How frog gets cooked in the cloud

How much computer power do we actually need for normal tasks? Does the difference in computational power influence the end-result? Can you tell a difference between an article written on a desktop or a laptop? The last question sounds incredibly silly, I know, and yet when I watch the tech YouTube videos there’s an impression that if you’re a “real pro” or a “power user”, you’ll need “MO PAWAH”. The poor-people tech made for the plebs just won’t cut it, you’ll need the shiny new thingy to keep up with the times. Only the 7nm node size will cut it.

Several things happened recently. First, a new Intel bug was discovered, possibly rendering modern Intel CPU machines vulnerable to attack unless you cripple the CPU by disabling almost everything on it. Second, America embargoed China by limiting access to all kinds of software and hardware technologies, from Android and Windows to x86 and ARM. If we add that to things that are already known, such as the Intel kill switch, and all kinds of technologies that make it theoretically possible for the manufacturer to brick the motherboard of your device remotely, on a low-level of access through the onboard networking hardware, BIOS and the chipset, because America put you on a list of “sanctioned” individuals, for whatever reason.

Microsoft is introducing a “politically correct” spelling-checker into Word. Online censorship is rampant. Witch hunts are out of control. I can easily imagine some AI identifying “politically incorrect” people online, through their cloud service logins, and I can easily imagine hardware and software manufacturers full of “social justice warriors” performing acts of “social activism”, for instance triggering a “stolen device kill switch” on your motherboard remotely if you write too much “right wing” or “racist” content online. If you think this is paranoid, imagine being Snowden or Assange, and imagine what can be done to their computers if they are identified remotely, and if it’s done by someone really powerful, like NSA, or Google, or Microsoft. Now imagine this being automated, delegated to an AI system that will check your login against a list, and then simply “deplatform” you by bricking your PC, because after all, Nazis can’t be allowed to speak.

All of this made me think: what would I do if I was targeted by something like that? Using a web browser made by a huge corporation is a vulnerability. Using cloud services is a vulnerability. Using an operating system made by a company that’s BFF with NSA is a vulnerability. Using Intel, and possibly even AMD CPU is a vulnerability. Using a motherboard with a chipset and a BIOS that isn’t made transparently is a vulnerability. So, if someone decided to brick my computers that run Windows and Mac OS on Intel, and my iPhone and iPad stop working, or at least stop connecting to the Internet and accepting my login into Apple services, what would I use to get online?

It turned out that I have one machine that is most likely to remain working: a Raspberry Pi 3B+ that I have under my desk running Linux, a machine I manually hardened and which runs 24/7 hosting mysql, ssh and apache. However, that’s not all. It also runs a LXDE GUI, with a complement of Office tools. But this is an extremely weak machine. Its CPU is a rounding error between two geekbench measurements of my main desktop PC, and I’m not even exaggerating much. Its “disk drive” is a micro SD card, and the entire computer can fit on my palm. However, there’s a catch. It is basically Android smartphone hardware converted to serve a different purpose and run a different OS. People use Android smartphones to do things online every day and don’t give it a second thought. But can you plug a smartphone board into a monitor, keyboard and mouse, run Linux and do normal tasks, like researching things online, taking screenshots, writing and article in OpenOffice, logging into a CMS and posting the article on your blog? Yes, you can.

In fact, it turns out that this small tiny computer is more powerful than the machines I used to write most of my books on. And I edited them in OpenOffice, printed them as PDF, and then used Linux command line tools to split the PDF into PNG images of individual pages, and then publish those on my website in the online reader form. I did all that on an IBM T43 laptop, which was less powerful than this Raspberry Pi thingy. Of course you can do it, and in fact that’s how I wrote this article; I connected the Raspberry Pi instead of my desktop computer, and used it to drive my usual peripherals. It doesn’t feel slower when you write the document; you can do most things just fine. I used computers with less power and memory for most of my career, because that’s what we had then. It’s actually quite smooth; I installed Gimp from the terminal while writing this article and not even a hiccough. Then I used Gimp to crop a screenshot and save it. It did it just fine. I just got used to computers that do the same things faster, that’s all. Using this thing didn’t degrade me into stone age. I could even plug my external HDD into it and process raw photos from my camera if I had to. I would use dcraw, rawtherapee and gimp instead of lightroom, the way I did for years, and guess what, you wouldn’t be able to tell the difference, because I did it exactly that way for five years and nobody could tell the difference between that and lightroom anyway. I just got suckered into using tools for lazy people, tools that make it very easy, but that don’t actually do anything I couldn’t do manually with some more effort. I could also do just fine without the online cloud services, and guess how I know? Because I was here before they were. I was on the Internet and finding my way around quite well before Google was a twinkle in its authors’ eyes. Some of those tools made things easier, but the price might be too great. Ease and comfort, apparently, can be weaponized as a vector of attack. You make it easy for people to access the same file from several devices and they don’t stop to think that their files are stored on someone else’s computer in unencrypted form. You make it easy for people to connect to other people online and they don’t stop to think and realize that their entire social life is now owned by a company whose primary motive is to sell you to the advertisers, and to control the entire experience as to be more presentable to the advertisers. Also, that they hire fuckwits who studied feminism and social justice and who want to change the world to be more like an American college: meaning, that it requires less thinking, more feeling good about yourself, and excluding everything that gets in the way of feeling good and not having to do any thinking.

However, someone bricking your PC might actually be a lesser concern. A greater concern might be someone blocking your Visa card because you’re on some political list. Also, the banks might not allow you to open an account. You might not be able to get a loan for a house or a car despite your stellar credit rating. Police might track your whereabouts using your phone, because you’re on a list of “extremists”. You might be stopped from boarding a plane. You might be taken off a plane in an islamic country that has you on some shitlist, because you criticised Islam online. Those threats are actually more real, and I’m actually not making those up; that shit actually happens now, as we speak. It’s just far less common than it could be, once the technology proliferates. So, sure, I used a PC made from a phone chip to write an article on the Web, big deal. I can maintain the same kind of online presence with rudimentary technology, and nobody would notice the difference. However, that proves one interesting point: that the advancement of technology in the last two decades was actually much less drammatic than one would think. We just got used to the fat and expensive tools that do basically the same job as the old lightweight free ones. Also, it means that America can cut the rest of the world from their technology, and the rest of the world could do just fine with Raspberry Pi boards made in China for $1, and they would actually be forced to get more creative with resources and possibly find better ways of doing things. Being reduced to simpler computers wouldn’t actually degrade life much, because faster and better computers didn’t upgrade it much. They just made it easier for stupid and incompetent people to think they are advanced, smart, trendy and techy, while in reality they are just stupid consumers.

So, what am I going to do now; continue using Raspberry Pi as my main PC? Hell no. It can only display a 1080p image on my 4K monitor, which makes everything blurry. Also, I have to pay attention to memory use because it only has one gig of ram, and so on. But I know one thing. If America does cut me off from American technology, I will find whatever piece of junk that runs Linux and connect to Russian-Chinese Internet, and I will do just fine. I used to write code on a potato when Web was an experiment on Tim Berners-Lee’s Next cube, I wrote books on computers that couldn’t walk and fart at the same time, and I can do it again if necessary. The only thing that’s actually scary for me is thinking how easy it was for me to get used to the idea of giving up privacy and security just to make things a tiny bit easier and more comfortable. Because of this, I might actually start intentionally giving up various online service that make things unnecessarily easy, but at a hidden cost. I will also give Linux a second chance.

However, if that is scary to me, there’s another thing that should be scary to the Americans, and that’s the idea of a smart person that’s comfortable using Linux tools on a Raspberry Pi instead of a Macbook. Because that person might understand that he can do just fine without all sorts of things that make him a slave. For instance, he might understand that the AGC computer that got people to the Moon was computationally much weaker than the toy I’m writing this article on, and that St. Augustine and Isaac Newton used ink and parchment.

10 thoughts on “How frog gets cooked in the cloud

    • Yeah, I’ll get it and try to install KDE on Raspbian, that might be fun. 🙂 Although, to be honest, the default LXDE is usable in a pinch, there’s nothing really wrong with it other than being visually simplistic, and the main problem of the 3B+ is speed (or lack thereof), RAM (or lack thereof) and 4k display (or lack thereof). 🙂

      • I was actually using RasPi 2 as a “main pc” for a short while and it was actually usable (default Raspbian). Slow, sure, but usable.
        But when I read 4th gen can output dual 4K streams, have bunch of ports can use 4gb RAM … this is serious. For 55$ in “top” configuration. I had to remember Apple at this point … sure, there is much more to Apple, but still. Apple sells rubbery watch band for more than that ffs.

        Anyway, shouldn’t this handle ANY Linux distro, making Raspbian just one option?

        • Anyway, shouldn’t this handle ANY Linux distro, making Raspbian just one option?

          Well, it has to be built from source for ARM, at least, and probably has to be made specifically for the hardware, but basically if it’s not on this list I wouldn’t try it: https://www.raspberrypi.org/downloads/
          I tried to install ubuntu mate but it was built for the older ARM variant and so didn’t work, and nothing else was interesting so I just installed mate-desktop-environment in raspbian, but that conflicted with some lxde daemon and got stuck until timeout every now and then so I removed it.

          • Apparently they fixed the ubuntu-mate build so it works on 3B+ now, which is good to know but I don’t care much about installing it now since I already configured all the services and I don’t want to do it again just for shits and giggles.

        • I ordered the Pi 4B with 4GB RAM today, I’m quite curious how it will do. In the meantime I installed KDE desktop on Pi 3B+. Let’s just say it looks like trying to run Windows 95 on a 386. That is to say, not fast. 🙂

          • Should be interesting.
            I think I will make something useful out of my RasPi 2, just for the sake of it 🙂
            I will also order 4B, I want to see how it can handle some of the server loads and how much of it exactly.

            • 3B+ is very nice and responsive as a LAN server for mysql; I’d have no problem with something like that for web/mail server online. Incredible for something that cheap. I’m deliberately testing the cheap options because the technology went so far that even those can be good. That a NUC is fast is obvious, but it’s also expensive. If a raspberry pi can do the job for a tenth of the price, it’s an awesome option. And you don’t always need a supercomputer for every job. For instance, having something that runs daily crontab sync jobs on the LAN, and also hosts my mysql mail archive is useful, and such tasks would be a waste of a NUC. But using it as a Linux desktop might be pushing it. I’m certainly curious.

              • Ok, so I am not completely crazy with this idea of replacing one huge x86 server with an army of small ARM devices.
                Right now, big server is running tons of virtual machines each dedicated to rather simple task.
                All of these require powerful hardware, but chopped to smaller jobs, I think 4 could do it.
                It’s not like I run Facebook or something and since RasPi 2 was already good for simple web and file server, I imagine 4 will be able to do anything, maybe even handling main web server which serves 100 clients with web api (not heavy workload).
                I can’t wait when I come to a client with two RasPi 4 and say – here, these are your main database and web servers.
                I can put 100% profit margin on them and still be dirty cheap 🙂

                • The main problem that I see is actually wearing out the SD card in the Pi. Those things are not nearly as well made as SSDs, and if you put a database on it, one with many updates, the card is going to wear out quickly. So I did a test on the 3B+ and copied the card to an SSD, and it works, the Pi boots from it just fine. So, with 4B and its USB3 ports, the speed would actually be quite good.

Leave a Reply