Hacker News new | past | comments | ask | show | jobs | submit login
I hereby pardon all junior engineers (taylor.town)
93 points by surprisetalk on Nov 29, 2023 | hide | past | favorite | 50 comments



> It seems like nobody knows how to build things anymore.

The Chinese, Japanese, Koreans, Germans, etc might disagree. There's a lot of people out there who seem to be able to build very high quality things fairly quickly. It seems like it's the Americans that are struggling. We have for a long time, actually.

I would argue we just have a shitty building culture. We chase trends, cheap out on materials and salaries, overcharge for services, ignore maintenance, and shun process. We're way more interested in the shiny and shooting by the hip than the boring careful measurement. Growth is more important than quality to us. Greed and the new are our gods. HN is the pulpit. Laziness is the ladder on which we climb toward the heavens, ignorance our fuel. Watch this, I'm going to reinvent the wheel! We won't be able to roll on it but we can sell it for a bundle before it collapses.

It doesn't seem this way because we have so much money keeping us afloat. Like Google, the company renowned for its tech culture, but really makes most of its money selling eyeballs to Nike and Kraft.


It's ultimately about practice: Spain's first high speed train was massively over budget, and seen as a boondoggle. If building had stopped there, it would have been a curiosity, but the government kept asking for more track, and with practice come better results.

It's difficult to succeed in early projects of all kinds, but it's not just people that learn: organizations do too. Set up good incentives, and be OK with early failures. And if you lack expertise, but someone else has the knowledge, just hire a few experts away.

On most problems where America is very inefficient, it's either a matter of lack of practice or horrible incentives. If there's a cultural failing, it's unwillingness to deal with those bad incentives.


Especially for big projects, I often get the sense that failing a couple times is seen as unthinkable in the US: so a lot of ambitious initiatives lose funding or become laughingstocks because of publicized early failures. As a result, there’s a tendency to over-engineer in some areas and reduce ambitions in others.


It has been my experience that there are good and bad engineers from all origins; Americans do not have a monopoly on it in either direction, nor does anyone else.


If you can't generalize then you can't think. Of course any affirmation can be turned down on the basis that it is not specific enough.

OP is pointing at differences in cultures, they do exist.


It sounded like the person I was responding to just wanted a lazy poo-poo soundbite. There are many examples of all of the countries mentioned making ridiculous/awful products and also amazing products.


> If you can't generalize then you can't think

If you blindly follow others' generalizations without question, you can't think.


Compression is a form of intelligence


Sure, but blindly following something without question, isn't compression, even though both happen to involve reduction.


Sure if it isn't a losing material information in that process. It is easy to compress anything to true/false if you're willing to live with information loss. I think material information was lost in the aforementioned "Country X can't build anything" hot-take.


So people compress to the point where the speed vs information loss tradeoff makes instinctive sense to them.

You can dramatically increase your computing power if you're not constantly deriving everything from the fundamental physics and instead rely on some stereotypes.

This ranges from "objects that look like that are chairs" to "I can sit on chairs" to "Americans can/can't do X"


It's amusing that this text chain is about derivations when in reality we can reference a direct quote. We're not deriving any meaning here.

I'll leave it as an exercise to you to find that direct quote, as it'll be sure you actually read the comment :)


Short good


Good and bad in the sense that they can achieve management's directives? Yes, anyone from any culture so long as they have a practical grasp of the language they're working in can adapt to the circumstance.

But the culture starts at the top. Everything within the U.S. spiritual ecosystem is geared towards a superficial approach to all things. One example in the vein of engineering is Ford vs. Toyota. Anyone who has owned both of these can attest to the difference. There is a craftsmanship culture within Japan that I'm sure we're all acquainted with, but there are also much more subtle values at play (f.e. duty, sacrifice, etc.) that color everything an individual, and the society he makes up, touches. The individuals of America do not have strong values or sensibilities. I want to say this is because of its quality of being a "cultural melting pot." But this aspect of it doesn't turn it into a stone soup, where all the different and varied ingredients work together to make something unique, with surprisingly complimentary flavors. Instead, I think it turns into an actual crucible, where people of all different mettle are melted down; and all of the extraneous spiritual slag is scooped off the top, before being poured out as a completely uniform metal, without a noticeable difference batch-to-batch.

Instead of refining and building upon a culture, as 99% of other nations do, America specializes in annihilating it, and integrating the stray quality here and there that it finds useful towards its ends. And its ends is something I cannot even begin to know.


"Everything within the U.S. spiritual ecosystem is geared towards a superficial approach to all things"

This is hyperbole, I don't think there will be any way to convince you.


I agree that there isn’t much of a chance I can be convinced otherwise, having travelled and having not been born in the U.S.

I disagree that it is hyperbole. The main motive of this country seems to be exerting power over others. Whether that be through military, monetary, or quasi-cultural means: onto other nations and onto its own citizens. It’s very Prussian in its pursuit; and nothing that requires sensitivity can flourish here.


Your assertion of "Everything within the U.S. spiritual ecosystem is geared towards a superficial approach to all things" is hyperbole since you qualify it with "Everything" and "all things".


> spiritual ecosystem

what does this mean?


Hard to define if it doesn’t “click.”

I’ll try though: the collection of intangible qualities that interact with the human spirit, that are found within a certain environment. The qualities themselves interact with one another to create a sort of “bubble”/“soul”/“vibe”/“mood”/etc. For example, if you go into a loud open office with harsh fluorescent lights, no windows, and stagnant air you could say the environment drains the spirit. From there it will have an effect on your psyche, and cascade into you acting a certain way that reinforces that draining quality and adds a certain extra “flavor.” E.g. if the environment drains you, you might become short-tempered, unaltruistic, or mean-spirited. In that case your actions will reinforce the overall displeasing atmosphere of the office, but now you’ve also added the aforementioned knock-on qualities. If we take this further, then perhaps this would lead to a work environment where no great work is done. People go in, do the bare minimum and tolerate it, but the possibility of any higher qualities like craftsmanship, care, and conscientiousness have been rendered null — due to the collection of qualities just mentioned. And from there it continues to cascade.


Thanks for the explanation!


pg's essay on the subject is a must-read: http://www.paulgraham.com/usa.html


America has the most productive economy in the world by dollar. We make everything from rockets to disposable forks. Don’t confuse meeting a price point with a lack of ability.


The US for sure knows how to build things. SpaceX alone delivered over 80% of mass to orbit while China’s entire space program can’t keep pace with a single American company. We are also killing it at EVs, again, all the great might of the Chinese economy is playing catch up to Tesla.

The thing is, we know how to build things.

Those things aren’t crappy block apartments or faster trains. They are: space ships, artificial intelligence, robotics, cutting edge medicine, etc. The name of the game is individuality, entrepreneurship, and freedom (The ingredients that built the most powerful economy in the world). These are antithetical to both their political system and their culture.


"As a result, this country has one of the worst economies in the world. When it gets down to it -- talking trade balances here -- once we've brain-drained all our technology into other countries, once things have evened out, they're making cars in Bolivia and microwave ovens in Tadzhikistan and selling them here -- once our edge in natural resources has been made irrelevant by giant Hong Kong ships and dirigibles that can ship North Dakota all the way to New Zealand for a nickel -- once the Invisible Hand has taken all those historical inequities and smeared them out into a broad global layer of what a Pakistani brickmaker would consider to be prosperity -- y'know what? There's only four things we do better than anyone else

music

movies

microcode (software)

high-speed pizza delivery"

-- Neal Stephenson, Snowcrash, 1992


Most of what you mentioned is research & development, not building. I don't know why you say we're killing it at EVs, when there's many competitors, and Tesla churned out dreck for years until it finally produced a decent car without defects, and had to be floated constantly on US government subsidies. SpaceX is the only thing that seems to work reliably, probably due to sheer will from passionate engineers and a ton of QA (and a whole lot of NASA experience).

Anyone can come up with a new sandwich, given enough time and money. Not everyone can churn out 1 million sandwiches, quickly, perfectly, every time. That's what building is about.


Pst; all of tech is floated by government. It wants to keep a pipeline of engineers around.

IBM refused to risk its old tube business on transistors so Uncle Sam paid for it, gifted tech to IBM.

Internet was a government defense project.

Capitalism is just people. The euphemism is getting too deep. We need waders.


I'd love to try the software products from these countries, but I can't deny that some of the very things spoken of in the article came from the origins of which you speak.

Samsung made Tizen, and many of the 'TVs with eyes and ears' out there.

BMW was apparently the flagship for heated seats as a service.

To be fair, Nintendo is pretty good at innovation in both hardware and software, but not so much at cutting edge power these days. Or, like, people trying to have fun with friends (whether online or in tournaments).

I can't speak for China on your list; would love to hear of some battle tested counterexamples though.

All of that said, greed is undoubtedly a leading cause for many of these industry wide ailments. Why innovate when you can coast? Why invent something in your spare time for nothing when someone can pay you good money to transform JSON?


Nintendo purposely doesn't focus on cutting-edge or power. Hasn't since the the mid-80s. Having fun with friends is a core concept of what they do. Hand a Switch to a group of people to play party games. It's really interactive and social.


Nintendo is good at social if all the players are in the same room: their internet services leave a lot to be desired.


China's Five-hundred-meter Aperture Spherical Telescope (FAST) is a nifty bit of now proven novel engineering with active shaping and nimble receiver positioning.

https://en.wikipedia.org/wiki/Five-hundred-meter_Aperture_Sp...


Ableton comes from Germany for instance - not my cup of tea personally but by all accounts great software.

Samsung meanwhile has a significant software development presence in California.


>> It seems like nobody knows how to build things anymore.

>The Chinese, Japanese, Koreans, Germans, etc might disagree.

It's funny, after riding the Chinese HSR network, this quote from the article had me shaking my head:

"California is spectacularly behind schedule and overbudget on its quest to build the world’s slowest high-speed rail.

It seems like nobody knows how to build things anymore."


Well China may have a HSR. But it's sort of bankrupt on arrival outside of the main metro's. China has the luck in this scenario of being free of oversight, and able to create it's own money. Most western Infrastructure isn't built in a vacuum that allows it to arrive as a gold plated entity.


I connected with this, I've had many of the same thoughts lately. It's a thoughtful essay and I have no idea why it's been flagged?

> When fewer than 5,000 programmers existed, they built Lisp (1960) and APL (1966) and sketchpad (1963). With over 5 million programmers today, the best we can do is Rust and React.

It feels like this has worsened even just in the last 15 years. When I started in the industry (2008ish) it felt like there was a lot of experimentation and new ideas compared to today. Memes like "language wars" and "javascript framework of the week" come from the recent past when there was not the same monoculture of like 3 big JS frameworks, 5 big languages, etc


> It feels like this has worsened even just in the last 15 years.

It is a question of Statistics/Probability and similar to statements like "The End of Science"(see John Morgan's book) i.e. all the main/major aspects of knowledge are already known. IMO the following are the reasons;

1) In the early days the field was populated with experts and the quality of the output was high. There was also a lot of unknowns to be mapped out and hence the quality and importance of inventions/discoveries were startling.

2) Once you have the fundamentals of a Scientific Theory with Principles and Guidelines laid out for a particular field almost all future developments in applications can be anticipated and will seem like "nothing new". Our very advances in Mathematics and Sciences negate future novelty. One good example is the current advances in AI/ML which is merely a hodge-podge of Statistical techniques/Computation models and lots of power.

3) With the rise of the Internet/other forms of content delivery, a very large number of non-experts now have access to a bullhorn to spread noise and amplify it in very many dimensions in the service of vested interests like business/power/money/etc. Hence the s/n ratio is very much diluted i.e. there is now a lot of junk to wade through.

4) Finally; Social mores, norms and values are very different today and hence discovering/inventing Knowledge has been made subservient to other goals/needs.


Errata/Addendum to above:

It is John Horgan (and not Morgan) and the book is "The End Of Science: Facing The Limits Of Knowledge In The Twilight Of The Scientific Age".


It has a distinctively LinkedIn style of writing, e.g. many one-sentence paragraphs. Could be why.


> How fast should programs boot? How long should it take to make bug-free software? How long should a smartphone stay relevant? How much mining should it take to live? How many steps should it take to get to my local grocery store?

None of this is really relevant. None of these questions 'need' answers.

What do the customers need? Where is the market for my goods/services? These questions are relevant.

People know and love to build things. The things they build and invent are then taken over by middle managers, project managers, high level executives, and politicians.

It's not the builders or makers that can't build things, it's the budget constraints, board-room death spirals, and project management to death that makes projects hot-garbage.

We can't build things because they won't let us. We show the powers that be a 7 course meal, and they make it a McRibb. Sure it's tasty, but it sure ain't what we wanted to build...


Taylor was one of my favorite people at strange loop this year. His outfits were fantastic. I love his blog even more. And he's a courageous and deep thinker.


Aww what kind words! Come stay at our place if you're ever in the Southern California area :)

If you want to copy my outfits, I find all of my strange clothes at Anime Expo every July! I just purchased an egg sweater from a vendor I found there:

https://shop.kinwamonster.com/products/cardigan-egg


That would be great! I hope to do that someday.

I'm trying to take my daughters to Japan in March. Let me know if you need anything direct from the source!


I don’t blame the junior engineers either. I blame their seniors and managers.

Leadership is especially awful these days and across most industries. Leaders tend to be too far removed from object level reality and just sit around blowing smoke up buttholes all day long. If you got laid off in the last year or two it’s because your leaders failed you with their shitty cargo cult approach to running a business. “Der, Google is hiring like crazy and dey lots of money and I want big big boat… so what if we hire like crazy and make radios out of coconuts?”

It’s easy to throw shade but when I was last at the Facebook HQ I saw Mark Zuckerberg down on the floor talking to a young designer. He also took personal responsibility for the slimy layer of middle management that emerged over the decade prior, the result of which contributed heavily to the need for layoffs in the first place.


> I don’t blame the junior engineers either. I blame their seniors and managers. Leadership is especially awful these days and across most industries. Leaders tend to be too far removed from object level reality and just sit around blowing smoke up buttholes all day long.

100% agree with this. "Modern Management" as practiced today is a major factor in the decline of quality creativity/research/engineering.


I liked this post, especially this quote:

> The norm is to battle software rot and add features ad infinitum.

Software engineering is so expansive, it is ideas built on ideas built on ideas. I learned Node.js in my free time because of the derision that is rampant for FE engineers. Node.js is built on C++, which compiles down to machine code which runs on a CPU that has it's own firmware. My Node.js application sits in a docker container on a Linux server written in C that somehow knows how to accept network requests and respond with data to a user's browser, using protocols and methods that were conceived of before I was born. The application got on the server because of my gitlab pipeline that first built my application and bundled it with Webpack, then ran unit tests with Jest, then ran e2e tests with selenium, then runs a bash script to git clone my code to build a docker image. This all to change a response header from a 200 to a 203 to align with best practices.

At any given day I am working on 30 levels of abstraction, trying to deal with elitists who were around when the internet was in its' "glory days" telling me now that the stuff I am working on means nothing when compared to the pristine web we had in 1987 when a gif of a dancing baby blew people's minds. Now the internet is layers upon layers of protocols and languages and code and tracking that if I take time to look back on and learn, the industry will have moved past me to Next or Nuxt or Nyxt or Naxt or whatever bullshit flavor of the month library I am required to master and have 3 years of experience in (it came out a month ago).

There is no extra space for regular people like me, we can't build things any more; at least not things of note, things that get used. And while building something that never gets used is great for your portfolio, it feels like making a painting and immediately sticking it in a safe. What is the value if no one even sees it?

> When fewer than 5,000 programmers existed, they built Lisp (1960) and APL (1966) and sketchpad (1963). With over 5 million programmers today, the best we can do is Rust and React.

Being a standout of 5000 people is extremely hard, being a standout in 5 million is a statistical anomaly even with specialization. And people do it in the same sense that people win the lottery: someone has to but your chance may as well be zero[1].

1. https://en.wikipedia.org/wiki/Lottery_paradox


This article touches on A LOT. I would say the focus is on why software projects are so large and complex. I think about Fallout 4, a remarkably buggy and inefficient piece of software; I think about the React + Redux web app that a previous employer sank millions of dollars into to succeed a perfectly functional, perfectly adequate C# MVC WebForms app; I think about how it seems today’s software engineers are valued not by what they built but by what they know.

1. Fallout 4, a game released by the notorious (for bugs) Bethesda Softworks in 2015. Despite the ample bugs, some of which were ignored by Bethesda and ultimately fixed by an unofficial, third-party patch, it won Game of the Year in 2016 from BAFTA. In 2015 I lacked the hardware required to run this game above 30 FPS. Today, I have a $2,000 monitor and a $1,000 graphics card, buuuuut Fallout 4 doesn’t recognize newer graphics cards and defaults the quality to “Low” with the lowest resolution possible. Additionally, it doesn’t support ultrawide monitors as many games these days do. Finally, it crashes every couple of minutes. I’m not sure why. It’s not a mod causing it. I’m using a normal aspect-ratio monitor now. So this game, not even 10 years old, has been abandoned by Bethesda. Like Fallout 3 before it, I’ll probably have to install a litany of unofficial patches and make edits to config files to get this game to run the way it should out of the box.

2. SS&C’s, Advent Software’s, Black Diamond’s “BD3.” The amount of effort that went into the UI was, to me, obscene. Probably 40 devs working on this. We had a Storybook as a style guide. We had internal tools (also in React and using Redux for… some reason). The web app looked beautiful, BUT always had bugs. Always. An area of the site would be enhanced, and then there’d be some bugs. I once spent about a week working through a bug that only happened on mobile Safari, and the CSS fix was ultimately hacky. To debug this, I had to have a Mac connected to an iPad, and use Safari’s debug tools to connect to the Safari on the iPad. I remember it didn’t “just work”. All this effort to make a web app whose primary purpose was to look good. I think the main concern was keeping up with the competition. Evidently custodians like BlackRock bounce around wealth management platforms and we were directed to cater to every whim of the large clients. If they wanted some bullshit little toggle on the benchmarking page to show grid lines, one of our dev teams would inevitably spend hundreds of man-hours making it happen. If they didn’t, we might lose BlackRock, which would be worse than spending $100,000 on adding grid lines to a chart.

3. I have actual, useful code on Github. I have a catalog of Medium articles I’ve written on subjects like ETL, AWS and Azure. My StackOverflow is in the top 5% this year for rep received. All of that means fuck all to any employer if they are looking for someone who is familiar with DataBricks and I’ve only used Spark and AWS Glue (for example). I get it, most people’s Github profiles are in a sad state… but wouldn’t you rather hire someone who clearly knows what they’re doing, than someone who say solves your Leetcode in O(log(N)) time?


Instead of complaining about others why don’t you build something great yourself?


Author here!

Totally agree. I'm secretly off in the wilderness trying to build something I think will be great:

[1] https://scrapscript.org

In the meantime, I thought some concrete complaints could inspire people to work together :)


Some nice things in the language, well done! I used content addressable code snippets stored in IPFS for my own system too. Content addressable code seemed to start with Unison, but now a lot of new programming languages are seeing the benefits of this technique!


Looks great, any sample app out there?


Nothing yet! Some folks from the community just stepped in to help me and we're now making incredible progress together. At this pace, I'm expecting a very rough first release in January.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: