching a 1h 55m video, and also learning a new form of "hexagonal" architecture), -- of not possessing a complexity-separating machine, of not operating in a task-oriented space (an interesting though not precisely informative story then climaxes in the insight, that misusing language makes people want to reply with fsck(8) and $#!@), of not testing his argument in reverse before making it, of causing magical fires, of arguing complexity doesn't meaningfully exist, and of starting a definition thing.

This post had equal comments and upvotes, and a high number of both (at a fast rate). It is rare that a "philosophical" question makes it to the front page and stays there, so it must have been pretty triggering. For comparison, a very well-received open source project gets 100 upvotes, and this has 50 upvotes. The Linux CoC debates had 100-150 upvotes (with many, many comments), and were (at the time) considered to "tear the community apart."

As to the author's response, I think this comment, here, seems very fair. Would you agree?

By the way, I thought what you wrote was very clear.

> On Sat, May 2, 2020 at 3:15 PM Mark Slee <xxxx@xxxxxx.xx> wrote:

Hah, yeah the reaction is pretty wild. I see I'm not the only person who invoked the laws of thermodynamics in responding to this.

I think the reaction is really a religious one, and it's a strong as it is because this guy is a part of the "in-group." If some clueless bozo who wasn't a good engineer wrote something like this, this group would probably shrug and ignore it, but it seems like this guy clearly is an experienced engineer.

Feels akin to if a member of a church stood up and said "look everyone, we're really just shuffling around uncertainty here, yes we might build and cultivate a strong faith, but that's not really reducing any cosmic spiritual doubt in the universe, is it? nothing we can do can alleviate that whatsoever, we're just pushing it around." I wouldn't expect the churchgoers to be very happy.

That's sort of what I see happening here: A great Unixbeard who's a member of a group of great Unixbeards, many of whom have presumably spent years developing a high level of skill in elegantly solving complex problems. And this guy seems to be clumsily handwaving the whole thing away as just shoveling $#!@ around (again, I don't think this was his actual intent, but the way he framed it all comes across like that). I think the reaction is as much to the clumsiness and the threat of the in-group norms (that skillful programming is a worthwhile endeavor which very much can actually meaningfully increase our problem solving-power and manage complexity).

Also - there seems to be a bit of self-contradiction here. If the point of his article is that "careful thinking" is valuable, it's sort of remarkable how he missed out on clarifying that he's talking about Human Complexity, not Software Complexity.


It is my opinion that now is the greatest time in history to be an artisanal programmer. You may have to use AI at work, but you might be able to see how AI is just trained on the work of artisanal programmers.

I have gotten immense value from participating, as a reader, in Lobsters. The community is very smart. Like many smart communities, it is prone to fracture. I care. While the benefits of caring are non-obvious, especially since the outcome is entirely outside of my control, the benefits of open source are obvious.

That is why I have become increasingly interested in programming during this tumultuous time. I personally love being tracked, because I love recieving relevant ads. I love being forced to click modals, because I love GDPR compliance. I love email, because I love the way it works, and I love DMARC, because I hate spam.

I was once a member of a school for computer programming, and am restarting it. I've already ordered my v100s and made an agreement to rent the electricity I'll use up on my friend's dusty old motherboard.

I worked with a great Unixbeard I met at the Berkeley Unix User Group to set up mumble along with six other hopefuls. We all failed. That was then. This is now. There is hope for my friend who wants to run GrapheneOS. There is hope for my friend who wants to "degoogle." There is hope for me as well, and I could care less about all that nonsense. My only goal is to eat enough metaphorical Cap'n Crunch that I can get the box tops to get the whistle. I just want to read writeups about O(n log n) and random seeds. We're at a fork() in the road. By the way, that was a really interesting paper, and the commentary both three years ago and in 2019 was FANTASTIC READING. I give a damn. I would appreciate if everyone would please just agree with me. Lobsters is not a replaceable resource.

Back then, it was a commonplace that so-called artifical intelligence was a series of IF statements, and so-called virtual reality was novel. I submit to you that reality has always been novel, virtual or otherwise, and intelligence has always been artificial. I figure if the IF statements you're writing at work are good enough to pass muster, then the IF statements I'm writing at home should therefore be, from a sufficiently naive perspective, equally liable to result positively.

What is more notable than the contents of anyone's particular chat is the digital world speaking to itself in a soundless echo chamber of inconceivable proportions, operating in a blinding rapidity at a microscopic scale.

The dalai lama is believed by devout theravadans to have a thousand eyes and a thousand arms with which he helps out the true believers. Also eleven heads. However these characteristics are the unmistakable mark of not only the object of Buddhist adoration but also of the cloud services emanating from the pure land, whose CDNs are as ubiquitous as the clear light of awareness. I, a dancer in the stream of events, know that the ultimate forms of marketing content are an aspect of buddha-nature, but that the particular buddha-natures which I experience as spam are only a reflection of my dependently arising complex constellation of thetans and midichlorines.

All of which is to say, tag or no tag, I respect the idea that there should be a circle drawn around the chaos, and I want to be able to pass in and out of it, and know which side I am on. If we keep the tag, we will be able to draw the circle. If we remove the tag, we lose the reason for removing it - keeping the people in who want all the garbage to stay out. It is my personal opinion, and it is decidedly neither professional nor unprofessional, that we should cave to the extremist view, and continue to advocate for the artisan at Lobsters.

If you are still unclear, let me be very direct about my opinions on artificial intelligence what it enables. It enables the fully automated luxury gay space communism &/or fascism, whether TESCREAL or HEPAITADIASTMT (humanist, entropian, presentist, anthropocentric, irrational to a degree, ineffectively altruistic, short-to-medium-termist — I'm just smashing words together!) that we all deserve. I will clarify my opinions with prose sufficiently purple that a child would understand it. Once upon a time, there were programmers...

I, an ever-roaming Najlorpa, came upon a dakini, who I trained to recognize cats by showing her one thousand cats and one thousand noncats. She, released from her karma, was able to be liberated from samsara by the force of nondual awareness. Upon the completion of her training, a tabby ran across her path. "Do you see that?" she said. "What?" I inquired. "A thousand cats and a thousand noncats," she said. Upon hearing this, I knew she was enlightened.

OM MA NI PEDME HNNNG!