Comments by "Winnetou17" (@Winnetou17) on "Theo - t3․gg" channel.

  1. On the Google doesn't support JPEG-XL because CPU costs... that's a cop-out. They've been asked multiple times, including by companies like Adobe, to implement it and they just straight up refused also acting like nobody cares about JPEG-XL. You can say whatever you want, but if you read the chromium thread about it, it's clear as day that Google DECIDED against both JPEG-XL and the community with no technical reasoning. Of course, they didn't say out loud why. Also, the CPU costs of JPEG-XL cannot be the only reason of rejecting it. There's loads of other use cases. Not to mention the forward looking of, JPEG XL is alone in the amount of features it provides/support that is outside of just some 300x300 images on a random website. You can have wide color gamut and HDR in it. People taking high-detail high resolution photos, can put the original images directly as jpegxl. But Google directly jumped in to say "nope, cannot see that directly in the browser. Think of the children in Africa!". I'll stop, my blood is starting to boil. I see that close to the end of the video these extra features were mentioned, so at least I'm at peace that Theo saw that. - I'm totally on the webp that it is good and it should still be supported as the bridge to JPEG XL. - AVIF has nice features, but it's a dead-end technology, we shouldn't bother putting more effort into it. - JPEG XL is truly the next generation of image formats and what we will use for decades. Hence why it's so infuriating of Google actively being against it and slowing massively its adoption. No, I'm not buying out the CPU cost argument. That's not the browser support problem. In a world where a website has a 6.6 MB JPEG with .png extension image, there's PLENTY of room for JPEG XL. Lastly, I really want to see benchmarks of just how slow JPEG XL supposedly is. I found out a benchmark, apparently from 2020 which unfortunately doesn't have a comparison to webp. And just citing X or Y MB/s is meaningless when you don't webp on the same exact hardware, as the numbers can differ massively from one CPU to another. Not to mention that 4 years have passed, things might've changed. EDIT: Initially mentioned about JPEGXL being lossless, but Theo corrected that very quickly. Also mentioned about PNG: PNG not being compressed ? Duuuude .... It's not a good compression for photographs (hence why people still use JPEG for that, it looks basically the same to a human, if you don't need to zoom in, and it is a smaller size as a JPEG). But PNG itself totally has compression. And multiple ways in which you can reduce the file size. There's a reason TinyPNG exists (or existed, didn't check recently)
    208
  2. One thing to mention: on things that should be on release, you can still have a little effect on press. Buttons are like that, to have 4 states: normal, hover, active (aka on pressed, no released) and disabled .... if I remember correctly. Having something like a color change on press would still maintain the feeling of very fast interaction.
    106
  3. At 1:08 the editor did not get the joke. There's 2 hard problems: naming things, OFF BY ONE ERRORS and live sync (cache invalidation in the version I knew)
    104
  4. The graph at 3:00 is classic NVidia shenanigans. These guys are totally incapable to ever produce a graph that's both correct and not misleading. And useful. Other than that, yeah, good corrections. There's even more minor corrections, if someone would be pedantic. I think that Theo deep down knows he's not accurate / properly correct, but overall he gets the big idea pretty well. Edit: one thing that I can give Theo is that he does listen to the suggestions and corrections, well, as much as he can take in a limited amount of time. You can see in this video that when he showed the fast inverse square root he didn't say that John Carmack invented it. That's an improvement on his part.
    28
  5. My stupid brain initially read that as "Cheese from Switzerland" 😄 I might be hungry...
    19
  6. While I hate Microsoft about as much as Apple nowadays, your take on Windows in this specific case is totally wrong and uncalled for. Only Windows was affected because only that update was affected. Windows doesn't have anything worse than Mac or Linux here. Also Mac and Linux could've been just as much affected. That whole section is very cringe. Edit: I mean the section starting at 3:35 The rest of the video is a-ok
    16
  7.  @privateanon7623  Yeah, but you need to see the bigger picture then. There are cases where it should be on release. Like allowing drag-and-drop (which Carmack mentioned). Whatever happens directly on press has be easily undoable (like on a virtual keyboard) or not that important.
    15
  8. Awesome, now that Theo switched to Zen, thanks to Firefox being so open and customizable, maybe now we can have more contributions to Firefox and somebody very known and vocal who can trash the websites that DECIDE to not support Firefox for absolutely no good reason (which is easy to prove when you change the user agent to appear like Chrome and see that evrything works perfectly). Theo, here's the formal request to revisit Firefox (like, say, when version 135 arrives) and do a video of what it is missing from it, what is wrong, what is insufficient and so. For me it's still my browser simply because of full open source and it works good enough for me. On the developer side, the fact that it has on the network developer tools the "Edit and Resend" makes it better than Chromium-based browsers. Not to mention that for mobile, it's the only one that supports extensions.
    12
  9.  @harrytsang1501  The hardware acceleration is a very good point, if it can be used it will be indeed be both very fast and energy efficient. Regarding the dead-end part, I might've jumped to conclusions too soon. Unfortunately I don't remember anymore the bigger list of limitations or downsides it has. From what I remember it was the max resolution and the lack of progressive decoding that were major hurdles. As in, it's almost as new as JPEG XL and it already has known downsides, why get used to them instead of jumping directly to the overall better format ? Now if these can be added in then... I'm not against having more formats. As long as we don't get to 14 of them :D
    11
  10. What I see at 4:04 if you pause and go frame by frame is that the AI started to change things it shouldn't too. First instance at 4:05 wheere it says it was $0.00340 to store in WorkerKV and the AI simply replaced it to $0.002 even though that $0.00340 wasn't part of the previous computations. Not gonna lie though, overall it's still neat, but I don't think I'll ever get over this "I have to recheck every little thing it changed, to be sure". Nice that it shows it with color highlighting.
    11
  11.  @Firestar-rm8df  Regarding that Memcached vs Redis... They're both in C. Please do the minimal amount of research before spewing such BS, you're wasting everyone's time. Also, Redis is single threaded especially for the speed it gives when you have no context switch from one thread to another. But memcached has other benefits.
    10
  12. Depends on which country in Europe. Western Europe has salaries bigger than 2000 €, for devs as associated roles. But in Eastern Europe (which have a lot of programmers because they can be hired remotely and are cheaper than US and Western Europe ones) it is about right. However, this being a donation means it's much likely that it will be taxed differently, at a lower rate. It certainly doesn't need a cut for a retirement pension. So if you take into account that it's probably 3000 € before taxes for a dev per month, 2000 will make it less than a month, so probably in the region of 5%. While the western europe might have this be 2% or lower of what a company spends on an individual developer. Still, the spirit of the idea stands. If you can't do the arbitrary fixed sum of $2000/year, just do 1% and call it a day. I'd say it's just as good, espcially for a company in a place with lower salaries than Silicon Valley.
    8
  13. That's so stupid I can't even ...
    8
  14. This needs more upvotes
    8
  15. This needs more upvotes
    8
  16.  @jabbruh  Yes you can. Since PHP 7.0 (which is many years old), though it wasn't everything at once, bits and pieces got added with newer versions. It's still not perfect, as Nick Hamilton above said, you still don't have typed arrays. But you have classes (first thing they added), you have the base (scalar) types like int or ?int (int or null aka nullable int), you have union types so you can more exactly define arguments and returns (like int|float aka int or float). How many decades ago have you last time checked PHP ?
    7
  17. Yup! Very well said. Sad to see all of that being kept "outside" because "CPU costs for old smartphones!"
    6
  18. Yeah, that confused me too. It's written the same exact as the color, WTF
    6
  19.  @andreas.111  It got that years ago
    5
  20. Theo: Check out Sam! Me: Sam who ? Theo: Yeah!
    4
  21. In the stream from which this video is made, Theo said that Firefox is a bad browser, but Safari is a toy, not an actual browser. So, yeah, even harder to love.
    4
  22. That's not the point of the video ... ??
    3
  23. JPG is lossy even at 100 quality.
    3
  24. You wanted to ensure they have an equal split, lol.
    3
  25. Regarding the javascript vs java at the end, can't there simply be a clause that whatever happens to JavaScript™ cannot be used in a case of Java™ not being taken care of ?
    3
  26. Yeah, the double standard rouding is a bit cringe. But you can tell he's really happy, and with those numbers, I'd be happy too.
    3
  27. Except he did the math wrong and forgot a 24 in the big seconds to month multiplications (I so hate how much he rushes) which would make the cost about $28k, when using the 500ms time. With, say, 100ms time, it would be $5k. I do wonder though, how would the serverless response times be. They would be bigger than 90ms, right ? Since there's spin-up time, well, at least for some of them. Is that spin-up time 10ms ? 100 ms ? 1 second ?
    3
  28. Here I place a formal request for Theo to STOP DOING MULTIPLE MULTIPLICATIONS AT ONCE. And in general to STOP RUSHING over delicate things/parts like that. The rather large and significant math mistake at the end would've been TRIVIALLY avoided by not rushing so much and doing the math in multiple steps. Like first stating how many requests are per minute, then per hour, per day then per month. Or at least something like 3600 seconds per hour, 720 hours per month to keep things still simple. I really mean it that it was really a rush, as what I requested above takes SECONDS extra to write. Can you NOT feel dumb for doing a big mistake because you wanted to save, say, 1 minute, out of a 74 minute video, kinda spoiling it at the end ? Also, Theo, please place the multiplication in the written part too. If you miss something, you AND US can see it there much more easily and not have to search in the calc app history. Not to mention that you switch things so fast there, it's very hard to keep up and spot the mistake right then (that is, without pausing), it's almost looking like you WANT to hide something (which I know you don't, I'm just stating how bad this is) After all, it was stated that serverless is more expensive and that the cost difference might not matter that much anyway, so the overall conclusion is still kind of there, but it's still so SOO frustrating to see (not the first time) such glaring mistakes that also might make someone less knowledgeable really question the video and conclusion. For something so easily avoided. STOP RUSHIIIIIIIII
    2
  29. Yup! Well summarized!
    2
  30. I don't get this video. I mean, the story is nice, and it makes sense. But it's a story of why Theo ditched Windows and went to Macs (initially because hardware, then because software aka MS made Windows bad). Linux was just a tangent. And other than maybe Photoshop, I don't see why the "gave up" on Linux, it simply looks like Theo didn't try it in an environment where it had to actually code. If that were to happen, I'm sure he would've loved Linux the same as Mac. I'll be waiting for Apple to force something to not work anymore or to be upgraded into an inferior version, for the love of Open Source to come back to Theo. Where nobody can force you to stop using something.... though in practice it does happen too (hard to use very old with very new programs at the same time, though AppImages and VMs can mostly solve that), but it's really REALLY rare that it will negatively impact you.
    2
  31. Yeah, not to mention how misleading is to show the single core improvements of CPUs compared to basically increase core counts on GPUs. Guess what, CPUs are getting more cores too. I can't think of any ANY graph made by NVidia to be correct and useful or insightful at the same time.
    2
  32.  @t3dotgg  But you can do both (though I'm ok with just having the long videos, since I like understanding the context and nuances). That is, AFTER you do the initial video, with all the details, you can make a summary video, a more distilled take. There's things that can be taken out like the chat interaction and tangents and the time it takes to write and compute. And you can make corrections too. Now, if you have the time and mood for a summary video, that's another topic. But it is (at least theo-retically) possible and not pointless.
    2
  33.  @Firestar-rm8df  Ok, I've been a bit of an ass, true. On the multi vs single threaded part, Redis can handle quite some requests, though, of course it does have a limit. And for using something like a Threadripper, you can use multiple instances. Might not be something very easy to do, if you already have an app, or have very specific data requirements, but in most cases this can be used, so you can have both blazingly fast™ response times and lots of users/data. Still memcached clearly has its place, if I'm not mistaken it can easily scale to be used on multiple servers at once, as a big pool of TB of RAM if you needed (and not in the sense of individual servers that are copies of a main server, neither manual sharding as I suggested with Redis)
    2
  34. gif is really niche nowadays, only usable for small simple images or small simple animations. And the animations part wasn't the topic of this video.
    2
  35. It totally is just you. Do you have extensions ? Maybe they're interferring or something. Does it happen with Firefox too ?
    2
  36. But is it dynamic across different resolutions ? That is, can you define it to look good on all resolutions without using multiple definitions (aka using media queries) ?
    2
  37. Well, at least it's in the name :))
    2
  38. He used the average of 2625 req/s (aka half of the peak of 5250), but yeah, that still amounts to $28k, effectively 10 times more expensive.
    2
  39. Don't know if he switched to something else, but I think it still is Arc. If I'm not mistaken, there's also a new one called Zen that also has this feature baked in, and it's based on Firefox (Arc is based on Chromium)
    2
  40. You over analyze and probably have too much time. Peace geeks ✌
    1
  41. Apparently this was recorded in like august (more or less) last year
    1
  42. Yeah, but nowadays gif is useful only in pretty niche cases, for small, simple images. Or animations, but the animation part wasn't the topic of this video clip.
    1
  43. That's quite extreme. If it's 15 hours per week, so 3 hours per day, so 1.5 hours per direction, you should get a permit as it is very wasteful. It should be below 10 hours per week (1 hour per direction per day)
    1
  44. Wait, so if I understand correctly what's said in Part 2, when you had 1 event before, if you want to have the same data, you have to use 2 events now ? That invalidates a bit the cost savings.
    1
  45.  @sgtyota  And progressive loading. And wide gamut + HDR support.
    1
  46. It's easy, we don't have a life, so we can watch things like this when we eat, for example 😀
    1
  47. I have this vague feeling this will happen more and more going forward.
    1
  48. Amen! A lot of the hate came from the pre PHP 5 and PHP 5 era and people haven't adjusted to how PHP 8 is today.
    1
  49. It was always worse than PHP
    1
  50. This was an interesting watch. Informative, but also flawed, in multiple points. Before going point by point, I'd like to say that I am a Linux user and, while I don't go out of my way to bash Electron apps, I'm not a fan of it either, and I prefer to not use them whenever I can. At this moment I think I only use one: Postman, which annoys the hell out of me when it hangs for like 30 seconds, and I will change it soon, once I have more time to check the alternatives and settle on one. Still, Postman I think is trash because it is made like that, not because of Electron. And it annoys me for other reasons too. But regardless, it is very much looking and feeling like a much heavier application for what it should be. I do have an 8 yeah old laptop, which was powerful in 2017, not so much now, but it's not trash either. One thing I must point out - I am using apparently a lot of software that has an Electron app. But I don't use the app, I use what is sane to me - a simple browser tab, in my browser. I was using this before Electron was a thing, so I kind of happened to be like that, before deciding to not use Electron if I can. Why not use Electron ? Some time ago MS Teams had a bug which let remote code execution within it, because of bad sandboxing. That's when I knew I'll never install it, as running it in the browser I can be much more sure that it doesn't have access to my system. Especially after finding out that Discord for YEARS used an extremely outdated version of Electron. Imagine how secure that was! And the performance penalty is there, I don't want it. By only having those directly in the browser (seriously, why are people not using this ?) I have almost 0 disk space used, much less security issues / stress, much less RAM used. So I feel quite immune to the "either Electron either nothing" threat, though I do understand it. For me it didn't enable me in any way to use Linux. As IDE I use JetBrains. On the performance side... that example with the SwiftUI and the conclusion that Electron might be faster than native ... I am soooo NOT buying that. I call 100% skill issue or just something related to SwiftUI at that point in time. Even the first update it says it swapped one component from SwiftUI Text to SwiftUI TextEditor. And apparently it does the rendering using the CPU ? Disable hardware GPU accelerated rendering on a browser and watch the same pain there too. You can't really say that native in general might be slower just from one example like that. Chrome, the only way to efficiently render text (without being John Carmack), really ? Talk about blowing things way out of proportion. In the end, I don't like Electron for the same reason I don't like Flatpaks and snaps. Having everything bundled. In case of flatpaks and snaps, I'd rather have a statically-linked executable. In case of Electron I'd rather simply use my browser. Having one embedded will always be a problem to me. Maybe with servo it will get to be less of an issue. Still, I'm not thrilled on having JS either, though for simple applications it is fine. I would still like something that can be more efficient. Not just in CPU usage, but also in RAM usage. A waste is still a waste. Even with the above, I don't hate Electron for existing. I do kind of hate it becoming the norm. It's like AGAIN we forget all the years of lessons we learned before, most of everything we've learnt and built for something that's short-term easier. Maybe PWAs will become a thing in this lifetime, and we can stop shipping an almost full browser with every app.
    1
  51.  @snowe..  Couldn't agree more that Firefox will still continue without Google's money too. Because a lot of that money is wasted on some type of people and useless projects (I'll let the read guess of which kind). Who knows, without Google's money, maybe Mozilla will get rid of those people and have a chance of getting relevant again when it has only engineers and ACTUAL free speech activits. Which might get them more funding in the future, if people see that their money isn't wasted anymore. And/or if they make a separate place/location/category to donate directly to the Firefox and Thunderbird devs, I'm sure it will actually attract enough funding to maintain the current level they have.
    1
  52. Apparently this was recorded in like august (more or less) last year
    1
  53. I do agree that what DoJ proposes is an overreach. And I think it's why it will not pass/continue. However, I totally disagree that it will kill the web or anything remotely close to that. There's plenty of innovation done outside Google. If Chrome disappears today, sites and products won't break, as Chromium can serve them with absolutely no issues. If Chrome disappears today, we won't be stuck in the current version of the web, it will continue to be improved. Maybe slower, but we'll be absolutely fine. I hate this type of argument so much. That if we don't sell our soul to the devil (figuratively) then we won't be able to do anything. Ok, maybe 10 or 15 years ago it might've been closer to the truth, but even then I don't agree that we wouldn't have have the benefits of the modern day browsers without Chrome. It would've taken longer most likely. But nobody would've died because of that. But it's even less of a concern now, with thousands of non-google engineers contributing to web standards and browser code. I wouold also not discount Firefox so quickly. It's true that most of Mozilla's revenue comes from Google, but a lot of the revenue is wasted on useless projects (usually DEI stuff) and on the woke staff itself that managed to get the leadership. Firefox developers don't get a lot of cash. Not getting the money from Google might actually be better for Mozilla and Firefox, as it might be freed from the woke management and steered back into competence and relevancy by engineers and ACTUAL free speech activists. And if they make a dedicated Firefox-devs only donation category, I'm sure that there will be people chiming in (myself included), enough to at least keep the current funding the developers get.
    1
  54. It's easy how he got there - 90ms mean response time, meaning you can do about 11 requests in a second. 10 if you round it. So, in order to have 0 extra latency, he just said that you can do 10 requests per second per core. It's true that in reality, that 90ms response time is more like 40ms CPU time and 50ms waiting for stuff to be loaded, like a database query response, or the SSD to return file data and so on. So a core could handle more than a request at a time, but it depends on the CPU time vs iowait time proportions just how much it can do. And after a point, it can do multiple at the same time, but there might be a bit of latency on all of them. Still, it's just easier to compute the way he did it, and also has some headroom, just in case.
    1
  55.  @olavrb  What latency ?
    1
  56. LoL, I made the same comment, only now I saw this one. Yeah, it's a programmer joke, I think I'd miss it too if I wasn't one.
    1
  57. I was thinking the same. If you have a base load, consolidate that on dedicated servers, but still have a way to use serverless for extra load and spikes. That would combine the best of both worlds. I wonder who's already doing that out there (surely there is somebody).
    1
  58. Edit: bad take, kinda off-topic, see next replies. That's the thing, SVG only makes sense in certain contexts. If you have any photo or complex drawing in your website you either can't use SVG or it will be a massive downside using it. Overall, it doesn't make sense to have a website just with SVG, as a general statement, so that's why it's a cross.
    1
  59.  @hopperelec  Hmm, actually, you're right, sorry about that, got things mixed up in my head. If that were the case I don't think we'd have fonts as vectors for decades now. Though fonts are probably a niche case where more optimisations can be had.
    1
  60.  @no_the_other_ariksquad  From what I know, no, though I'm not 100% sure.
    1
  61. Drink a shot everytime you hear masonry. Then realize this isn't a video about Civ 7.
    1
  62. It's just Java that the problem there. And I say that as someone who hates Java.
    1
  63. It's rather new.
    1
  64. Yeah, for simple websites, it's still perfectly ok, no need to complicate things so much. $0 tooling
    1
  65.  @Ben-ts7ut  Yup. But if you already transplaced the results into text, then it is a bit tedious to change it. Technically you can have in a Word/Google Docs document values from an Excel/Google Sheets spreadsheet, but that's quite tedious to set up as well.
    1
  66.  @trumpetbob15  Interesting, good to know, thanks!
    1
  67.  @igoralmeida9136  Or instead of going bankrupt, they might boot off the woke management which are wasting a lot of that Google's money into useless projects (usually DEI) and have the revenue go to the devs. Which might make them get even more revenue later, if people see that they're no longer wasted on propaganda.
    1
  68. Functional or procedural ? I have this vague feeling that you're actually referring to procedural.
    1
  69. LoL! And MOOTools ... whatever version they had + 1
    1
  70.  @thewhitefalcon8539  It's both true and untrue. Yeah, compilers are somewhat (and I think have always been) nondeterministic and unpredictable. But it will never put things like remote API calls in your code. The amount of things an AI can put that's unpredictable is an order or magnitude more. With compilers, the amount of wrong it can do is much more contained.
    1
  71. It is. Though it is from 2011, and Apple was using ARM for the iPhones since 2009, right ? They could've had some input on that, though the wikipedia article about big.LITTLE doesn't say anything about Apple in the initial development.
    1
  72. I'm not that convinced. First, Google already showed how much they love ads, so that alone is in direct contradiction. At most they can say that Chrome team doesn't have that goal. But something that I haven't seen anyone address (didn't search much, to be honest) is: why the 30k limit ? If it's not in bad faith they should allow a much higher limit. Maybe mention when you install/enable the extension that it has whatever number and it might slow you down. But don't strictly limit it.
    1
  73. Agreed
    1
  74. Yeah, I was confused about that too. Well, he does say that it's ok in other PHP land, and this is pretty squarely just a Drupal 7 thing. I heard about Drupal 9 several years ago, I'm amazed that Drupal 7 is still a thing on websites that have people working on them. Hopefully all the breaking changes in Drupal 8 are worth it and the next upgrades easier.
    1
  75. Eh, if you read the comments, you realize he messed up a lot of details. And I'd say that even the broader picture (pun not intended) is a bit off.
    1