Man I feel this. Every time I used to express interest in learning C or ASM, I'd get responses that ranged from "why?" to "don't," and of course, "learn python instead."
I really miss those days when America was BOLD enough to take on CRAZY challenges like the Moon landings! Sure, America was going through a VERY turbulent social crisis at the time, but they did it! Now? Now everything is lazy and outsourced to more capable countries. And, even though I can't stand Trump, he does have a point on how America has stopped innovating, and would rather indulge in mindless consumption... Well... Things ARE changing now! Post COVID... Elon, and others, are taking bold strides forward again, against a backdrop of a society that is going quietly mad... So much so, that even gender itself is becoming 'indeterminant' and the youth, drugged up on TikTok! It would appear History is repeating itself once again... But this time... The stakes are MUCH higher!
Because they don't know it themselves. And instead of admitting they don't know anything, they'll tell you to go down the same route as they did, validating themselves and the choices they made in life.
Yea. The fact that these were inflicted on us by the hostile, racially alien, power-holders of the day, certainly outs the bronze age collapse in perspective too. One wonders what really happened then.
yeah, within three years the globalism order of trade that complex supply chains depend on is collapsing on one hand, the demise of China was in the cards this decade due their demographic catastrophe. But now their never-ending Covid lock downs and the cancelation of Russian exports from the global market are acting like pouring gasoline on a smoldering fire, rapidly accelerating the collapse.
Its funny because it says there are "2 replies" to your comment right here, but i can't open the replies and see the replies to your comment so now somehow even the replies in youtube are bugged and dont work i refreshed the page, still the same issue, hoping that now im commenting on it the problem somehow gets fixed and i can read the replies, jesus christ
I'm learning about how to work in an Agile environment, and now I can say it's part of the problem. The whole idea of sprints, constant change and never being finished is really toxic to writing good software. It's about adding new features as fast as possible without considering the quality of the code, and when things get ugly people just jump ship, and it is up to someone else with no idea about the code to fix it while continue adding more bloat.
yeah, very much frowned on when an item needs to continue being worked on in the next sprint because quality-wise it's not really where it needs to be. My company is going to shorter sprints, as if this is going to improve anything. Certainly it will make it harder to do complex features that have to ultimately span multiple sprints - but the agile stats point an accusing finger at teams that aren't getting items fully completed in a single sprint. So the culture of agile is indeed a toxic one for software developers that want to be doing quality work.
I just watched a talk about Agile by one of the original signers of the "Manifest for Agile Software Development", Dave Thomas. In short, the whole concept of Agile as we know it shouldn't exist, because it was never thought as a methodology but as guidelines to add flexibility to software development. It was corporate who came in and made it into a thing on itself and gave it the name "Agile", making it into a rigid way to develop software, defeating the whole purpose of the manifest.
@@dienand_ A piece of software dies when no one uses it, not when it is done. And yes, you can finish software, you don't need to continue adding features for the sake of doing it in the hopes of covering every use case, that just make it more bloated, increases the possibility of bugs and just makes things worse in general, and is a terrible way to develop software.
Hmm. "Maybe there was some environmental stressor to kick it off, maybe some huge drought, possibly also some floods". "People attacking people" Feels some the impacts are getting closer
@@Martinit0 I think that we are pretty resilient to those types of stuff today luckily. IMO the scariest thing so far is the EU energy crisis. That really will have MASSIVE effects on literally all sectors and might cause some important sectors to become entirely unprofitable until it's fixed. Scary.
This is weird, because when I was a child I remember software crashing way more than it does now. And everything was harder to configure and use. But I'm a software engineer now, and I also kind of feel like everything is more complex than it needed to be.
I think there are a lot less hard crashes than there used to be, but more logic bugs and weird unexpected behavior from all the layers upon layers of dependencies that everything has.
We're going to need teams of people to write hello world if we keep up with this complexity orgy. 😔 There's so many moving parts now...I find it incredible anything gets done. 😏
You're right. Computers have never been faster, more stable, or capable. Blow's a clown. You couldn't even play a game while burning a CD in the 90s, or your computer would hang.
Another parallel is in the shipbuilding industry. In 2000 Swan Hunter cot a contract to build two Bay class landing ships for the Royal Fleet Auxiliary. The yard hadn't built a ship in years, so a lot of the knowledge of how to build these ships had been lost. Costs went through the roof, the yard went bust and the second vessel was transferred to BAe systems at Govan for completion.
@@LuccDev the Olkiluoto III nuclear reactor facility in Finland is like the 8th most expensive building in the world. It was supposed to get finished in 2008 IIRC, but it got done just before 2023. There were many reasons but the only one I thought was valid was the disappeared industry skill and knowledge.
@@LuccDevspare a thought for the UK nuclear industry, which was not just allowed to die but was actively killed. Imagine going from being a pioneer of a technology to needing to pay the Chinese government to do it for you.
Another great example of this would be the resurrection of AMD's GPU division. Their new GPUs (from 2015 onward) are amazing on the hardware spectrum, they're more efficient, fast, etc.. But the new "Radeon ReLive" software driver thing is so slow, buggy, unusable and unstable that I just don't get it. Really magnifies the whole "software is piggybacking off of hardware" thing he talked about in the beginning.
Software has DEFINITELY been free riding on hardware. Because memory allocation and garbage collection and resource management are tedious and people don't want to do it, we just use languages that don't require you to think about that stuff. Our databases don't necessarily need to be well designed because we can compute millions of lines in a matter of moments. We've been spoiled with crazy computing power.
I don't know much about the software thing, but this was really amazing. I think we can relate this sort of "slow-fall" problem to other things too. We're slowly losing active thinkers... By the way, I think The Witness was the perfect game to describe science and education. Thank you so much for giving us that game!
We're not losing active thinkers. It's just that all of those active thinkers are being wasted on trying to understand their 30 millionth Javascript library rather than actually trying to solve real problems.
You're right about active thinkers decline (in the broad sense, not just computer space). The worse being it's not a bug, it's a "feature". Globalists certainly don't need much active thinkers, besides the very top dogs which are a handful, that's why for decades we can witness dumb decision after dumb decision which many common people had predicted the outcome. Tech has played its part into this partition, which is an inevitable side effect if not actively fought : humans being lazy, automation indeed makes lot of people losing/not learning useful basic skills (which often can have an impact on other skills too). AI... While I have a hard time believing it will get sentient, I can easily picture people relying on it instead of using their own skills, which will make the AI (relatively speaking) look to many as "intelligent", "magic", which is a vicious circle... one more toward an obvious downfall. I think for anyone that counter intuitively, it's the best time for a student to build his lower level tech skills, whether it be for high tech or not, the chain of knowledge that got us there need to be maintained, and exercised. Those who will learn low level high tech or low tech won't have any problem finding work in few years, once the AI bubble will have popped, IMO. Even before that by proposing vastly more efficient solutions resource wise.
Like many things, there is a Star Trek episode about this concept. STTNG: When the Bough Breaks. Dr. Crusher says "Your people have forgotten how everything works"
38:20 is so interesting & I'd never even considered it. It really makes me want to go try out templeos where everything is in ring0 so I can see what it's like in crazy land
I feel very heavily for the question at 59:53 I feel like soo many people are on board with this concept of "Let's learn how to do things the 'right way', let's write better code!" but it's extremely difficult to make that shift. It'll be years before you make anything potentially worthwhile, even rendering pixels on a screen. I get where Jonathan Blow is coming from but it sucks that there's no real solution even attempted here.
Yes, stuff was crashing all the time necessitating hard reboots and / or running out of memory. Maybe the stuff banks were running was more reliable but consumer software was buggy even then.
but it also doesn't make sense either. Because no one wants a fragile operating system or piece of software. That's also bad for business. Therefore there is usually a balance, and because companies *really need a stable system* when running their billion dollar applications, stability is actually really highly valued and prized and therefore has been achieved. How long do servers run? Basically non-stop, under heavy loads. That's really impressive - both software and hardware parts. His laundry list of problems with software is just one long anecdote. It is not a serious statistical analysis...so yup. Hard to take this really seriously
I work in software development and given how stingy programmers are with helping others or passing on experience, the loss of generational knowledge does not surprise me.
I would say it has been "riding on hardware" for decades. However, hardware has started to plateau. And it's not just with computers that you "can't just" - it's pretty much everything has gotten gratuitously complex.
Even with the capability to do ray tracing, doing it is going to be slower than not doing it. Its not something that needs to be universally applied everywhere.
@@josh1234567892 Taiwan is an independent country. The US isn't going to go to the lengths it did to minimize its interaction with Taiwan like it did for Ukraine. Taiwan is too strategically important because of its place in the CPU market.
Putting my few cents here about learning, people generations and knowledge flow. When I started learning to code there were things in older programmers that they know something because technology or some idea was known and iterated over years while they were learning/working in industry. But when I was just starting to learn to code I was seeing technologies in their most recent shape. I had lack of this years of background and iterations over technologies and all this "why's?" behind changes in all IT field. And personally I disagree with people saying that "IT (at least coding) is easier and easier to jump into" I think that it's reverse because of lack of knowledge flow I mentioned before but maybe I'm just not seeing this bright side they mention?
I think a lot of what was said here about loss of knowledge is true, but the problem is twofold, and i dont think he addresses these whatsoever. 1. Economics. We live in a capitalist system that is heavily reliant on cheap labor and cheap products. Until labor standards increase for developing countries, that will continue. Companies are focused on easy profits, not reducing complexity of systems, and they are also focused on monopolizing the markets they are in so someone with a less complex system could ever overtake them. But I doubt Blow wants to go there because he has rather conservative politics. He asks a lot of questions in this talk and never answers them. Why is the market like this? Well, Blow, it's like this because people like you think capitalism is a good system. 2. Software engineers dont know how to communicate. Its incredible to me just how bad my coworkers are at basic english, grammar, and communication, and how hard it is to find someone who actually wants to teach as part of their job. They dont know how or dont want to write documentation. Its not really even the lacking of social skills that is stereotypical of software engineers that im talking about here. You can be socially inept but be good at documenting things. Code comments nowadays are frowned upon by many of my recent clients, instead wanting "self documenting code". No one writes details explanations and absolutely no one wants to read. I don't necessarily blame engineers themselves for this, this problem is largely a byproduct of point 1, the economics of it, but I also think it has to do with the culture of software engineering as well. Software engineers are just trying to get out of writing anything BUT code, and then wonder why our code is so hard to maintain, build upon, and pass down to others. We don't speak to each other in code, we speak in English, and that is for a good reason. It should be mandatory for any CS degree to have gone through some writing and communications classes, like public speaking, debate, technical writing or writing in general. And I say this as someone who has struggled their entire life with social anxiety. Blow is no exception here. Despite being known primarily for his outspokenness on a range of topics, he makes no attempt to provide a solution here that is workable under our systems, and he certainly doesn't advocate for changing those systems.
I know engineers who insist on taking what they know to their grave for different reasons... Off of the top of my head, they are: 1) they don't feel it's their problem/job to teach, 2) they don't want more competition in their field, and 3) they're being forced to train their lower-paid replacements who are often foreign. For reasons #1 & #2, I disagree. For reason #3, I 3/4 agree... let's see if that's allowed.
@@lordlucan529 I think this plays a larger role than most people want to admit. The general consensus (not just with younger programmers!) seems to be, "Well, why should we bother learning all that stuff? That's just a waste of time, we can just use such and such program, or we can use this game engine, or we can eventually rely on AI to do all this stuff." Very short-sighted and naive thinking, in my opinion.
Very inspiring talk. The point that "programmers aren't as productive as they used to be" is interesting to me. I'm self-taught of 3 years, with 2 years of professional experience, with almost all of my experience in front end development. The thing is, I want to be one of those engineers that "gets it", that understands the foundational ideas and tools of our craft. But I have no idea what to do to get this type of knowledge that he's talking about. Is a computer science curriculum a good start? Then what? What resources would any of you recommend for someone who is interested in learning how to do things the right way? What sorts of things can I do?
I would suggest checking out the Handmade community. They're devoted to reviving the "old ways" of programming. The roll your sleeves up and get your hands dirty kinda stuff. I've learned a lot from Casey Muratori. He's making a game from scratch in C. In what amounts to a day of work, he had a code hot swapping system working where he could recompile the game logic and see the results on screen without restarting the game. It's deceptively less complicated than you'd think to do things more or less from scratch and it's an incredible learning experience.
@@chordogg That code hot-swap thing is incredible. And it's like 3 hours total of video time with him explaining every line of code. I'm sure someone could implement that system in an hour now.
Find people who believe in this concept. Find people who aren't just smart, but are geniuses. Study their code. Decide if their reasons are valid. I know people will throw rotten fruit at me for saying this, but the Linux Kernel is a great place to learn some things. There are other even simpler (not that Linux is simple) OSes to look at like ucos, and the bsd guys are pretty serious about KISS as well. Have a look at this list, pick one and port it to one of the many SOCs. en.wikipedia.org/wiki/List_of_operating_systems#Other_embedded. The arduino hardware is great because it is cheap. Whatever you do, do not use the Arduino software, it is a perfect example of What Not To Do, then giving you an API that has almost no relationship to the hardware. C is key. I am sure there is a kernel not written in C somewhere, but I've never seen it. C is good on hardware because you generally know exactly what is happening. I've written applications with a 512 Byte (less than 1K) stack using C. It was a BLDC motor controller that used SPI commands to set acceleration, ramp up/down, distance, kept track of the optical encoder and ran all 3 phases with their hall effects. It also had a serial debug terminal. It all fit in a 4K flash on some 28 pin AVR part. You can't use printf - you have to write your own. You can't even use a C library because it's too big. You want to learn something, do something like that and you will learn a lot about simplicity, hardware, and the cost of bloat.
51:30 "Almost nobody knows what is happening on a CPU" - I posit there is another area where it's even more critical: There are only two companies worldwide that know how to manufacture 5nm node silicon chips (TSMC and Samsung) and there is one ONE company (ASML) that can manufacture the EUV-lithography machines that is needed for this manufacturing process - AND - ASML has each only a sole-source supplier for the optical assembly and the driving laser, both critical components of said machine. The amount of knowledge that goes into making that EUV machine and into making the 5nm manufacturing process work is insane. Insane! And it's heavily guarded knowledge. Who know whether there even exists a person that knows it all.
I know I'm late to this, but it almost sounds like *gasp* knowledge not being shared is to the detriment of LITERALLY EVERYONE. These companies may make their money right now by doing this, but sooner or later they won't be around and one HAS to hope that the knowledge that is necessary to build these instruments is still available to those who may need it. The whole idea of intellectual property is absolutely baffling to me but the corporate machines cogs just keep turning, business as usual. (I'll add that IP definitely makes sense in the current system of economy, it just doesn't in a humane one) I hate to sound like a anarchocommiewhatever, but huge companies carving out their niche in such important parts of our industry is so dangerous, not only to the industry itself but to society as a whole. The same problems WILL emerge in the software/hardware industry as they do in every other industry where the profit incentive gets in the way of the greater good like we've seen with housing and finance which don't even have the slightest veneer of passion, just cynicism. Not angry at you, just a man shouting at clouds here. You make a very valid point.
@@styrofoamsoldier You sound like you could go either way, anarchocommie or anarchocapitalist. The only thing that leans toward "commie" is the word "shared." Personally I don't think knowledge should be freely shared, but it definitely shouldn't be locked down by feds with guns! If some kid makes something in his garage and wants to sell it, let society figure out whether they want to support that kid or the mega corporation that would otherwise sue his ass into the stone age and ruin his life for his mistake of not knowing someone else made something similar.
This applies to the universal hate for Delphi. Unlike most other advance and capable languages, Delphi doesn't need a million APIs to work. It just works.
You need root access to fix society's kernel bugs. If you do that, you don't have to build an empire of workarounds. Everything flows from the source. Societal change starts within each individual. I.e. spiritual wisdom teachings and cultivation practices are crucially important.
Tried GitHub Codespaces, but it was really buggy, so I switched to TextMate, which was worse, so I tried NeoVim, but it had graphical glitches everywhere, so tried VSCode, which runs like dogshit. It's incredible.
Unnecessary interruptions occur due to using tools tile Unity and Unreal but I think the time it would take to write everything from scratch is such huge overhead. If it takes solo devs up to a year to make a game I feel like it would take 5 years to get your engine up and running with rendering, input, lighting, sound, etc...
I would imagine that if more people were working with the lower level APIs then there would be more pressure to simplify and standardize them. Like the shader language problem Jon mentioned. So it would be easier if there wasn't such a reliance on the 'big engines' to deal with these problems for you.
True, but an engine made by a studio won't have the scope Unity or Unreal has because isn't a product that you wanna sell to most people, is a tool you will use for a specific purpose. Remember that "back in the day" videogames used to be developed in months, not years, i think doom or quake was developed in like 8 months or something like that. I'm pretty sure that was because they knew the ins and outs of the engine they were using and most important the engine was made to develop games, not to create movies and many other things.
Anyone ever tried video editing? Unknown coded, can't read file format, duplicate frames, audio desynced, video length incorrect, video file corrupt, video freezes for 2 seconds at the beginning... And most of these bugs happen for no apparent reason.
I agree with a lot of it, but not the part about Language Server Protocol. old editors WERE NOT able to do what new editors can do (variable renames, variable usage lookups, other structural refactors), and for the longest time, only SOME editors could do SOME language-related analysis for the programmer. the point of LSP is that ALL editors can do ALL the analysis. the only part of his criticism that is reasonable is that LSP should be a library, not a server. it's a fair point, but there are obvious reasons why a server was chosen instead of a library (the server needs / is better off holding some internal state).
>What happens nobody can get to stackoverflow and copy paste thier code Well, we will be buying monthly snapshots of whole stackoverflow on a flashdrive from a local dealer
I work in IT security, and I know exactly what this guy's talking about, however, when the programmers started asking question I got the impression they had no clue what this guy just said..... What the fuck is going on here. You can tell they don't understand because they are asking the wrong questions. At this point I'm not confident people will be able to read what I just typed... The comment section of this video is horrible.
Gotta take into account that this was in Russia and there is a language barrier in place. This is a very hard phenomenon to reason about (like Jonathan said, it's hard to see it from the inside so everything is conjecture) and having to do it in your second (or third) language that you might not speak daily is a hard ask. Especially if they didn't really know what talk they were getting before hand.
No one asked following questions: 1. If we would have ability to copy a program as a single file to another program so that would work fine, then how we should manage our dynamically linking libraries? I mean that one program may depend on plenty of shared code (libraries) that other programs may possibly depend on too. You need to invent a new concept of dynamic libraries or some kind of enitity (i.e. runtime chosen components we can't predict). DLLs are using a simple concept but it forces us to make some kind of layer in the OS that will "install" those for us to work. If we get rid of the concept of "installation/deployment/whatever" we have to keep in mind that we want a flexibility (like dynamically linked plugins in runtime) and keep space on disk (i.e. not take all those libraries in each binary). I don't know how to expain well what i mean, maybe someone understands me 2. If we get rid of complexities in the OSes then how we can pretend those OSes would be secure? I mean that if the OS is capable of writing "raw" pixels on screen then how we can ensure that some software would not abuse that feature to disrupt our user experience (for example a software that fills entire screen with a color and disturb us from killing that process using the system capabilities). You can't protect your systems from those kinds of abusing with other layers of the OS. I also think that security features are adding complexities in the applications we write and this is inevitable Good talk by the way, I have some kinds of inconveniences that exist nowadays in software like inablity to send a file from iOs to Android over the air without internet and cloud storages or messengers, however, both of them have bluetooth that meant to be working the same across all devices
40:00 yeah, that dream of "worked everywhere" just was never true at all. You even had to have separate settings and drivers for every sound card even. (and the list changed for every game!) Also in the past consoles were not x86. So growing up with DOS & Atari/NES for consoles... I'm not sure where this came from. And Boy, did we have a lot of hard crashes and things like IRQ and out-of-memory problems. Though otherwise there are valid points there.
It was a completely asinine argument. You can trivially "just draw pixels to the screen." Obtuse toads like him pretend that things like SDL2 don't exist because they don't meet his ignorant and arbitrary concept of what software should be. I feel sorry for rubes that are tricked by transparent charlatans like blow. "You can't just run an unsigned program." Yeah no kidding. Just like in the 80s, you couldn't just code an 8bit NES game, burn it to some ROMs, and plug it into a cartridge pcb you had lying around because it was all proprietary junk you couldn't even acquire. Software has never been faster, more reliable, more accessible, or more stable in human history. This video is complete garbage. Literally everything in it is wrong.
Great talk, I hope we find a way to get back to pushing systems to their limits while also pushing to make software more user friendly and approachable to new talent!
3:21 "we went from nothing to all that stuff in something like 12 years"... gee... The Third Reich wasn't "at nothing" in 1939..... sounds to me like Operation Paperclip took that long to gear up what was already in motion in The Third Reich... need i explain the detrimental costs of lost inertia in engineering projects or manufacturing? jeez, we're at a stage now where people are delusional enough to think the "everybody helped create all technology that we deem valuable today, no matter which arbitrary thing becomes the popular ~~~~social~~~~ topic". I agree with most things he's saying in this speech, and they should be addressed, but we need to be honest about the history of this stuff; the first programmable computers were made in The Third Reich and the transistor magically appeared at Bell Labs, with no paper trail as to how it was researched, 2 years after Operation Paperclip. People who claim to love high technology are frauds if they're not examining these things. 4:08 the trajectory went downwards because the USA started doing what the Politburo did in the Soviet Union did, and used it as a paycheck generator. It also makes sense that there was a downward trend when you factor in the vulture "scientists" like Einstein and Edward Teller who seemed to siphon off research from The Third Reich and brand it as their own, and both of these "scientists" are written about as if "oh this one man research lab came out of nowhere and showed us all how it was done!" Lies. The concept of why these projects go down hill here is akin to a flying saucer landing, us learning to ride it around, but never being able to reproduce it. It gets used up and then we never again can use those capabilities. 6:05 elon musk is an unelected and extralegal US president whose "businesses" have never turned a profit. he's purely siphoning off US taxdollars and we're just supposed to believe he invented mars and electric cars and all kinds of other nonsense. that man belongs in a prison cell for life, and never to be seen or heard from again. 8:16 yeah.... we didn't forget how to do it... the Reichers that were working on these technologies were killed or covered up because no one wanted to admit what a horrible thing America did in world war 2 attacking their ancient homelands. 6:44 this makes me cringe as well; talking about "great achievements that have been lost". immediately I can hear idiots saying that elon musk invented the Egyptian pyramids or something or attributing other nonsensical ideas to that. again, this man belongs in a jail cell. 12:08 yes, this stuff doesn't happen overnight, which goes back to why the Operation Paperclip is so meaningful. I'm going to try getting off this topic now though, but I wanted to drive the point home. one more thing 14:58 is exactly what I'm talking about with trying to pluck the scientists out of The Third Reich, and why it's bad that America attacked their homeland of Germany, and i'm thankful to hear another human being say all of this. I was going nuts having these thoughts and hearing no one repeat them after I told them and the ideas weren't catching on 29:49 in relation to them talking about it being too expensive to fix something: there is so much tech debt in almost all the code i've ever ran into at a job that they have to start over. and to be frank: they shouldn't have hired programmers with hotel management degrees, chemistry degrees, i've even seen people who went to med school get hired to work as programmers and it was ridiculous. but at the same time the university I went to clearly had five times as many people as it could handle and the quality of education wasn't there, even before the flood of wannabe programmers seeking big paychecks and social praise. 31:13 yes, exactly. the point is the market niche. Even google seems to have done this. they clearly have problems with their code. gmail takes 60 seconds to load on my desktop and i just stopped using outside of exceptional situations. i cannot even get some of google's web apps to load in firefox and it's been that way since 2016. just recently I think this was addressed and I can now load them, but i don't even bother trying because of how little confidence I have in it working, PLUS they downgraded their interface from what it used to be and INTENTIONALLY REMOVED USEFUL FEATURES. i have never been able to comprehend what happened at google, starting in mid 2012 - besides diversity hiring (and it's been aaaaaaaall down hill from there). Freaking microsoft has been able to get the inside track on AI in its search, and bing's AI definitely makes it better than google search. google search is another piece of google software that doesn't work, and I can pinpoint the exact update and time when it happened; it was the Panda 2.0 update in the last week of august 2013 - i was forced to stop using google search because i could not find things. 36:31 on Jon's comment about "how much waste there must be": this goes back to what I was saying about the pay check generator that is NASA. 44:28 this is a disaster and is completely stupid, and the next thing you know you'll have recruiter or interviewer calling you and telling you you're not skilled enough to be a programmer or your skills are "out of date" because you don't know how to code for this completely nonsense architecture they made.... there is NO reason anyone should know how to do this - as jon says, there should be no reason anyone even thought this up in the first place.... this has to be the product of some nefarious actor reusing some scam they've pulled off before, like corporations making shell companies and proxy companies. 47:37 man who is this guy, Jon Blow? He is knocking it out of the park here. he's saying so much stuff that i've thought about and only even heard in small flashes in the pan online that i'm having some serious catharsis. this whole slide is exactly what the deal is. 48:39 and i've been saying how it's a shame that games don't have their own flavor anymore; they're all using the same engines and they feel the same. THEY LACK SOUL. 57:28 omg and this thought here about letting the AI handle the software... is this guy an ex google employee? this guy is like a copy of me. 59:04 agreed. (about the malfunctioning tools)
A very interesting talk, but I'm not sure how I feel about the takeaway. Ultimately he claims complexity is the root of the problem we need to make things simpler. I think it is more correct to say that a lack of quality and a need to rush is more the problem, which does result in extra complexity, but causes other problems as well. Sometimes it creates the opposite problem where there is not enough necessary complexity. Robert C Martin explains how programming evolved as new paradigms emerged as just taking something away at each step. Procedural took away gotos, OO took away function references, functional took away assignment. What is left to take away? Well I think he is asking a very good question, but the answer to all of this is more complex than that. In order to take something else away we have to add something else. Elegant complexity is needed in order to deal with the amount of necessary complexity now. You need to make something so elegant at every level, that something amazing can emerge from it, something that is more than the sum of its parts, like an emergent property, like how consciousness is for example. In order to achieve this we need to be able to specify more information about the types when programming. I call this type oriented programming (TOP). Every function should have all the parameters and return types be so specific that it is almost impossible for the implementation to be wrong. We currently don't narrow down what the parameters can be nearly enough. When taking in a whole number we always use the int type, even if it is intended to not work with 0 or negative numbers or if it will malfunction with number higher than 5,000. Strings can be anything, including empty. An empty string shouldn't even be representable. What is a string of characters with no characters? Nothing, the compiler should not allow it, the programmer should represent that state with null, not an empty string. The same thing with an empty array, it should not be possible, use null instead. Then with the new languages catching null reference exceptions at compile/static analysis time, so many bad states would be caught by the compiler/static analysis. So what is left that can be taken away? Implementation. When the types are specific enough, the implementation can be decided by the compiler or some static analysis tool. You would still have to provide additional hints like which methods/functions it is allowed to call to implement it. Then the tests would be auto generated. Your IDE could show you a bunch of examples of inputs and outputs and if any look wrong to you, you know you specified something wrong about the types or about how it is allowed to calculate the return type. Then you can pin certain examples that you see are correct, and it will save these as part of the specification, knowing something is wrong if those combinations of input and output ever stopped going together. Then the computer could work endlessly, trying to come up with faster and faster implementations while you sleep that satisfy the requirements of the types.
@@kodingamedev Thank you. This has motivated me to look into rust a little more. So far, the most similar languages I have found to what I am talking about are Ada (or better yet, SPARK), the latest version of PHP paired with Psalm, and now maybe Rust. Here is my current impression on the state all 3 of these, in comparison to my ideal language that I described. Ada and SPARK seem to be the safest and seem to have the most robustness around their safety. This is probably the closest you will come to what I am describing if yoiu want it to actually work the way it is supposed to, if you want it to have good tooling support, and without building your own language and tools. It seems that most programmers either don't know anything about them or are under the false impression that these are outdated languages that nobody would want to use for real programming, that it would only be forced on you if you worked in some field that required them. I think these are more modern and nice languages to work with than most programmers realize, and that is a bit unfair. But there is some truth to the idea. There are not enough modern libraries and frameworks available yet for these languages yet that would be useful for typical modern types of programming projects. But Ada does have a good package manager now, and the number of available open source libraries is growing. As far as the language itself, it seems to still be a bit outdated. There are many annoying limitations it has to make the static analysis more feasible, and I think a modern language rebuilt from scratch, but inspired by these languages could have less limitations while still being just as safe. I have seen people comparing Rust to Ada, and it seems they do have some similarities. It seems that most people who are into Rust don't know much about Ada and don't realize they are similar in some way. Rust is probably more modern and less annoying to work with. It is extremely popular right now, and I imagine there are a lot of modern, open source libraries and frameworks you can use with it. It probably also cannot provide the same level of safety you can get from Ada and/or SPARK. PHP is not a statically typed language. It started out as a templating language. They kept tacking on more and more features, becoming a worse and worse design, all the way up to and including the point where they added object oriented programming. The design only started improving at PHP 5, where they fixed many issues in the OO part of the design. Ever since then the language has continued getting better and better. They have a nice system and community around evolving the language in a positive way. It is still not a statically typed language, but you can now specify the types of class properties and function parameters. Psalm is a static analysis tool that allows you to specify many more things about your code, including but not limited to more specific types. It will validate if the things you specified are true or not. It's a bit annoying to use PHP and Psalm together, because you can't specify everything you want directly in PHP code, you need to use DocBlock annotations all over the place. But other advanced type systems have similarly verbose ways of specifying a lot of extra information. I just don't like the way I have to repeat myself, with a type in PHP and an even more specific type in the DocBlock. But Psalm does provide a very advanced type system, similar to what I described. You can actually specify that something is a non-empty-string, or a non-empty-list, or an int (an int in the range of 1-9). You can also specify that a function is pure or that a class is immutable, or that different methods follow different levels of immutability rules. It can also help validate some aspects of security. But since it is just a tool that is not part of PHP, and not made by the PHP core developers, it is a bit janky. It has tons of bugs. Using it as an LSP server is not going to provide you with an IDE like experience that you would get with a statically typed language. I think what we need is a team of people to build a new language from the ground up that learns from all 3 of these, and is designed to be more modern and less limited than SPARK while still offering the same amount of safety.
@@kodingamedev I’ve looked more into Rust now, and I’m going to say it seems you are completely wrong about this one. The Rust type system is not that great. It is nothing like what I described. Rust simply is statically typed and has some good protection in place to prevent certain types of bugs that only come up when doing concurrency. But the Rust type system doesn’t do much to help ensure the basic logic is correct. The types are not specific enough, and so you cannot create nice libraries. When you look at a function defintion, the types are so vague, you don’t even know what parameters it can really work with or what it can return. It is just a nice alternative to C and C++ that can be good for both system and higher level programming. It’s basically a successful version of D.
I mean... At least part of the complexity IS given by the hardware. You all know what a mess the modern x86_64 instruction set is, unless you know all that nitty gritty stuff about SIMD and whatever else, you can't really beat a compiler in writing more efficient code. I tried, lol. RISC-V is something different, thanks to it I understand at least the basics of writing assembly, and how a cpu works in general.
We need to build some kind of "knowledge eggs", containing both raw human knowledge and an AI capable of learning a human language it hears that would transfer various levels of technology knowledge to the ones who listen, from agriculture all the way to being able to craft an AI and build a new "egg".
This is actually an excellent motivator and use case for PKM systems to preserve the knowledge. The practice of maintaining a Personal Knowledge Base which contains the non trivial, deep knowledge related to topics that can be dispersed and used to preserve ideas
Literally sums up all the reasons for all the problems across all industries. That person who managed to blag their way to the top job but doesn't have a clue what's going on?! The UK is run by hundreds of these people.
Most of the issues he points out are just bugs. A) you can never test every single possibility in a website/program. You test as much as you can and fix bugs as people find them. That being said, some of the bugs were just developers not prioritizing testing and having dedicated, good QA engineers on the team. This is more an issue of most companies rushing to create programs, websites, etc., without leaving enough time for testing. That is a big issue in CS companies.
King Terry was one of the smartest programmers who ever lived. Too many people couldn't grasp his genius and misjudged him. He was a good and honest person. He thought things through. A skill, that seems to be getting rarer and rarer as society "evolves".
Also we will end up worhsipping these old game engines and OS's like the mechanicus in 40k. The knowledge was lost, so now we pray to them and they have "machine souls".
You also have to look at the 40k political climate. There's the top brass that just want the machines to work so there's no incentive to learn how all machines work if A plugs to B to produce C that was needed to battle this creature you just need to know that. I doubt there's budget for R&D in the 40k universe
I think one might be able to say "business" got involved. And like all things business touches, the incentives become perverse i.e. "the market doesn't reward robust software".
Just now getting into learning the very very basic bottom barrel concepts for programming, and honestly this presentation was eye opening but not really surprising, software today is fucking atrocious, the phones we got now suck eggs and I stg my iphone 5 was the greatest one they ever made, not only was the OS (pre ios7) great but the durability of the phone is something ive not seen since. It actually still works. Anyways, thanks for giving me an existential dread I didn't know I needed to stoke the flames of ingenuity.
A lot of good points, but to trash the LSP solution is short sighted. He isn't thinking about the alternative, where every editor has to implement a library to handle each software language. Or alternatively every software language team has to implement a plugin for each editor. This is the simplest / cleanest way that has been thought of thus far. As complicated as developing an LSP is, and I am looking in to it personally right now for my own language I am designing, the alternative is worse.
One of the best talks i've seen in a looong time. Very entertaining. Edit: it's really good and makes me think about everything i learned about software in the past 5 years.
I'm on board with the general idea, but not with the idea that software is getting shittier. It's simply not. The complexity has increased by orders of magnitude since the early days, and even then, the experience of working on a computer nowadays is insanely smoother than it used to be in the early days of PCs. Endless driver issues, installers simply not working, etc. I don't know why this dude doesn't remember that, but I certainly do. Everything is very much plug and play these days with minimal issues. The reason visual studio sucks is that it's a massive piece of software that tries to do too much and does nothing well as a result. Again, there was -nothing- comparable back in the days. Same with Steam. I remember struggling like a maniac to get most games running in 1999. Half the time installers would finish with errors, required libraries weren't packaged with the game, etc. Steam and games in general are insanely smoother than they used to be. Even if you tried to make the argument that they're not, most of the games we have nowadays that aren't just written by a random guy in his basement are massively more complex than the original Doom or Quake. You can't talk about decrease in quality without talking about the skyrocketing complexity of software. The software argument is just simply not true.
Technology may have advanced enough to release civilization from the confines of the second law of thermodynamics. These confines were imposed during Victorian England's scientific and religious cultural fascination with steam engines. The second law is behind modern refgeration needing electrical energy to compress the refrigerent to force it to release as waste the heat that it has removed from the refrigerator's service interior in the cooling part of the refrigerent's circulation. There is also discarded heat from mechanical friction and electrical resistance. Refrigeration by the principle that energy is conserved should produce electricity instead of consuming it. It makes more sense that refrigerators should yield electricity because energy is widely known to change form with no ultimate path of energy gain or loss being found. Therefore any form of fully recyclable energy can be cycled endlessly in any quantity. In an extreme case senario, full heat recycling, all electric, very isolated underground, undersea, or space communities would be highly survivable with self sufficient EMP resistant LED light banks, automated vertical farms, thaw resistant frozen food storehouses, factories, dwellings, and self contained elevators and horizontal transports. In a flourishing civillization senario, small self sufficient electric or cooling devices of many kinds and styles like lamps smartphones, hotplates, water heaters, cooler chests, fans, radios, TVs, cameras, security devices. power hand tools, pumps, and personal transports, would be available for immediate use incrementally anywhere as people see fit. Equipment groups would be on local networks. If a high majority thinks our civilization should geoengineer gigatons or teratons of carbon dioxide out of our etnvironment, instalations using devices that convert ambient heat into electricity can hypothetically be scaled up do it with a choice of comsequences including many beneficial ones. Energy sensible refrigerators that absorb heat and yield electricity would complement computers as computing consumes electricitsy and yields heat. Computing would be free. Chips could have energy recycling built in. A simple rectifier crystal can, iust short of a replicatable long term demonstration of a powerful prototype, almost certainly filter the random thermal motioren of electrons or discrete positiive charged voids called holes so the electric current flowing in one direction predominates. At low system voltage a filtrate of one polarity predominates only a little but there is always usable electrical power derived from the source, which is Johnson Nyquest thermal electrical noise. This net electrical filtrate can be aggregated in a group of separate diodes in consistent alignment parallel creating widely scalable electrical power. The maximum energy is converted from ambient heat to productive electricity when the electrical load is matched to the array impeadence. Matched impeadence output (watts) is k (Boltzman's constant, 1.38^-23, times T (tempeature Kelvin) times bandwidth (0 Hz to a natural limit ~2 THz @ 290 K) times rectification halving and nanowatt power level rectification efficiency times the number of diodes in the array. For reference, there are a billion cells of 1000 square nanometer area each per square millimeter, 100 billion per square centimeter. Order is imposed on the random thermal motion of electrons by the structual orderlyness of a diode array made of diodes made within a slab: ______________________ - Out 🔻🔻🔻🔻🔻🔻🔻 ______________________ + Out All the P type semiconductor anodes abut a metal conductive plane deposited on the top face of the slab with nonrectifying joins; all the N type semiconductor cathodes abut the bottom face. As the polarity filtered electrical energy is exported, the amount of thermal energy in the group of diodes decreases. This group cooling will draw heat in from the surrounding ambient heat at a rate depending on the filtering rate and thermal resistance between the group and ambient gas, liquid, or solid warmer than absolute zero. There is a lot of ambient heat on our planet, more in equatorial dry desert summer days and less in polar desert winter nights. Focusing on explaining the electronic behavior of one composition of simple diode, a near flawless crystal of silicon is modified by implanting a small amount of phosphorus (N type)on one side from a ohmic contact end to a junction where the additive is suddenly and completely changed to boron (P type) with minimal disturbance of the crystal lattice. The crystal then continues to another ohmic contact. A region of high electrical resistance forms at the junction in this type of diode when the phosphorous near the ĵunction donates electrons that are free to move elsewhere while leaving phosphorus ions held in the crystal while the boron ions donate holes which are similalarly free to move. The two types of mobile charges mutually clear each other away near the junction leaving little electrical conductivity. An equlibrium width of this region is settled between the phosphorus, boron, electrons, and holes. Thermal noise is beyond steady state equlibrium. Thermal noise transients, where mobile electrons move from the phosphorus added side to the boron added side ride transient extra conductivity so the forward moving electrons are preferentally filtered into the external circuit. Electrons are units of electric current. They lose their thermal energy of motion and gain electromotive force, another name for voltage, as they transition between the junction and the array electrical tap. Inside the diode, heat is absorbed: outside the diode, an attached electrical circuit is energized. Understanding diodes is one way to become convinced that Johnson Nyquest thermal electrical noise can be rectified and aggregated. Self assembling development teams may find many ways to accomplish this wide mission. Taxonomically there should be many ways to convert heat directly into electricity. A practical device may use an array of Au needles in a SiO2 matrix abutting N type GaAs. These were made in the 1970s when registration technology was poor so it was easier to fabricate arrays and select one diode than just make one diode. There are other plausible breeches of the second law of thermodynamics. Hopefully a lot of people will join in expanding the breech. Please share the successes or setbacks of your efforts. These devices would probably become segmented commodities sold with minimal margin over supply cost. They would be manufactured by advanced automation that does not need financial incentive. Applicable best practices would be adopted. Business details would be open public knowledge. Associated people should move as negotiated and freely and honestly talk. Commerce would be a planetary scale unified conglomerate of diverse local cooperatives. There is no need of wealth extracting top commanders. We do not need often token philanthropy from the top if the wide majority can afford to be generous. Aloha Charles M Brown Kilauea Kauai Hawaii 96754
Yeah, like, defunding public schools, making universities more expensive and less accessible, creating an economy composed of completely decentralized and competing standards, all contribute to the decline of technology and, eventually, the collapse of civilization.
what an interesting talk. i wonder if windows move to arm will help or hinder as well as llm's ability to parse code as a means to simplify the underlying substructure or the coding process thereby correcting avoidable conflicts or dll type problems. so glad i found this vid. if nothing else it got me to thinking
I do think rules are different. A data breach can be deadly for a company. Governments do hava a look at what kind of data you own with who do you share it? The perception of good and bad people(and I am not talking about hackers) is coming up on the internet. Then there is the pay as you go model(basically subscription, how complicated can it be to identify an human, identify his bank and take his money today compared to before). Then how complicated is it to run ads on millions of websites and all of those ads are provided by one provider? How had is it to serve 4k video. What about live streaming. How hard is it to build a virtual reality? Now lets ask the really question, why does windows terminals sucks hard. I have to google to delete a folder recursively then type a hundred Cap in the prompt.
You're not paid to design good software and think it through. You're paid to implement exactly what and how your stakeholder/manager requests it. If want to enhance something or if you say anything against the plan/desciption you're not complying with your leaders and you're making extra hassle - so either face HR meetings, where you have to justify your good intentions - or implement garbage to stay in line with everyone else and get your paycheck. I loved clean programming - challenging myself to make it as performant and reliable as possible no matter what you throw at it - it was beautiful - until I got forced into a different coding style and got money for it. I was in multiple areas of aviation, industry, game dev and online social services as senior fullstack dev, dev ops and teamleader. Its the same everywhere, so that the non-techies and not-so-intelligent-programmers all feel equally heard and responsible for the profit and all opinions, thoughts and decisions are worth the same. Its cringy theatre but you get money for it. It's easy: You're not trusted to make most of the meaningful decisions, but so you aren't responsible for their broken outcome as well - even if you already saw it 3 months prior to the development in the flawed design and brought it up 10 times since then with background and examples. If the boss wants to see his money burn, because he thinks that everything is good and that this is how its supposed to be, then it will be like that. You'll get fired if you tell him how he could increase his efficiency, because nobody tells the boss what to do - especially not the coding peasants that should be down in the code mines, mining features, like they're supposed to. Software development was a virtue of academic elites, who wanted to create or do something, because they knew or wanted to know or do something - now it's just a money making grind on the leash with more interest in responsibility than why something is the way it is. Philosophy got turned into economy, so technical quality became profitable trash. Good software isn't written with money. Good software is written in freetime out of personal motivation, knowledge and skills. Solution: Give some good software devs (= with personal interest in coding) more money than they need (not tied to worktime or outcome), so they don't worry about it. Give them all the information they need to understand a system and what its made of. Tell them your desires and let them come up with a way to accomplish that. Let them take as long as they want and they'll be done faster than any current team and it will just work.
Not many programmers today, just coders. Biggest deception to IT (or rather whole academic/higher level education) is that people were convinced that anyone can go and be "programmer", or in terms of education that everyone not only can but should go to collage, university or polytechnic. Were it partially true as you can make nearly anyone to code. You can even make some proportionally less people to be programmer (teach concepts and issues which have to be solved), but you can't make someone engineer if he/she doesn't have engineers mind. Big tech today are just money greedy molochs today run by business guys, which have bunch or random guys doing "coding" according to templates passed to them by others. Meanwhile real, amazingly smart, passionate and wholehearted engineers just quit jobs as that can't stand these madness, or just give up and do shit without improving anything as no one seems to care about it.
For correctness - this idea that science does not automatically get better was not invented by Musk. Musk often emits ideas he read somewhere else without mentioning the source. We can give him credit because he used the money he lucked into from the startup world to work on cool things. But let's not elevate him to some sort of genius. All the work is done by people much smarter than him.
Well he and most people use windows and he has to get the point across in a reasonable amount of time. I've definitely seen him rant about Linux on his stream before.
No. Linux is not this mystical infallible, super secure OS that's cursed with lack of software because it doesn't play ball with those evil corporations. Desktop linux is an infuriating piece of shit, and therefore barely anyone bothers to release software for it or compromise its security. And it will remain so as long as simply installing and customizing it remains a rite of passage to weed out those "unworthy". And believe me, I tired. Ubuntu, Mint, #!, Puppy, MX... Windows is a decent OS buried under tons of bloat imposed by suits. Desktop Linux is a trash fire propped up by tremendous amounts of goodwill. Of course, both are rapidly getting worse over time, each in its own way.
