Robin: Hello and welcome to the first episode of our new podcast Syscadia. I am Robin and I am here with Trevor
Robin: And we are going to talk to you about technology. We are going to get in depth. We are going to go over the things that we care about and we might get kind of nerdy, but it should be a fun time for all involved.
About this Transcript
This transcript is provided for content accessibility for a variety of users, and hopefully can enhance your enjoyment of Syscadia! It was created using computer speech recognition with light human editing, and as such may contain small errors or inaccuracies.
Robin: Hello and welcome to the first episode of our new podcast Syscadia. I am Robin and I am here with Trevor
Robin: And we are going to talk to you about technology. We are going to get in depth. We are going to go over the things that we care about and we might get kind of nerdy, but it should be a fun time for all involved.
Trevor: Just pretty much a conversation between us with kind of a loose agenda.
Robin: Yes. There is a conversation with a conspicuously placed microphone… and nothing else is unusual.
Trevor: I have really been wanting to talk to you about that ARM stuff a little bit more in depth because that I’m honestly excited about. And I don’t really get excited about technical things that easily. I’m like, eh, they’ll probably screw it up, but, no, this is important. And you know, the future of that with windows, and working with everything else using it.
Robin: Yeah, so we’re talking about the new MacBook Air and the new MacBook Pro. Apple released their new ARM based Macs -which they’re not calling ARM because they licensed the stuff from ARM- but they don’t put ARM on their box anywhere. They call it Apple Silicon. Apple Silicon is their trademark.
Trevor: Oh, so they’re not actually calling them ARM.
Robin: They have never used the ARM brand that I’m aware of.
Trevor: Yeah. So, does that mean, can you use existing ARM tool chains?
Robin: Yeah! You can pay a licensing fee and basically buy that instruction set and you get that version of ARM. Now it’s from Nvidia because they bought it. And you get the permission to use that instruction set. So [Apple has] like a super set of the ARM instruction set, and that’s what they use for their Apple Silicon.
Trevor: Oh, wow. Okay. Um, so one, I had no idea that ARM was bought by Nvidia. I’m not sure how I feel about that.
Robin: There’s a fair bit of incentive for Nvidia to continue to include Apple and ecosystem because Apple tends to have good ideas for things to include in the instruction set.
WebKit and the Future of Firefox
Robin: Kind of like what Apple did with WebKit. We all use Apple’s browser now. Because Chrome is WebKit and Edge is WebKit. And, I think even now, Firefox is discussing moving to WebKit. So…
Trevor: so, I know Firefox has been discussing that forever. They’re still on gecko. I think I heard something about opera wanting to do something like that. Or is Opera just already WebKit?
Robin: I’d have to look now
Trevor: I think they did with 10.1 or something, because that was when that was when they started supporting like modern
Robin: Yeah. I like how I searched just for opera. I was, for some reason I expected that to be the right word. Um, Opera browser. There we go. You can tell that techies are running Wikipedia because you can search for the page “opera.” It’s a common word, like, billions of people know what that is.
And the very first line in this article about the Western art form… “Did you mean the web browser? See opera, web browser. “
Trevor: I appreciate that.
Robin: [Opera] moved [to chromium] seven years ago.
Trevor: I’m out of date.
Robin: Do you remember when it was like a BitTorrent client and a mail client. It was all that for like five megabytes.
Trevor: No, I don’t remember. I was not involved in the early Mac OSX era. Uh, the most I remember about it was it powered the Wii browser, which meant it was a pile of garbage.
Robin: This is PC era. Even as far back as like, I think 95 or 98 people would download it because it was like an all in one.
Trevor: Oh, that’s why Firefox was doing that. I was okay. So, wondering why that happened to Firefox and like its early years and they completely dropped it where they were like, we’re going to cram Thunderbird in here and we’re going to cram transmission in here and everything.
Robin: Firefox itself never had that. Mozilla back when it was forked from Netscape. Netscape existed, was forked, and Netscape navigator existed since everything was open source. they quickly formed Mozilla and mozilla.org. It had the beautiful red and black, (almost soviet looking color scheme, honestly), with the big, Mozilla kind of dragon and, scriptural sounding Bible verses about Mozilla talking about like… mamon. so that all existed from the early days.
And that had a mail client, a chat client in the form of chatZilla.
Trevor: what I was thinking.
Robin: And it also had a new script service and a few other things, all kind of bundled together into this one engine, and you could install themes and all was grand in the world, but Mozilla realized that people largely didn’t care about the other things. They just wanted a browser. And it was a lot of work to upkeep all those components together. Mozilla was renamed, Mozilla Suite. Just the gecko engine and the chrome and as in the browser interface language XUL was pulled out into Firefox.
Trevor: I can’t tell you how confusing it is. Whenever I try to look up like making extension stuff for Firefox, and I’m seeing references to Chrome and I’m like, what, you talking about?
Robin: Yeah, and everything is still NS for Netscape, NS dot whatever. It’s very ancient now. It’s kind of cute. But effectively that was the start of Firefox. It was called Phoenix initially. But other people were quickly like, you know what, there’s already a Phoenix. You shouldn’t do that.
And Mozilla was like, “Hey, we’ll choose a word that no one could possibly be using, Firebird “ yeah, of course, Firebird exist as a database engine.
It caused confusion. It wasn’t good for SEO. So, then the third time was the charm, and it became Firefox. I was around to watch that decision happen, which is kind of cool. I was UID 8 on their new website.
Trevor: Your Mozilla tenure’s legitimately kind of like terrifying. You’ve influenced the actual development of the internet.
Robin: To some degree. But, that’s when Firefox was born and a lot of excitement happened, there was a huge amount of, attention around it, and Google noticed the attention and basically hired all the well-known people from the Firefox community.
Trevor: That might be a good jumping off point for the state of Mozilla as it is now, which is kind of weird. I’m not sure what the heck they’re doing
Robin: Mozilla had this phase where. In a weird way. Google somewhat destroyed Mozilla as a, it was very scrappy. They had very little money. They required people to donate their time and donate their money. There was this period where you they would do fundraising rounds where you would put in a couple of hundred dollars and get like a, like a Firefox plushy, which were very cute, I have one.
Then when Google became more of a force, they started paying them a ton of money to include Google as the default search engine. So Mozilla started getting millions of dollars. They started hiring a lot of people. They started to build more of a corporate structure. Mozilla corporation was formed.
Mozilla up to that point was nonprofit. Then they formed this other corporation. And so the corporation was a fully owned subsidiary of as a foundation. But then they had a lot of head count and they started adding more and more people. They started adding corporate structure.
Suddenly there was less open-source contribution. Most people who worked on the browser were employees. Mozilla hired a CEO who had more experience and background in being more of social causes CEO.
Trevor: Oh, that explains. A lot of things, like why they’ve constantly had so much messaging around that, which I’m not saying it’s a bad thing. I kind of liked that they’re trying to do stuff to change the internet, but it is sort of funny to see that juxtaposed against the current need for funding.
Robin: Yeah. They, they kind of stopped being a technology company because all the leadership in charge of the company was started to get focused around activism. Um, but they were already people in that space because on the legal side, EFF does a much better job. On the, on the standardization side, we already had the worldwide web consortium, W3C.
Trevor: I was honestly really surprised that they’d never got into the standardization thing more because I, I felt like for a while, there were a couple of years where they were really trying to push that, like our browsers faster, where we have the standardization of extensions and we’re making this format that can work for both Chrome and Firefox and.
They were really trying to push that for a while. And then that just kind of dropped off And I don’t know where it went.
Robin: They’re downsizing rapidly now. My belief is that as they’re laying off people, they’ll probably look for some way to reduce the cost of having engineers at Mozilla. So they’ll probably look into ways to just make a chromium based browser, call it Firefox. But yeah, I don’t think they see themselves as technical anymore.
Trevor: Well, see, I’m not so sure about that because they have recently released a quantum, I think is what they’re calling it. And it’s their new engine. It’s really good at multitasking memory management, supposedly whatever. I don’t really know. I just know it runs faster.
Robin: Yeah, I don’t think they’ll be doing anything with quantum. I really suspect the next year they’ll switch to chromium.
Trevor: I hate that. If that happens, I actually hate that because there’s no competition for the bloated crap that is chromium. Not that it’s terrible. Chromium is great. And it certainly set a lot of standards, but Oh my God, is it bloated!? Even using the open source version is like one gigantic pain. It takes over an hour to compile on a laptop, it shouldn’t take that long.
And I’m speaking from experience of using Slack (which was a terrible mistake because I wasted so much of my time).I just didn’t have the tolerance for it, but that means that your program, your project is really, really big and bloated if it takes that long to compile.
Robin: Yeah. Yeah. there’s probably more potential to be ahead in making forks of chromium that are lighter. Then there are in any sort of attempts at making something different.
Trevor: I guess
Robin Maybe do the hipster thing and fork from WebKit rather than chromium.
Trevor: I feel like that would be a good compromise because. Maintaining and improving gecko must be a huge drain on them because, you know, the web never stops. It’s like, Hey, we constantly demand more features. which to me is like, when is that going to stop? Are all web engines eventually going to end up this gigantic, bloated messes, because we want to make an operating system in browser. But doing that, forking WebKit or just using WebKit might be a good middle step where like, I like how relatively snappy the Firefox Java scripts and, um, interfaces. But I want the rendering consistency of WebKit. Because for all of its like pros, Firefox is rendering is a little bit weird.
Robin: Oh yeah
Robin: There was a, a lot of diversity inside of browsers, but by and large, there’s been a big centralization and no one’s had a lot of vision to do anything new. The closest I’ve seen for totally new things has been more brave, like the brave browser. it is a Chrome fork, but it has, some memory management going on. But now I doubt it’s better than edge
Trevor: Yeah, I feel weird hearing you say that. Firefox has gotten better with that because I’ve noticed that they do the thing where they don’t load the times if you haven’t. Um, so if, because I am a horrible tab person, I will relaunch my browser and I’ll have like five windows each with 10 tabs on them. Um, none of them will launch until I click on them. So it’ll actually like sit there and retrieve the whole webpage as if I had just gone to a new URL and opened a new tab, which is really nice because it means that the initial RAM usage is good. But it does make me wonder how much overhead there is underneath that. I can’t really see because memory is a weird thing.
That’s actually not that easy to track.
Robin: The battery life using edge on windows is really impressive compared to vector Chrome, but that’s purely a benefit of them kind of knowing how to cheat that.
Trevor: Right. wonder how good edge on Linux would be.
Robin: Uh, I have no idea. I know it’s good on Mac for what it’s worth. I would still use Safari instead. Because Safari gets those, you know, first party privileges on Mac.
Back to ARM!
Trevor: it’s not a bad segue because we’re talking about first party benefits and now we can go back to ARM.
Robin: I’ll put some chapters in here so that people can go back and forth. Obviously, arm’s been more in the mobile space for awhile and is very power efficient. I don’t think anyone thinks of ARM as fast, though.
It’s usually just doing simple tasks on the computer in your pocket, but it has been getting much faster over time. Apple and SBC have been very willing to create better dye processes to really ramp ARM.
Trevor: Okay. See that actually, that is really interesting to me because, that’s a really good idea, like investing better parts in it, but I do feel like, that could really speak to why Arum has been so perceived as cheap as, because they’re manufactured speed in these crappy little devices that are not supposed to use large amounts of battery.
In fact, they want to hammer it as much as possible. So you don’t burn your battery up.
Robin: It’s also why you can get a flagship Android phone and an iPhone, and the iPhone will have a smaller battery.
People will make fun of that for valid reasons. I think that would be kind of cool to have a little bit thicker iPhone with more battery. But it will still last a long time because the CPU is very efficient.
They’re super aggressive with sleeping applications in the background. If you’re an app I think you only get 5 or 10 seconds to store your state and exit before the manager will kill you.
Android likes to just let apps chill in the background, especially if they’re a system app. Of course if you’re on a Samsung half your apps are system apps.
Trevor: I hate that so much. I can’t express to you how much I hate it. It comes with Facebook. Pre-installed. Unfortunately pre-installed almost always means system app, which means you can only disable it and can’t remove it.
Robin: The funny thing is with Android, Facebook Lite exists. As an iOS user I would love to be able to get Facebook Lite. I’m envious of a lot of the lite apps on Android because Android sells these really crappy phones. Legitimately, it’s like someone drove over some garbage phone parts and one of these phones spat out costing $20.
Trevor: I remember back when I saw those, they were still using Android, like 2.3 0.4. Are those up to Kit-Kat now at least.
Robin: Google makes a version of the OS just for those phones called Android go
Trevor: Oh, I haven’t seen that
Robin: It’s normal Android but with basically everything stripped down to the bare bones, and it doesn’t even come with Gmail, it comes with Gmail Go.
And it’s basically just a progressive web app instead of an app, because they try to be as light as possible. photos go, which is the photos app. That’s almost all the same features, but weighs like a 10th of the size and uses a 10th of the memory.
Trevor: That just makes me think they should be using photos. Go everywhere.
Robin: can download with it. You can grab it for normal Android. This is what I was going to say. Um, if they, if this is why I don’t like the system’s app concept, because you’d be able to rip out camera, go, Grupo camera installed camera, go or rip out Gmail and sell Gmail go or Facebook light because these apps exist for really constrained devices.
But if you have a fast phone, they run like murder. because they take up nothing as far as efforts concerned for the larger, amount of memory and CPU you have. Yeah.
Trevor: Slight tangent. Can I ask you what runs like murder means?
Robin: I guess if you, if you murder someone, you want to run away, like that’s how fast go is.
Trevor: I just wanted to make sure you heard of any of those. And that actually makes me hate system apps even more because you just have no choice, especially you have no choice if you’re using like HTC and you’re not on one of the sanctioned phones that get the unlocks.
It makes me wonder if there’s like a gaps installer that has go in them now, because I know that the, when you flash your phone, you choose what version of gaps you want. And I’ve been using like micro and nano forever because; God, like I need Google photos on my phone, sucking up all my data.
Robin: So on the XNA developers forum, some real-time follow-up, they do have gaps go editions, that someone posted looks like they’ve been updating it since 2018. It was a few versions of it. And someone said, “some of us want the benefits of Android go faster for performance, faster speed and whatnot. Being frustrated, I decided to make my own”, um, includes, face lock, Google calendar, sync adapter, Google, Text to speech files, Google assistant go. Oh yeah. Google assistant go was a thing too. Go. All that fun stuff.
Trevor: I’m really curious about Google Assistant Go because I really hate how creepy Google assistant is, but I still want it sometimes if I’m like in the shower and I hate the song I’m listening to.
Robin: Yeah, there was one night when I was really not able to sleep all that well. And I was looking at YouTube videos of people testing Go, and they were using Go on these really crappy phones that had almost no power.
Trevor: Huh. Okay. That’s really interesting. I would love if that was made open source because I would love really, really small Android that doesn’t burn the crap out of my battery. That is, yes. That.
Okay. So another thing that I really hate about system apps, and I guess just Android and Google in general. The whole Google play framework is the worst. It is so heavyweight, Google play store just turns on whenever it wants and it decides, Hey, I’m going to destroy your battery now. And there’s nothing you can do about it except to come to my wins. Hey, you have updates. Did you know that you have updates?
I’m going to wake up every five minutes and tell you about!
Robin: Yeah. as an iOS user, I don’t get to experience any of that framework pain… but I get my own special sorts of pain. So I don’t see either platform as like, particularly amazing, but. I love hearing Android complaints. You just complained about things that like, I can’t really conceive of that happening in iOS. Like a background app is launching on its own?!
Trevor: That is something that I really admire about the iOS ecosystem, is that they have such tight control over what goes on on your phone. That. They can force you to be super efficient, that’s really a positive point of that. But at the same time, that comes with the downsides of you have to pay for the developer license.
You are completely subject to their terms. And if you want to install your own apps will screw you. You have to jailbreak there’s no, just installing unsigned apps.
Robin: Yeah. But there also aren’t like ads that will just randomly try and download an APK to your phone. When you do that.
Trevor: Oh yeah. Um, in fairness though, I’ve never had that happen.
Robin: Yeah, when I used to use, so I used to use a nexus 4, and that was unfortunately a somewhat common thing.
Trevor: So that’s, that’s quite a while ago.
Robin: Yep. It still runs lineageOS because the nexus 4 was unkillable.
Trevor: Oh my God.
Robin: It has NFC. It has wireless Qi charging.
Trevor: My current phone doesn’t have Qi charging.
Which I’m actually confused about because I, okay. So I have a one plus seven pro it’s a pro, let me tell you it doesn’t have wireless charging. Not that I care that much because I do just plug it in every night and the battery life is fantastic and it lasts me all day, but you’d be really nice to have.
Robin: Yeah. The Nexus four doesn’t have LTE, it’s that old, but it has Qi charging and it supports Android pay. And like, it still has enough memory and processor to run,
Trevor: Uh, yeah, I feel like there was definitely a heyday of Android phones. That’s kind of. I definitely would not trust any of the big flagship companies anymore to buy one of their phones. Like the last Samsung I got, I was extremely disappointed. Every person I’ve talked to has gotten like the latest nexus, pixel, whatever.
They’ve been disappointed for the price they paid.
Robin: The pixels have not been good phones.
Trevor: but my next seven, my next seven was something that I used for years until the port broke because micro USB is
Robin: yeah, could probably still repair it.
Trevor: I tried
Robin: Oh, okay. Nevermind.
Trevor: I broke it more when I did
Robin: But I was going to say, let’s talk about the people whose fault this is for phones kind of sucking.
Not really because the, the Android operating system has flaws. Sure. But it’s not necessarily why Android phones, aren’t making great strides as far as performance to battery ratio. I feel the main reason is just putting so much pricing pressure on the CPUs being put out by Qualcomm and Broadcom.
They, so they’re the ones that typically make a lot of the kinds of processors that are ARM-based.
Robin: I think it was a Qualcomm that made the Nintendo switch CPU.
Trevor: Um, so the switch is basically an Nvidia shield.
You know how much it’s basically an Nvidia shield is that all the debugging stuff is still there and that’s how you hack it.
Yup. I’ve considered doing that to my switch because I don’t know, because it would be fun. I just think it’s absolutely hilarious how you do it because you can just CHAM a paperclip into the right joy con to bridge some pins and that’ll turn on debugging mode.
Robin: Right. Yeah, I think, um, I think for Tegra, I think Nvidia makes their own, or at least, I don’t know if they’re the ones that print them or fab them, but I think in the video designed them themselves, it appears that way. The first product to use the Tegra processor was Microsoft Zune, HD media player. Oh, and the Tesla Instrument cluster.
Tesla’s Instrument Cluster Woes
Trevor: Wait what?
Robin: Touch screen and instrument cluster. Yeah.
Trevor: But why. Okay. What resolution is the Tesla cluster?
Robin: I’m not sure. All I do know is that the highway safety administration believes that 20 to 30% of all Teslas will fail within three years. But primarily because the memory was written wrong, they use a really like low end consumer mobile phone storage. And then write Sysd logs to it.
Trevor: So that means there’s a compound failure here, which means: One thats really garbage flash memory for a car. Two, they wrote system logs to the garbage flash memory. Three. They, as I read, they used IEX T3 for their garbage flash
Trevor: please. No, please don’t please. I don’t
Robin: The good news is that soon they will fix it. Which mean… it’s not been fixed yet. Uh, but there was a note that they were planning on using like proper caching techniques to not write things directly to the flash. Yeah.
Trevor: How are they not!? That’s built into every kernel with everything.
Robin: They’re innovation too hard for that, I guess.
Trevor: turned off the…
Robin: Innovating so hard that they just couldn’t couldn’t do it. I remember back in like 97, 98, maybe Noppix. The, the Linux system that would run off a CD rom, Kappic? Noxix? It had memory level cache for disc buffering, because you would run COPICS off a CD or off a, a flash drive and they wanted the flash drive to run as long as possible.
Trevor: I would like to know when this is, because now everything can run off a flash drive
Robin: Yeah, but this was back when that wasn’t the main thing.
Trevor: Wait what?
Robin: It was Debian when it came out I think.
Trevor: Oh. Oh, Oh, okay. I didn’t understand because of your pronunciation. KNOPPIX um, I don’t know. It’s K N O P P I X.
Robin: I think I was making it a hard K because of KDE…
Robin: Yeah. I remember, Linux format magazine used to ship with the KNOPPIX CD attached to the magazine.
Trevor: When you said it the first time I actually thought it was spelled like, KOPIX or something like a kenotic jar.
Robin: I totally butchered it. I remember the KNOPPIX, and the Damn Small Linux days, you know.
Trevor: that that was really fun. Back when I had, when my largest flash drive was a gigabyte.
Yeah. I did this for my Bitcoin miners. Back when I had those, I’ve ran my next off of flash drive properly, made them not burn themselves out. So I would love to know why they disabled, right. Caching, because that’s, because why would you right. That’s the thing that I’ve really been said about with self-driving cars, like, why are you taking the half measure of you must pay attention. There’s no way anybody’s going to pay attention to the car is doing everything for them. You expect them to monitor the car.
Robin: So I have, uh, like level one or something. Self-driving with the most basic version of it for, for my Jeep. And all it does is lane keeping and smart cruise control and that’s enough to do automatic breaking and that gets to a certain level.
Trevor: That’s a half-step that I can respect because you have to drive the car still.
Robin: Yeah, it will tell you eventually like, make sure your hands are on the wheel. It’s like they are, but I’ve go to issue commands because I’ve not moved them in while.
Trevor: feel like that’s the crappy part of that half step is like, you’re going to nag me, huh? Okay. Sure.
Robin: So, yeah, there are certain stretches of straight road where like halfway through, it’s like, please pay attention to the road. Anyway…
Can We Have Better ARM Chips? Plz?
but basically because of the driving pressure is more towards costs than performance and a lot of the Android, the flagships not withstanding, a lot of the actual production of processors for ARM tend to not be super powerful. So when you do find they’re really powerful versions, they end up in things like the, Microsoft Surface devices that have arm.
Trevor: Oh, well, okay. If that’s the standard.
Robin: The Switch is actually quite powerful. The switch and a surface would not be that different. The difference kind of being that there’s more video cores.
Trevor: Also it just has straight up has two CPUs.
Robin: Yes. Well,
Trevor: It is, it is completely two because one of them is ARM seven.
Yes. Um, whenever you press the home button, that is the ARM 7. So I’m probably getting this wrong. One of them is a lower version and one of them is ARM 11. Um, the lower version, which is something that you see on lots of old crappy Android phones is the thing that runs the home menu UI. This is also why some games, Splatoon, cough, cough are not coded very well to communicate with the home menu.
So if you press the home menu button and you’re doing literally anything in Splatoon, it will just say connection interrupted and kick you back to the Plaza. It’s very annoying.
Uh, I guess so. Yeah, I mean, I was really impressed with what this which has done overall. Like breath of the wild was gorgeous. They got the Witcher three to run on it with some advanced compression stuff going on. Um, you know, it’s not as amazing as a computer could do, but still the fact that they got that powerhouse of a game running on the switch is.
Robin: Yeah. And I know that when people started seeing the latest pixels, also the iPad pro and things like that, they were running benchmark scores on them and started realizing that these things were getting close to what the average PC would have. Like why can this device run for 10 hours beat my laptop that runs for 2 and run just as fast in benchmarks. Some questions started getting asked around that.
There’s not really anything that prevents any company from licensing arm or the companies that already have like Qualcomm Broadcom, and Nvidia, uh, that really owns ARM now. So they’re extra in this realm, but, nothing prevents them from doing the same thing Apple and SBC did and going to a smaller nanometer process.
And, uh, I think Apple M1 has four fast cores and two slow cores, which is more than the iPad pro if memory serves.
Trevor: I think it’s really, uh, an interesting approach in particular for the fast and slow cores, because like we’ve talked about with the switch, that’s a really interesting process and it makes sure that certain experiences are super seamless. Um, the funny thing is that I hope we don’t do is go too far.
Like the PS three, where you have three different architectures and six weird cores, and no one can develop for the thing because no one knows all the three different architectures and the thing is a barbecue.
Robin: I believe, the cores are managed by the OS, so there’s no separate process or architecture between the two. It’s just, some of them are faster than others. The OS shunts apps off where it makes sense.
This doesn’t prevent anyone including like Microsoft cough, cough from trying to outdo Apple here and do something bigger and better. The impressive thing is just how fast the M one is compared to previous things.
In fact, they refreshed some of their products recently. To include the new Intel chips, between cycles.
Trevor: I would love, I would like to know why they did that though, because from what I’ve seen, that the latest Intel’s just really, hasn’t been an improvement ever since Rizen came out and they’ve been like, Oh God, we can’t do hyper-threading anymore because we suck. Why bother refreshing.
Robin: I think it was just a matter of packing in more transistors. Uh, yeah. So there are some comparisons between the M1 and some AMD processors, which have better, uh, floating point number handling, I guess, in Intel. and that was one place where certain AMD processors could beat the M1 but they were using 49 Watts of power.
And the M1 is seven to eight Watts. So dramatically more power through that same operation.
Trevor: Oh, yeah. And I mean, AMD already is pretty impressive with their TDP. Like 49. Watts is really good in the world of CPS compared to like 75 or a hundred for Intel.
Trevor: I’m probably, I’m sure I’m getting that wrong. My brain is saying 175 watts and I’m certain that’s wrong. Maybe that’s like an i9 or something,
New Macs are Really Good, Not Great
Robin: That’s pretty intense. The thing that I find interesting, and the Apple’s kind of already admitted in some of their interviews is this is kind of a crappy version of all the Apple Silicon, too. the Mac mini that was released the, MacBook air and the MacBook pro, uh, this is replacing the lowest version MacBook pro the one that used to be sold within a touch bar.
So like the, the low end Mac book pro is the one they were targeting here. All three devices have effectively the same motherboard and CPU layout. so it’s the same kind of device into three different shells. the overall the performance characteristics are kind of weird in that the Mac mini supports having one external 6k display over display port and one external 4k display over HTMI, which matches what the Intel could do.
But on the Intel, you could also just have three 4k displays and the M1 can’t do that. You can only ever have two displays, but one can be up to six K. Because of limitations I’m assuming in how, displayed processing works in the arm stack. something they’ll have to definitely fix before they go into the higher end machines.
If you get an iMac pro or something, you can run two additional 5k displays, I think, off that sucker and people would be unhappy if that, if you couldn’t do that.
Robin: and the also the Mac book, air and MacBook pro that released they each only have two USB-C ports, which is the same as the lowend Macbook Pro used to have, but you can also only have one external monitor.
Trevor: Oh, yeah. That would make sense. I, man, after having to use a MacBook for work so much, I’m really missing just having more USB ports like on the Thinkpad that I’m actually running. Now, I admit to, it’s a huge pain. I have to plug in three different ports or whatever. But I don’t have to worry about the stupid little dongle that might overheat or might fall off the desk or whatever.
Or if it won’t fall off the desk, it will be really big and require its own power supply. So I don’t like the, for all that I like USBC. I do not like dongle world.
Robin: It’s only got 2 USB-C ports.
Trevor: Well, at least with three, you could do something like you got an ASC adapter for a KVM, you get two USB-C monitors. And one of the monitors has power in that might be all you need compared to, Oh, we’re a Mac. We have one USB ports
Robin: But still it’s not great. The, the one complaint I see, as well as like why they’re both on the same side. which is interesting. there’s a good reason for that from a motherboard location perspective, because you don’t want to have like an entire board going the whole way across the computer.
And also if you have one of those nicer non-thunderbolt docks, they typically use both USB-C ports.
Um, yeah, you can get them that plug into both of them and then downstream, you can put in multiple monitors and things.
Trevor: If only docks stopped being like 150 to $200,
Robin: yeah, and the, the good Thunderbolt three docs are easily 200 or 300.
Why are USB-C/Thunderbolt 3 Hubs so Darn Expensive?
Trevor: So I’d like to ask you. why are these things so expensive? Because even the switch docs are like $60, even though it’s just a USB-C port with three yeah. Three USB ports, one HTMI a power in and that’s it. And the rest of it is just plastic.
Robin: So there is a component you can get for USB-C, this pre-packaged ready to go component and with USB-PD, HDMI and a USB-3 header.
Trevor: So you’re talking about as far as manufacturing, this would be the ports that the, the parts they could commonly
Robin: Right. So my assumption is they’re grabbing this chip, putting a USB, a two hub in front of the USB header, putting one of those into a, uh, an audio. USB device, giving you the other couple, two or three, whatever you mentioned there as outputs in the front of it. So just a normal four port hub, and just one of the ports has an audio device crammed into it on, on the PCB for power throughput. So it sounds like you’re using that. There’s a couple of versions of that chips. They often have a USB-A, USB-C and an HDMI on it.
It’s why so many of them follow that layout. and it’s just because it’s really cheap to manufacture. it took a long time before we got more than that for the Thunderbolt because you had to have license from Intel
Trevor: so you’re saying the majority of manufacturers. Okay. This explains a lot. It all comes down to licensing again,
Robin: there is effectively a Thunderbolt controller, whatever you want to call it. But yeah, there is this downstream thing, which is taking that PCI express lane from the TB3 connection and, and doing its bullshit with it.
Trevor: HWhy, why are there not been competitors until now?
Robin: it’s like when you get a graphics card and you have to wait for EVGA or whoever to make a card based on the newest Nvidia or ATI reference.
Trevor: This would explain why I see such a big Gulf between like, The adapters, the adapters that I tend to get, they’re like 40 bucks, which I’ve very few ports and it’s extremely hard to get more ports on them. And they generally don’t have great bandwidth. And then there’s like a $50 jump. And then suddenly you get three mini display ports and a display port and blah, blah, blah, blah, blah.
Robin: the $300 one I linked the ports actually list the differences between if it’s plugged into USB-C 3.1 or Thunderbolt.
Trevor: Yeah, I see that. They explain it, compared to most things.
Robin: Yeah. And you can also compare the video resolutions. So this is the reason why devices for USBC hubs tend to suck, the cost of the actual, uh, chip set or whatever you want to call it. I’m thinking I’m using the wrong word with chip set and probably going to get feedback.
Trevor: I think that’s close enough.
Robin: but yeah, the cost of it is quite high to get a good one. And the cheaper ones that are kind of knockoffs of it tend to have a high failure rate or do weird things where like they’ll stop working or you have to unplug and plug them back in constantly.
This is very much Intel’s fault. Yeah. Looking around. It does appear that they are the only ones making the actual controllers
Trevor: I wish the chip fabbing, wasn’t such a horrendous business to get into because you really need. A lot of capital, because there’s not a lot of competition. And as we’ve seen with the huge shortages that happens, I guess like two, three years ago now with Ram, if one of those businesses goes down the entire industry, like implodes
Robin: I’m not sure there’s a business incentive for you to start fabbing your own stuff. So you’re going to use one of these off the shelves. You’re gonna choose a price point. You want to make a $30 USBC dongle. So you’re gonna, you’re gonna grab one of the cheap USBC hubs, and you’re gonna use that and it’s going to overheat and break on you.
Trevor: And then like, it’s just, it’s so not easy to get into because you have to set up factories of your own and they all have extremely expensive equipment. You need extremely talented people to run them you need engineers to even design the blueprints compared to something like. I don’t know, this is a terrible comparison, but comparing let’s run cloud servers versus let’s run our own virtualized servers.
That’s an even easier case to make them let’s fab our own chips. So that’s partly why Apple and SBC doing what they’re doing is insane.
Robin: Yeah. And the only reason they got to this point is because they’re able to have enough overlap because they’re producing stupid amounts of iPhones and kind of interesting amounts of iPads. And honestly, not that many desktops, the laptop desktop business for Apple has not been much of their margin for quite a while.
Most of their money is made through, the phone and through services like the app store. but yeah, because they’re already kind of doing this for one device and there’s a lot of overlap for them to expand the line.
Trevor: This makes me wonder this will probably never happen, but. Every now, and then I get a little bit of a wind of like some little sub community. That’s like we designed a new chip architecture. That’s theoretically like 5,000 times faster, whatever. And it makes me wonder about the potential of things like that.
That will never happen because it takes a giant on the scale of Apple who just has this much money sitting around to say, you know, it’d be cool if we made our own chips.
Robin: Yeah. At one point, someone did the math and Apple had enough cash on hand to buy the biggest us movie studio in cash.
Robin: Um, So, I don’t think that’s the case anymore because they’ve made acquisitions and things. But they still have like a crazy amount of money for things like this.
And Nvidia has never been shown to be good at like innovation. Um, it would require them to maybe develop the skill, but if they were to realize the potential and the marketplace now sees that, you know, Intel not doing that great. if they were to go through and adapt their knowledge to making..
With their video knowledge and with ARM and they were to go through and make a couple of high-end APUs and sell them to other systems. So like a Dell or a Microsoft could then use that chipset to build a computer around it. That would have to be the direction you have to go and you’d have to have a company like that.
It really requires someone like Nvidia. Or, Amazon might decide, you know what? We don’t like paying Intel for processors for AWS. They’ve already started going in this direction a little bit, but they could go a whole hog and say, We’re making all of our own processors for our own hardware for AWS by default.
Trevor: No, when I saw how quickly AWS adapted to how Intel like messed everything up repeatedly, Just seeing that, like they had their EC2 instances, like immediately, and they had ARM instances, like immediately, that was crazy.
Robin: They could just move it all to arm. And Intel becomes a specialty thing.
Trevor: That would be interesting. I mean, they definitely would have to keep some of it around to be like, you know, our, our software was made 20 years ago and we can’t get it to compile on arm because no one makes it anymore. Please give us Intel.
Robin: Yeah. And I think that would be still doable, but I can definitely see all of their services running fully on arm. do we want to start wrapping up here? Anything else you want us to cover right now?
Trevor: I think we should start wrapping up partly because of this took a little longer than I expected, but I have really enjoyed it. it’s been nice. Just kind of. Bullshitting about technology
Robin: Well, yeah,
Trevor: I don’t really, I don’t really get to do that a lot. And also pandemic. Yeah.
Robin: This is our, our Alpha technology preview version. I’m sure we will leave it more organized and used to this in the future.
Trevor: Ooh boy. Here we go.