Reality has a surprising amount of detail (2017) | Hacker News Reader

Reality has a surprising amount of detail (2017)

johnsalvatier.org

397 pointstosh136 comments

I blame this phenomenon for why people feel mathematics is not useful. It is, it's just that reality is more complex than high school mathematics can properly prepare you for. To understand the mathematical solution, should you encounter it, does however require some foundational knowledge so if you don't know any then you can't even begin to understand it (possibly even failing to identify it is mathematical).

I mean take something practical like a mortgage. It's fairly easy to calculate the annuities using a geometric series, but that places it beyond most people's mathematical skills. Sure you could use a special calculator, or if you're adventurous look up the formula (beware though, Wikipedia is bound to lead to errors by defining the monthly rate as yearly rate / 12), and then it probably may not like doing mathematics at all but neither do you really understand what's happening.

"If people do not believe that mathematics is simple, it is only because they do not realize how complicated life is" - John von Neumann [1]

Saw this quote on an art blog showing SEMs of diatoms [2]

[1] https://butdoesitfloat.com/Diatoms [2] https://en.wikipedia.org/wiki/Diatom

edit link

Yay John von Neumann, genius of geniuses, also said Young man, in mathematics you don't understand things. You just get used to them.

Yep, 10x that for real life.

Then you can start layering on the complexity by estimating repayments based on the expected variability of future interest rates, convert to Real Dollars based on inflation and expected wage growth, then compare with alternative options such as renting + investing, etc...

It's astounding how some of math's first practical applications is to calculate movement of the stars. Absence of frictions and close bodies make physics much easier to approximate than anything on earth. Maybe to have students appreciate math, we should have them predict eclipse one again. God create astronomy to teach humans mathematics.

The oldest writing is numbers... for accounting. I'm pretty sure Pythagoras' theorem was for property boundaries.

Although the oldest mathematical work I'm aware of on accounting is early medieval (al khwarizmi).

I dunno. It seems to me like high school math is a pretty decent foundation for understanding many important aspects of the world (derivatives, integrals, etc.).

And I think that one can clearly and substantively “get” and intuitively understand and intelligently work with an amortizing loan without needing to understand whatever level of math is required to comprehend why pressing (g) (12i) on an HP 12c is not an atomically precise representation of of the answer out to the 12th decimal place.

> beware though, Wikipedia is bound to lead to errors by defining the monthly rate as yearly rate / 12

Looks like all the calculators do that as well. What's the right way?

My mortgage company seems to use the same formula.

I assume to take compounding into account, you want the 12th root. Instead of 12% → 1%, 1.12^(1/12) gives around 0.95%.

The effective interest rate should be calculated and then converted to the monthly rate:

$$ \left( 1 + \frac{i_a}{n_a} \right)^{n_a} = \left( 1 + \frac{i_b}{n_b} \right)^{n_b} = \left( 1 + \frac{i_{\text{annual}} }{1} \right)^{1} $$

https://www.investopedia.com/terms/e/effectiveinterest.asp

My mortgage company charges interest based on the number of days in the month divided by 365. I have replicated their calculation like this. I'm not sure what they do in leap years - there are at least 3 distinct approaches.

It could be correct, it just depends. Also according to the comments banks do all kinds of weird things.

This is a bit of a rabbit hole. The right way is whatever is agreed, and there are many ways to agree it.

https://en.m.wikipedia.org/wiki/Day_count_convention

Yet you will find that mortage interests charged by the bank don't match your series, because they use an accrual schedule with a special way to measure years fractions, then calculate an equivalent notional for those dates, and remap that to your payment calendar.

Turns out there is a surprinsing amount of details here as well. And you don't have all parameters for the calculation, like often in life.

> (beware though, Wikipedia is bound to lead to errors by defining the monthly rate as yearly rate / 12)

Lead to errors relative to what? The mathematical idealisation or the actual practice?

Are you referring to this page?

https://en.wikipedia.org/wiki/Mortgage_calculator

"Since the quoted yearly percentage rate is not a compounded rate, the monthly percentage rate is simply the yearly percentage rate divided by 12."

Do you think that the explanation above is wrong?

In practice interest payments are calculated in many "wrong" ways, but that's what it is:

https://en.wikipedia.org/wiki/Day_count_convention

“All models are wrong, but some are useful”

https://en.m.wikipedia.org/wiki/All_models_are_wrong

I was looking at: https://en.wikipedia.org/wiki/Mortgage#Principal_and_interes...

It's not necessarily wrong, but it's missing any disclaimer about which interest rates they're talking about so if you don't know what you're doing it will lead to mistakes.

Ok, so defining the monthly rate as yearly rate / 12 will lead to mistakes if the yearly rate is not twelve times the monthly rate.

(Of course any other definition of monthly rate will also lead to mistakes when it's inconsistent with the definition of yearly rate.)

Cant you just use 365 instead of using something less accurate like 12 months (months are unevenly distributed)

You are correct about the mortgage payments, I had a 1-2 cent error every few months, after a days struggle, I simply added a “Penny” column to add or subtract as necessary so that they would match!

Aside: Building stairs that way is a really bad idea. Don't use angle brackets – you cut stringers. And besides the first and last step, stringers are pretty straightforward to layout with a carpenter's square with some stair guides (two little posts that screw to the square). And, generally you just layout 8" rise, 12" deep stairs, so you really don't lay out anything at all.

But, really, don't build stairs like OP suggested.

Somehow as easy as it is conceptually, I still find cutting 3 stringers that line up well to be surprisingly difficult. In part because I’ve only has to do it a handful of times in my life.

Either you use jigs or you tie them together temporarily and cut the stack.

pretty amazing how the simplest of solutions seems to be so elusive. i've been guilty on more than one occassion of making something seemingly simple as difficult as wrangling cats.

It's just experience. Everything you've done before or that you've seen done by people that knew what they were doing is trivial, everything you haven't done before is difficult.

Yup. If you're doing almost anything in wood, a set of clamps is more useful than you'd think.

I made a table. The legs are all exactly the same length, but don't ask me what that exact length is - I eyeballed the height I wanted and then clamped all the legs together before cutting them.

Another tip for building things with legs - a 3-legged object is stable on any uneven surface while a 4-legged objects will wobble on uneven surfaces.

This is because 3 points make a plane, adding a 4th point that is not on that plane introduces a wobble.

Easiest way to avoid that wobble: get the legs to within 1 mm, add felt pad to the bottom of the legs. The pressure of the table will compress the felt and all four legs now contact the ground.

> Easiest way to avoid that wobble: get the legs to within 1 mm, add felt pad to the bottom of the legs. The pressure of the table will compress the felt and all four legs now contact the ground.

On a level floor, sure. But for something that will be on uneven surfaces, like outside, 3 legs are stable, even if the legs are all different sizes[1].

[1] Three points make a plane, the plane itself may not be level, but because all points of contact are on the same plane there is no wobble. When there are more than 3 legs, some points may not be on the same plane, producing a wobble.

Yep. That's the whole principle behind stablity mechanisms such as tripods and easels

A sister comment to mine gives you one strategy, but you can also use the first stringer as a template to the rest.

You can create a template online, print it out, and use it to get perfect results:

https://www.blocklayer.com/stairs/straighteng, click "Show Notching Template", then "Diagrams to PDF"

I also don't like that fixtures at top and bottom are SPOF. Why not add pillars to the top and ~3rd from bottom steps, then tie with another beam, like an H overlapping a ladder?

That way, there will be far less chances that bolts spontaneously rip off and stairs go down, nor the lengths of pillars to have to be critical dimensions. The pillars can optionally be cut to precise lengths to be screwed through both steps and slopes for maximum Apple-ness, or, can be lazily cut to long-enough lengths, nailed through to steps from the "outside", and used as base for handrails as if it had been the plan.

(dc: not an engineering advise. consult a real engineer for safety. Also add cross beams in width-height plane, they help tremendously)

The author should just have learnt how to build stairs in a craft school instead of wasting time on unprofessional trial and error.

I'm not sure that author in that times was able to afford craft school..

Yeah, having the steps only connected by some small screw brackets is borderline redneck engineering, someone's gonna get killed when two screws come loose.

What OP was really making were misused ladders.

Also related to the ending, I‘ve come to realise more and more that most people reason out of belief first and arguments second on a lot of things (maybe most things?) Climate change deniers are an obvious example. But also more nuanced views such as "the software I make makes the world a better place" (I belief this myself), "me being a healthcare worker is a great thing for humanity", or "I need to game at least once a week to relax." And it makes sense. Many things are extremely complex and so picking one side and going for that will make sense in many cases. It’s often necessary to be able to make decisions. However, the risk is getting too stuck in certain ideas. It’s easier to accept a long held belief and reject opposing information than to re-evaluate.

Professional deniers are propagandists, not good-faith debaters. Lack of nuance is one problem, but there's a far bigger problem with groups of people knowingly trying to poison and undermine anything that resembles reality-based consensus.

Absolutely - PR firms, lobbyists, astroturfers, troll farms - I'm in full agreement, and there's a lot of them around.

Also a problem are those who label the undecided or free thinking among us as not being good faith actors. And unfortunately it's a tough position to be in, as you get attacked from all sides for not choosing a side.

"reality-based" and "consensus" is an interesting combo, in general but particularly if one considers the history of consensus of "reality".

For some reason this really speaks to me.

I think it's because we create a story about every part of us and our life. Every thing we do (sometimes we even do it after the fact).

I feel like those beliefs are what's keeping us grounded in the sense of understanding the world, of being in control. so it makes sense that we would start with them even if we are not aware of it.

And it makes sense that it's hard to let them go because without them you get a sense that you are just floating away and don't have any to "hang" your assumptions on.

You can reason a lot but there comes a point when things get bigger than you and you have to trust some other authority or just trust your gut.

There's not many people that could comfortably reason from first principles and be satisfied with where they end up. (I don't think I could do that)

Actually correct take: You cannot slow down climate change without mass depopulation, and most climate change alarmism is overhyped. It will impose a cost to civilization and it will have to be dealt with by adapting to it, rather than some inane scheme to create carbon tax credits to justify more ways for the hyper-wealthy to stay wealthy.

> belief first and arguments second on a lot of things

"support rather than illumination"

> I‘ve come to realise more and more that most people reason out of belief first and arguments second on a lot of things (maybe most things?) Climate change deniers are an obvious example

Do you think most climate change believers reasoned themselves into that position? I think for >95% of people, belief in climate change is a purely social phenomenon. The overwhelming majority of people don't understand enough about physical phenomena like blackbody radiation to have any intelligent opinion about climate change one way or the other.

I do not think that. This similarly applies to both sides of some other contentious scientific/medical topics which can easily get someone labeled an “antivaxxer” as an ad-hominem directed without solid reasoning. One imagines what labels were given to those who did not support widespread leech therapy at its peak…

The untrue explanations are often the simpler ones. They are easier for people to "understand" even if they are wrong. And people don't like to be perceived as dumb, therefore they keep on arguing that they are right.

Example: Flat Earth. The horizon looks flat, therefore the Earth must be flat.

> Also related to the ending, I‘ve come to realise more and more that most people reason out of belief first and arguments second on a lot of things (maybe most things?)

This is the thesis to the introduction to Jonathan Haidt's The Righteous Mind. Boiling it down, he argues that self-justification is the most fundamental human reaction. We reason after, not before, and our reasoning flows to align with our own justification.

(Luther would be proud.)

Lossy compression.

This book touches on similar ideas: https://en.wikipedia.org/wiki/The_Scout_Mindset

I’m a data scientist, and it always surprises me how much work goes into fields that do not require a degree — easily just as much (or more) thought is required to solve “daily” work tasks.

I needed to redo a closet in my house where the original wireframe shelves were not installed correctly and fell out as soon as any weight was put on them. First you’ve got to patch the holes. But spackle will show flashing. Ok, so I’ll use joint compound instead. But there’s 27 brands and types, all for slightly different purposes. And there might still be flashing if you don’t use the right type of PVA primer. And you need the right sanding tools and, oh, it turns out I suck at applying joint compound so I have to call someone anyway to make it look like years of expertise went into that.

On to the new shelves. Well I don’t want the closet to look like I did it myself, so I’ll get some floating shelves to make it look nice and modern. Wait, $1,200 for a piece of wood?? Can’t I just find a tree in the backyard and cut that myself? Oh... I need specialized equipment to make that look nice too. Whatever I’ll just order the shelves.

Have the shelves, but now it turns out the tolerances on the walls in the closet weren’t that tight when it was built. It seems like the width of the closet varies quite a bit as you move away from the wall. So now there’s a gap. And the floating shelf metal bracket misses the nearest wall stud by 1/8th of an inch, which would halve its load capacity. I’ll just drill another hole in it. Oh, it’s stainless steel and that requires specialized equipment and fluid to prevent overheating and...

So yeah, I totally agree with this article in every way. It’s amazing anything works at all in society.

I enjoyed the stair story here quite a bit - I find that I always have a newfound appreciation for the “real” jobs when I find a fine craftsman at work.

We had a curved staircase fabricated for our house and it was absolutely fascinating to watch the planning and execution go into this task. The original measurements were plugged into a cad drawing where we could adjust the number of treads and the position of the top and bottom steps based on our preference. Then they fabricated the entire assembly offsite and brought it in via crane in one piece. Even building the handrail was fascinating as there was a jig built in place to bend the wood to the curve required.

The physical constraints and fractal complexity is fascinating for someone like me who is used to deterministic machines moving bits on a daily basis.

Previous threads with 100+ comments:

https://news.ycombinator.com/item?id=29429385

https://news.ycombinator.com/item?id=22020495

https://news.ycombinator.com/item?id=16184255

Thanks! Macroexpanded:

Reality has a surprising amount of detail (2017) - https://news.ycombinator.com/item?id=38304840 - Nov 2023 (1 comment)

Reality has a surprising amount of detail - https://news.ycombinator.com/item?id=36309597 - June 2023 (1 comment)

Reality has a surprising amount of detail (2017) - https://news.ycombinator.com/item?id=29429385 - Dec 2021 (118 comments)

Reality has a surprising amount of detail (2017) - https://news.ycombinator.com/item?id=28006256 - July 2021 (1 comment)

Reality has a surprising amount of detail (2017) - https://news.ycombinator.com/item?id=22020495 - Jan 2020 (115 comments)

Reality has a surprising amount of detail (2017) - https://news.ycombinator.com/item?id=16184255 - Jan 2018 (294 comments)

> If you’re trying to do impossible things, this effect should chill you to your bones.

> It means you could be intellectually stuck right at this very moment, with the evidence right in front of your face and you just can’t see it.

This speaks to me. It's terrifying knowing deep in your soul that something is possible but getting stuck trying to make it happen. When you try to talk to people, you seem a bit unhinged. Like some crazy person who's creating his own problems by avoiding the easy established solutions. People will straight up tell you it can't be done, they will ask you why you can't do it in the normal way. It's tempting to take their word for it and just give up. However, if you persevere a little and ask the right questions, they might just see that it is possible to do things differently. If you persevere a lot, you might figure out all the details and end up proving that it was possible after all.

I achieved a small victory just like this a few days ago and I'm really proud of it, I'm actually preparing to Show HN.

This is an unusually rare, worthwhile piece. It's rare for somebody to be both very grounded, with a personal history of hands-on practical competence, yet simultaneously abstracting his experience to a high philosophic level.

A nitpick is this statement: "The massive difference in weight between a rocket full of fuel and an empty one means that a reusable rocket can’t hover if it can’t throttle down to a very small fraction of its original thrust, which in turn means it must plan its trajectory very precisely to achieve 0 velocity at exactly the moment it reaches the ground."

I don't think this is how SpaceX does it. When you have closed loop feedback control using velocity and distance measurements from radar and vectoring thrusters, it's no longer an impossibly difficult ballistic problem. I suppose that highlights the necessity of not making assumptions about solution methods.

The original statement is correct: the rocket cannot throttle down to a small fraction, to an equivalent thrust to the mass of the rocket. SpaceX rockets indeed do not "hover" for any significant length of time. The trajectory is planned to achieve 0 velocity at just above ground level. The feedback control comes in because it's impossible to know the exact performance and atmospheric characteristics, so if you tried to do it by dead reckoning it would be off vertically by tens of meters, so the trajectory is continuously adjusted.

Stolen from previous discussion https://wikipedia.org/wiki/Coastline_paradox

The gif of Great Britain looks like it freezes at the end, but zoom in... it's just adding in finer detail.

I thought about fractals while reading the article too.

And I'm starting to worry that maybe fractal-like concepts not only apply to geometry (as in, "how long is the coastline", or "what direction is the coastline at point (x,y)?"), but to truths in general.

Just like the article describes, "rough" facts like "water boils at 100C" can have tonnes of nuance, and when you drill down into the details, you might discover that the truth is so much more complicated that the notion of truth and falsity is lost among the vast sea of complex details. (eg. https://en.wikipedia.org/wiki/Phase_diagram#/media/File:Phas... ... who'd have thought something taught in elementary school would be so complicated :-/ )

So is it possible that most "truths" in which we have absolute conviction are actually just rough understandings, and if we dig deep enough we'd find that we're actually mostly wrong? It seems insane to think that way, but I have a strong intuition that this might actually be the case.

I'm personally convinced that a lot of arguments in modern life boil down to not going to a sufficient level of detail. Mostly I find that both parties in an disagreement have valid points and simply haven't defined their terms precisely enough or been clear that they are talking about a p95 vs a p5 situation.

I agree, most of the time you need to refine and quantify the question.

People are resistant to that though. If I'm being generous I think they think it's just saving time, because they 'know' the right answer. But the melancholy cynic in me wonders if it is just ego.

I don't think it's cynicism to realize that people are imperfect creatures subject to the whims of insecurity and ego, not being well slept, yadda yada. To me that's just realism.

it can be hard for me to admit that my entrenched beliefs may be flawed but enhanced with an updated worldview…now me thinks this is my ego…actually i am now sure of it…a detail i had not previously flushed out…

This basic idea was the foundation of the logical atomist philosophy of Bertrand Russell and the early Wittgenstein.

Agreed. A lot of the past conflict in my life boils down to this as well.

Conversely, I'm now in the habit of asking big "obvious" or "dumb" questions; questions that will highlight assumptions being made by each party. I'll often go slow at first to establish that these large unspoken thoughts are aligned and then work my way down to lower levels of detail until we hit an actual disagreement, rather than a mere misunderstanding.

It drives me crazy to spend 30 minutes arguing with someone only to find that "Wait, you're talking about THAT? I thought we were talking about THIS."

It's too common for people to find themselves arguing in the direction of the same general concept, but differences in their understanding/interpretation will diverge more and more the farther you get into the discussion. It's so helpful to lay out terms at the start when possible.

Agreed - and this is why I find it _so_ frustrating when I try to take a step back (or down, or whichever direction seems most intuitive to you) and actually define terms, only to get met with a "you're just nitpicking about terminology, I'm trying to communicate ideas!"

If we aren't using the same terminology, we can't possibly communicate ideas.

Absolutely. I summarise it as "abstraction is the curse of the age".

My wife works in governance, and when a proposed new policy is being discussed, she is always the one to say "so, how would the policy work in this situation, with these people involved? [Outlines a situation.] Tell me the steps."

And, of course, the elegant, abstract policy is shown to be unworkable or lead to manifestly unfair or absurd outcomes.

I admire my wife's patience greatly.

I usually phrase it as arguing parties discussing a high-dimensional idea by projecting it down to a much lower dimensional space, and if they can't seem to agree, it means they're projecting to widely different bases - considering different aspects.

Visual analogy: take a cylinder floating in the middle of the room, and a single light source you move around. As you move the light around, it'll cast shadows of very different shapes. This is merely projecting a 3D object to 2D, but if all anyone in the argument has is a single snapshot of a shadow, you could see how a person seeing a circle might have trouble agreeing with a person seeing a rectangle - and yet they're still talking about the same thing.

(The biggest hurdle is to get people to realize their view on a topic is a low-dimensional projection of a high-dimensional thing, and not just the thing - and that to do something productive with it, one must be willing to "walk around" a bit, project to different bases.)

After much initial resistance, I've become a fan of pair programming. When things are clicking, ideas spark back and forth. Frames collide and interfere, bits of partial insight are exchanged, and you and your partner arrive at a solution that would have been impossible for either of you to get to working individually.

There is the famous psychological notion of flow. The thing is, this can happen among multiple people. Think of a lovely choir or orchestra. Or a basketball team where every team member just seems to be reading other team members' minds. The book (and presumably the upcoming movie) "The Boys in the Boat" really emphasizes this and crystallizes this.

As the linked article mentions, the author and his father butted heads for three hours because they were not listening to each other and pairing very effectively! Really good pairing does take work and practice. But when it starts to work there can be moments when it feels like magic. It is a great way to grapple with the myriad of unforeseen "details" that are actually life-or-death critical issues to project success.

I actually think it's far better to pair program with ChatGPT instead of another human.

The stairs in the article are probably not code compliant, even with all that work. The uppermost tread is cut short. Most locations require all treads to be the same depth, so his half-tread probably won't pass inspection.

The stringers either need to land a little further away from the wall, or the stairs the need to be at a slightly steeper angle (but check local building code for the max permitted angle!).

I built stairs with a friend years ago, and there are a lot of constraints to solve for if you want to build code-compliant stairs. A good parametric CAD system was very helpful. We could specify overall dimensions: stair rise/run, number of treads, thickness and width of the stock. It'd solve for the critical dimensions to verify, like tread rise, stair angle, tread depth, and the angles to cut the stringer ends to.

What software could a DIYer use for that?

I used (and still use, often) Fusion 360

Excel!

I think the main message is this here: "If you’re a programmer, you might think that the fiddliness of programming is a special feature of programming, but really it’s that everything is fiddly, but you only notice the fiddliness when you’re new, and in programming you do new things more often.".

I've done programming and construction and more than often construction is like programming, especially when you are renovating.

This is non-intuitive to me because programming languages (above the level of assembly, probably) are made by humans to represent human concepts, while construction is based on physics.

While writing that sentence I realized that programming does have a lot to do with logic, which is a construct of reality, and not humanity. So I guess it does makes sense that programming would be fiddly.

I am not a programmer.

Also, construction designs, materials, and tools are made by humans.

Programming languages can theoretically abstract things away to the extent a properly formulated phrase will get working code from an LLM. Construction designs and materials are still concretely limited by what physics will allow.

Well, programming languages are limited by what logic will allow, and computers (and by extension programming languages and LLMs) are also limited by what physics will allow.

Yes, but the compiler or LLM can theoretically handle that. In the construction case the humans are a large part of the compiler.

Humans are a large part of the compiler and LLM in programming too.

Human choices are, yeah. My comments on this thread ultimately derive from an intro to programming with Java course 10+ years ago that I got a B- in because I never submitted the final project worth 20% of the grade. I bit off way more than I could chew in the design of the project, but ultimately gave up because of a human design choice called "type erasure" that my limited programming skills couldn't work around.

I've always kind of seen programming above the level of electrical engineering as working in a social system designed with decisions that seemed like a good idea to a person who isn't me. This point was driven home a year ago when taking an R course and reading the literature discussing the grammar of R.

But yeah, in addition there are physical limits for programming as well. That's what just came as a mild insight for me here. I theoretically had been aware of it, but as an end-user / non-programmer just hadn't really put it together and thought about it before.

Details. :)

You keep using programming languages instead of programming as an analogy. Is that thought through or do you just throw these two concepts in the same basket?

It's thought through: https://news.ycombinator.com/item?id=38410024

But it's thought through as a person who is an amateur at best. I've taken literal introductory courses to programming in Fortran, Java, and R, have hacked together two or three VBscript programs, had a very slight one-day introduction to assembler in the Java course, and as a child was briefly in a summer program that had us "programming" a very simple robot (telling it to go certain lengths in certain directions). A former roommate said my pseudocode (when taking the Fortran course) looked like Pascal, and I used to watch him do some Forth programming.

So in terms of programming philosophy I am really, really clueless. But in terms of programming languages what I do have is a basic understanding of how incredibly diverse the languages are. And that a huge amount of this diversity is based on human preferences.

As an analogy to construction. Yes, there are a diverse number of techniques that can be used in constructions (yurts, hay-bail housing, etcetera). One could also draw an analogy to pre-fab construction as modular programming and libraries as kits (or whatever). But for the most part, outside of scale, it seems that most construction effectively standardized on a few "languages", if not a single language.

It's not just logic. Storage is in practice never infinite, everything has a speed limit. Those are real things, not some human concept.

I was thinking about the human-facing end. Not whatever constraints the compiler is under to translate the human-facing end to machine-operable code.

Well, programming is a lot easier if you don't have to deal with reality.

That tends to not happen so often.

One thing your initial reasoning left out is that human concepts are not simple, either. Money may seem simple until you start to consider currencies, rounding, taxes, fees, etc. Text may seem simple until you consider punctuation, different alphabets, math symbols, emoji, and other special symbols, glyphs, or characters. Names may seem simple until you consider suffixes, names that consist of multiple words, middle names, cultures that place the family name first, nicknames, name changes, various accents and other "special" characters (that probably are perfectly normal to the person with the name), etc. As a programmer you often have to deal with other programmers' solutions to handle all this complexity, which may be flawed or simply different.

I mention all this not to pick holes in your logic but because it sounded like you were trying to look more closely at your intuition.

> As a programmer you often have to deal with other programmers' solutions to handle all this complexity

That's entirely my point. Human concepts can be incredibly complicated, or just as bad based on ideas that are foreign to others. Some people like figuring out human complexity, but no one likes figuring out everyone's personal complexity.

Many disciplines are still more constrained by our limited understanding of, and ability to manipulate, physical constraints. The more abstract programming languages become, the more personally complex they become. Like money and verbal languages.

I think this distinction is artificial. Programming is made of and constrained by physics: the electric field in the transistors is a real physical phenomenon. While constructions are made of concepts; just ask an architect about "negative space".

And yet the humans doing the construction still have to measure angles in non-obvious manners regardless of the kind of staircase they are installing.

> in programming you do new things more often

Only if you don't do new things often outside of programming. Everything certainly is fiddly and anything new you do is going be technical if you care about doing it well. I spent half a day last week learning to paint-fill engraving to make labeling. Conceptually it was trivial (coat with paint, wipe off excess, sand back when dry) but there was still plenty of nuance.

> If you’re trying to do impossible things, this effect should chill you to your bones. It means you could be intellectually stuck right at this very moment, with the evidence right in front of your face and you just can’t see it.

It has always chilled me to my bones. Whenever I tried to do a slightly non trivial things, I am afraid I miss the super obvious easy solution and get stuck for years before getting it. And it already happened several times too! Really, the only solution I can come up with is "experience"... and a lot of asking from different perspective.

The interesting thing is that we somehow manage to cope with all this complexity, both as individuals, and as a society, and survive. When you look at all the details, and all the things that can go wrong, it looks like maybe it wouldn't be possible.

Also at a natural (non-man made) level, the complexity of biochemistry and the need to get 30 trillion or so cells working together as a single animal, in the presence of another 30 trillion or so single-celled organisms of random species, doesn't really seem like something that's going to succeed.

> cope with all this complexity, both as individuals, and as a society, and survive.

Surviving is not thriving. Like roman construction, we just have very big error margings and margin of safety. Things do not have to work perfectly to finction

> to get 30 trillion or so cells working together as a single animal

It's no coincidence then that your brain has a circuit to create backwards explanations for behavior that you didn't initiate, it helps create the illusion that you are a single animal, when in practice, you mostly aren't.

Let's build an API.

I need a CI/CD pipeline to deploy it.

I should use a static code analysis tool.

I need an observability tool!

Let's to security code reviews.

Do I have a SBOM?

Which MFA should I use?

Should I put it inside a container or go serverless?

Better generate a OpenAPI spec!

Do I have enough tests?

Oh, I need blue/green deployments!

Should I deploy to the edge?

Adding to the previous discussions on counter-steering bicycles: you can see it in tyre tracks. e.g. ride through a puddle; or, ride from rain to under cover.

Consider: how do you steer riding no-hands? Shifting your body weight makes the front wheel turn (because of the "rake"), but now your bodyweight is on the outer side of the turn! Hold it for a (brief!) moment, then lean the bicycle the other way and you have a nice turn in the opposite direction - which is what you wanted all along.

This is a critical skill on motorcycles, and most people learn it subconsciously, but real masters can throw the bike into turns in ways that make no sense unless you take countersteering into account. I'd love to be good enough someday to make physics seem confusing...

Motorcycle licence tests cover counter-steering, as an obstacle avoidance technique.

I think gyroscopic effects are also in play, and more pronounced on a motorcycle, with its wheels having greater mass and angular velocity.

Not in the US, or at least not in Tennessee. It's taught, but not tested on.

Absolutely. They're very unique vehicles, and I find them to be a fun mechanical exercise.

As a cyclist and motorcyclist, I've always found that bicycles will let you get away with not knowing you're countersteering, but on motorcycles you really need to be explicit about "push harder to turn tighter".

> The first problem you’ll encounter is that cutting your 2x12s to the right angle is a bit complicated because there’s no obvious way to trace the correct angles. You can either get creative (there is a way to trace it), or you can bust out your trig book and figure out how to calculate the angle and position of the cuts.

My first thought would be using a string: Fix a string to the floor at one end and to the wall with the other end, at the places where you want the upper edge of the 2x12 to meet the floor and wall (the exact locations are "input parameters" that depend on your design, i.e. on how steep you want the stairs to be - you can't trace or calculate them)

Then pull the string taut and now you can measure the exact length that is suspended in the air (which will be the length of the 2x12's upper edge) as well as the inner angles it has with the ground and wall (which will be the cutting angles from the upper edge towards the lower edge).

Of course even this is easier said than done. I'd assume that keeping the string taut and not hanging through might be a problem (catenary curve etc).

This was a brilliant read for me back in the day (still is!) and gave me the inspiration to write a post[0] about detail abstraction, market forces and why you can't fit big cups in coffee machines.

[0]: https://omarkama.li/blog/the-market-decides-your-project-sco...

No matter how closely you look, there is always more to it. However you think things are, good or bad, you are "wrong". But you may have sufficient precision to make some specific technique operate. (The technique is predicated on an implicit theory simplifying how things are. Both the technique and theory are "wrong".)

In finding correlations between two things, a tricky bit is finding which two things (a thing can be an arbitrary composites of other things). When we do find a correlation, that becomes how we see it, and we think in terms of this particular simplifying theory of reality.

i.e. there are two issues: (1) reality is complex. (2) we simplify it.

> However you think things are, good or bad, you are "wrong".

Does this principle apply to this sentence? ;)

Every time this is posted I want to rebut the water boiling portion. It's finally time.

The author gets the details of boiling water wrong. He commits a fairly common error: using his experience of a practical and common phenomena to make/guess a technical definition. The technical definition of boiling water is simply the temperature beyond which liquid water will not go (let's ignore super heating- too much detail!). The rest of the exposition is unnecessary and wrong given the true definition (although could probably be reworked from a cooking perspective)

That's fair to point out. However, I don't think it detracts from the point: the definition is in some sense a detail, because there is a connection between "bubbles forming" and "the temperature water will not rise beyond", but it's subtle and assuming that one equals the other leads to error. The very reason we have that definition is that it's more precise, because in the real world there's details that escape narrow definitions as seen in school or in the kitchen.

As an aside: I love the bubbling patterns of water as it heats, and it's actually a useful skill in the world of tea-making (especially in the gongfu tradition) to estimate the temperature of water based on the amount and size of bubbles. Nowadays I use a temperature-controlled kettle, but in the past I would get by with a clear borosilicate glass kettle and paying attention to the details, which is fun and rewarding in itself—if far less precise.

The way I read it, that’s why you’re supposed to imagine it’s the 1800s. All you have is the simple-sounding concept of “boiling”, and you’re trying to codify it by finding the temperature beyond which water will not go. At which point you start really paying attention and find the phenomenon is way more complex than your initial model of “water is either boiling or it isn’t”.

But it rather isn't. The temperature keeps going up as you add more fire. And then it stops going up even if you add more fire. That point is the temperature of interest and was from the very start. It isn't hard to find. It isn't hard to understand. It's just a bad example for this essay

> The technical definition of boiling water

is just the start. He starts with "water boils at 100 °C, right?" which I expect you'd agree with.

He then _continues_ to work out that, despite the simple definition, the everyday-life _experience_ of boiling water is nowhere near as simple. Why is it not so simple? Because, details.

Which is the point of the whole essay.

> gets the details of boiling water wrong

Can you give an example of a detail he got wrong?

> The rest of the exposition is unnecessary and wrong given the true definition

His _goal_ is to explore the details of the everyday-life experience of "boiling water" in the everyday sense of "boiling" and show how it's more complicated than what the simple (technical) definition would lead you to believe.

Part of the trouble is that the word "boiling" means two distinct but related things:

1. Heating liquid so much that all of it transitions to gas (your technical definition, restated to be a verb form -- "boiling water", a gerund phrase, cannot be a temperature, as you proposed to define it)

2. Heating liquid so that _some_ of it is transitioning to gas, while the rest of the liquid can be used for cooking

The second sense is what people actually mean when they say "I'm going to boil some water" or "Is the water boiling yet?"

If the discussion was about boiling pasta water I'd be on board (e.g. a discussion of the every day). But it's wearing the clothing of science and so falls flat (for me). None of the details turn out to matter for the real technical definition. Amusingly the one detail that does matter- air pressure and by proxy elevation- is omitted.

I have this comic on my wall at my desk.[1] It's about how much work goes into every detail of the everyday things we take for granted. I practice law, mostly patent litigation, and often my job entails untangling one of these stories.

[1]: https://xkcd.com/1741/

This was a beautiful read until I got to the part about disagreements. I think it may distract from the main message about the complexity of tasks and the value of experience.

That's the thing with details; sometimes you don't know the extent or context in which they will occur.

Your perceptiveness at a particular moment could determine a good or not-so-good outcome. However, you may not have the expertise or familiarity in a domain to spot micro, mini, or normal details.

My personal takeaway is to be mindful when you are on autopilot to avoid missing details and to avoid keeping your blinders on.

I agree that looking at the details really helps when reasoning about things. But that is hard as it is always easier to look at things from 10,000 feet

Except when it comes to driverless cars & so-called AI. That’s easier done than said, supposedly

To me, it's not so much that reality has a surprising amount of detail, it's that our brains make us feel like whatever model we have is complete. Same thing, of course, but focusing on the brain gets us closer to the cause.

>Frames are made out of the details that seem important to you.

Beautiful, and explains so much.

> In retrospect, I think I saw the fundamental difficulty in what we were doing and I don’t think he appreciated it (look at the stairs picture and see if you can figure it out)

What was it?

You can’t rest the ladder against the wall and trace where to cut. Because the wall will stop the ladder.

The biggest problem I see is that the part of the board you need to cut off intersects the ground. If you're leaning the board on the ground, you need to somehow trace a line parallel to the ground several inches above it. Hopefully you can solve that by using a spare chunk of 2x6 and tracing along that. Similar story for the vertical joint.

I don't know if that's what the author meant, though.

atan2() to the rescue.

Interestingly, as crude as the diagram is, you can feed it to ChatGPT's data analysis plugin and it will not only recognize what's depicted, but offer to calculate the cut angles for the 2x12s given the rise and run. I was pretty impressed that it recognized the problem without any prompting other than the image itself.

Trig was already mentioned. It's definitely not the answer to the question GP is asking.

There's a reason a huge part of technical architecture drawings is referred to as "detailing".

Feel all depends how one looks at things. For example, the vacuum, supposedly there should be nothing there. However, uncertainty principle allows fluctuations of all kinds which are not physically directly observable. If one looks at things at the level of principle, it is simple. However, if one tries to look into the actual fluctuations, there is infinite and unknown amount of complexity.

The link doesn't seem to work.

The link doesn't seem to work

Yes, because... Physics and you know, Math.

This reminds of Dunning Kruger Effect https://www.youtube.com/watch?v=4FGnb2lgPBA

people often tend to over estimate their capabilities and then only later once they start working on a project after some time realise that things are not that straightforward.

Seeing this here again, I was wondering if it's related to current events. I hope it isn't too soon, sorry to bother you in that case.

Can't you see? Let me explain.

The point of the post is that when you try to translate ideas to the physical world, there are plenty of quirks and you quickly find out what's the difference between theory and practice. That's why many people think that the chatting performance of an AI isn't a good measure of its ability to conquest the world.

And that's why some are so fed up with the doom prophecies, they seem disconnected from reality. But I believe the general opinion has swung too much in the other direction and maybe, just maybe, it's not wise to completely stop thinking about safety.

If this comment is the best LLMs can do then really we have nothing to worry about

"The first problem you’ll encounter is that cutting your 2x12s to the right angle is a bit complicated because there’s no obvious way to trace the correct angles."

Rubbish or as we say here: bollocks. Chippies have been doing first fit for centuries.

Its lovely that someone is discovering how wood and homes work. All power to you. Tomorrow - bricks?

Another surprising detail is parents are exempt from child labor laws. /s

Reply on news.ycombinator.com