Hackaday Links: January 10, 2021

You know that feeling when your previously niche hobby goes mainstream, and suddenly you’re not interested in it anymore because it was once quirky and weird but now it’s trendy and all the newcomers are going to come in and ruin it? That just happened to retrocomputing. The article is pretty standard New York Times fare, and gives a bit of attention to the usual suspects of retrocomputing, like Amiga, Atari, and the Holy Grail search for an original Apple I. There’s little technically interesting in it, but we figured that we should probably note it since prices for retrocomputing gear are likely to go up soon. Buy ’em while you can.

Remember the video of the dancing Boston Dynamics robots? We actually had intended to cover that in Links last week, but Editor-in-Chief Mike Szczys beat us to the punch, in an article that garnered a host of surprisingly negative comments. Yes, we understand that this was just showboating, and that the robots were just following a set of preprogrammed routines. Some commenters derided that as not dancing, which we find confusing since human dancing is just following preprogrammed routines. Nevertheless, IEEE Spectrum had an interview this week with Boston Dynamics’ VP of Engineering talking about how the robot dance was put together. There’s a fair amount of doublespeak and couched terms, likely to protect BD’s intellectual property, but it’s still an interesting read. The take-home message is that despite some commenters’ assertions, the routines were apparently not just motion-captured from human dancers, but put together from a suite of moves Atlas, Spot, and Handle had already been trained on. That and the fact that BD worked with a human choreographer to work out the routines.

Looks like 2021 is already trying to give 2020 a run for its money, at least in the marketplace of crazy ideas. The story, released in Guitar World of all places, goes that some conspiracy-minded people in Italy started sharing around a schematic of what they purported to be the “5G chip” that’s supposedly included in the SARS-CoV-2 vaccine. The reason Guitar World picked it up is that eagle-eyed guitar gear collectors noticed that the schematic was actually that of the Boss MetalZone-2 effects pedal, complete with a section labeled “5G Freq.” That was apparently enough to trigger someone, and to ignore the op-amps, potentiometers, and 1/4″ phone jacks on the rest of the schematic. All of which would certainly smart going into the arm, no doubt, but seriously, if it could make us shred like this, we wouldn’t mind getting shot up with it.

Remember the first time you saw a Kindle with an e-ink display? The thing was amazing — the clarity and fine detail of the characters were unlike anything possible with an LCD or CRT display, and the fact that the display stayed on while the reader was off was a little mind-blowing at the time. Since then, e-ink technology has come considerably down market, commoditized to the point where they can be used for price tags on store shelves. But now it looks like they’re scaling up to desktop display sizes, with the announcement of a 25.3″ desktop e-ink monitor by Dasung. Dubbed the Paperlike 253, the 3200 x 1800 pixel display will be able to show 16 shades of gray with no backlighting. The videos of the monitor in action are pretty low resolution, so it’s hard to say what the refresh rate will be, but given the technology it’s going to be limited. This might be a great option as a second or third monitor for those who can work with the low refresh rate and don’t want an LCD monitor backlight blasting them in the face all day.

And finally, if it feels like time has been flying lately, you’re right. The Earth’s rotation on its axis has been shortening lately, such that days are spinning by not in a glacial 24 hours, but at a frenetic pace of 23 hours, 59 minutes, and 59.9998927 seconds. Laugh if you will, but those microseconds add up over the year. Leap seconds have been applied at least 27 times in the last 50 years to correct for Earth’s slowing rotation, but we’re now in a regime that may require a “negative leap second” be applied. Scientists will be studying the trends in Earth’s rotation and make a decision at the World Radiocommunication Conference in 2023, so our 24-hour day is safe until then.

36 thoughts on “Hackaday Links: January 10, 2021

  1. “You know that feeling when your previously niche hobby goes mainstream, and suddenly you’re not interested in it anymore because it was once quirky and weird but now it’s trendy and all the newcomers are going to come in and ruin it? ”

    Wow, you just described my life.

    1. Let’s see, vinyl is ruined, craft beer is ruined, vintage computers, just ruined… ah.. “Hey man, check out my vintage StarTAC..”

      It’s a crying shame I tell you, I’ll just have to dump my collection at highly inflated prices, shamefully high prices, so high that I almost feel bad for not throwing in a mousemat with my phone number on in case their friends want to buy one of the others.

    2. I think a subtle but important distinction was left out of that description: “Hobby” vs “collecting”

      Collecting is trivially ruined all the time this way, as ultimately anyone with more time and money is going to out-collect those with less. This isn’t to say collecting isn’t a hobby or anything, just that if your stature to others in your hobby is directly proportional to your wealth, you just need to accept someone else will have a larger and more complete collection than you, no matter what. If that’s the game, you lost before you started.

      Outside of the collecting hobby, I don’t think this is a universal truth.
      I wouldn’t mind at all sharing my knowledge and experience from growing up in the era with newcomers.
      And I love sharing *new* experiences I had once thought would no longer be possible for an era I believed to be long past.

      Sure, more demand when there is a limited and no longer produced supply can suck. But look what that did to the c64 community, it resulted in others taking up the mantle and manufacturing new replacement parts, from enclosures down to PBCs, to making new accessories to bridge the ages, such as a modem-emulator that connects over wifi and the Internet.

    3. I know the feeling.
      On the bright side, some of the systems that were out of my reach when I was a kid (because scrappers would pay more for them to extract the gold from) will be saved instead of destroyed and melted down for scrap as long as this ‘fad’ lasts.

      Thankfully for better or for worse, I still have a ton of old microcomputers I saved from the heap or collected as a kid. Now the challenge is to organize the mess, figure out what needs work, figure out what’s beyond saving, do the work on them, and display them whenever possible for people who wish to have a go or play with them, and keep enough that I can give away items to people who really express interest, but the market being silly has passed them by as being an affordable thing to play with.
      COVID-19 has really put a dent in my effort to do those things.
      An old computer is best experienced beat up, cranky, with flaws, and at your disposal to do silly things with instead of a no touch perfection museum exhibit.

      I think it’s neat that people think these pieces of junk are cool even though it now limits my ability to get spare parts monetarily.

      No one’s making new actual SID/VIC chips, 8088’s, unibus ethernet cards, AppleIIGS motherboards, 5×86 133’s, FDIV bugged pentium 75’s, 68040 mac quadras, athlon 550’s or MFM hard drive spinny disks, so if as much of the current supply as there is survives the garbage heap, that’s a good thing.

      Also documentation I could have only dreamed of when I started wanting my own computer and building them up from people’s discarded junk is now available, as well as alternate methods of producing programming and loading files into said machines. I still remember a time of worrying about the difference between 720KB double density 3 1/2″ disks and the high density 1.44MB, use the 1.44MB to write things on that 720KB floppy and depending on the exact scenario and how forgiving that 720KB drive is you are gonna have a very bad day!

      It’s really fun when you realize that the single board computer you have emulating something like an MFM hard drive is hundreds of times more powerful than the computer you have it hooked up to.
      (Shoutout to David Gesswein and his excellent MFM emulator! https://www.pdp8.net/mfm/ )

      I for one welcome our crazy collector overlords, may they enjoy at the least collecting these crazy pieces of history, or more hopefully may them enjoy fooling around with them and learning something from them instead!

  2. Re: Apple 1: The ONLY distinction the Apple 1 has over the ][, is that there were fewer of them built. This is just silly. And as for the big Ataris, 68k Macintoshes, and Amigas, these might as well just be under-powered mainstream PCs. If you want retro, it’s gotta be 8-bit.

    1. … and there is also enthusiasm for underpowered mainstream PCs, up to the Athlon era at least. Particularly around the first generation of true GPUs, original Radeons or GeForce, with thunderbird Athlons, Willamette P4s, or tualatin PIIIs…then the Voodoo fan club has always been a thing.

  3. For the record, I wasn’t part of the “so what” cohort in the comments for the Boston Dynamics dance video. The article explains why: they essentially had to make sumo wrestlers perform ballet. Hell yeah, it was hard, and hell yeah, they had to tweak moves over and over, just like human dancers do.

    1. >and hell yeah, they had to tweak moves over and over, just like human dancers do.

      The difference is, a human dancer learns new motor skills and improves their control in general by this “fine tuning”, so the ballet-dancing sumo wrestler can then easily adapt to other ballets than “swan lake”, or other dances entirely.

      Whereas the robot is just running a special program for that special occasion, which doesn’t extrapolate – except through the improving skills of its programmers to make new special programs for other special cases. In other words, if you expect to have the robot perform task X for you, you should expect to hire the entire team to adapt the robot for that task – because they’re the real “AI” that’s running the robot.

      1. In general, I’m seeing a trend which started with AI after the 70’s, which is to attack the nominal problem instead of the actual problem, because you have no idea how to even define what you’re trying to do.

        If the problem is to make a robot that walks, instead of figuring out what walking is or requires, you whittle the problem down to, “ambulating in any sense of the word”, and then pretend that you’re solving the bigger question by producing a number of partial solutions for a number of special cases.

        The outcome is a number of party tricks that can be applied for show, but not for any actual application where you need the robot to walk to perform some other function, like rescuing people out of a collapsed building. It’s no good if the robot is suddenly unable to walk because it would have to walk in a way that is not pre-programmed. It would be the proverbial blonde who has to stop to chew bubble gum.

      2. Maybe you don’t understand: nobody is claiming these are self-teaching robots. Nor are they autonomous. They are TOOLS. The videos show what they CAN do, not what they will do out of the box. So yes, teaching “them” to dance is really teaching the engineers how to make them dance. But so what? Programming is like math: once you’ve done something, you can apply that knowledge to any future application.

  4. >the routines were apparently not just motion-captured from human dancers, but put together from a suite of moves Atlas, Spot, and Handle had already been trained on.

    What’s the difference? Repeating the previous comment:

    “The dance routine has exactly defined paths and the robot’s task is to follow those paths, which means someone has already calculated the dynamic balancing problem in advance (…) it is left to the robot to correct the difference between the already planned motions and their physical execution, which is a far simpler problem.”

    Boston Dynamics may use previously computed programs and paths to pull off some parts of the dance routine, but it’s still the same thing – going through a planned routine that is specially designed and honed to stay within the robot’s dynamic envelope – to stay within the limits of what the robot can compensate for. It’s like a Rube Goldberg machine that you carefully fine tune until it performs its trick despite everything being on a hair trigger and balancing on razor blades.

    All I’m saying is, we don’t see how many takes it took to film it. We don’t see how many times the robot fell down. We don’t see from the video how narrow the operating envelope really is – which is why it is showboating. Boston Dynamics has a bad habit of this.

    1. I don’t remember which robotics lab it was; I saw some other team doing a humanoid robot that can jump over boxes, and they added the blooper reel where they showed this exact point. In the promotional material, we see the robot making a running jump over a box – while in the blooper reel we see it hitting the box four or five times before it clears it just once and everyone cheers.

    2. You “push up” on the control pad joystick, and the robot walks forward, goes up and down stairs, climbs, and walks over obstacles to execute a single “go forward” instruction provided by a human.
      You were given the “making of” video link last time. Why are you still pretending you don’t know or that walking is scripted, when you know it isn’t?

      1. I know how the walking works. It’s an entirely different routine.

        >the robot walks forward, goes up and down stairs, climbs, and walks over obstacles to execute a single “go forward” instruction

        No it doesn’t. Each of those is a different routine that is initiated when the operator decides it’s time to do that. This is exactly what I’m talking about: people having illusions about the ability of the robot – which is exactly why Boston Dynamics is doing these videos.

        1. “No it doesn’t. Each of those is a different routine that is initiated when the operator decides it’s time to do that.”

          In September I was given a 20-ish minute opportunity to pilot a Spot robot, a purchased unit that was not on boston dynamics labs or mfg shop.

          Seeing as I was the operator you’re speaking of, I did not have to do any such thing. Pushing up on the joystick, and in autonomous mode simply tapping the touch screen, cause spot to move to that point.
          I didn’t even have to let go and re-push the joystick up. I held it there. It successfully climbed up and down a stairs obstacle, and it strafed to avoid surfaces it couldn’t get footing on (an unsecured board)

          That you could make such a statement shows you haven’t seen one of these being controlled, let alone actually took control of one.

      2. Furthermore, I’m not saying that the robots are simply running a fixed rigid script, but that the routines are limited in what they do and how much or what they compensate for.

        Again, like the dog walking routine that always does the diagonal support thing, so when it’s in walking mode it keeps trotting and can’t take individual steps with its feet. In order to step over stuff, it switches to passive balance mode where it keeps the center of gravity within the base of support and ONLY moves one foot at a time… etc.

        It’s basically a bag of tricks; one trick solves one carefully limited problem. The machine doesn’t walk in a general sense, it performs A walk.

  5. The sole fact of fake schematics alone does not imply that there are no nefarious intents coming with covid vaccination frenzy. In fact I consider this rumpus could very probably be a false flag defamation operation performed by vaxers, as there is no reason a anti-vaxer would use a random, possibly well known schematics to support his theories.

    1. @ Solipso, you’ve got it backwards. It’s not about the vaccine, it’s about discrediting all criticism of 5G. These days if you criticise any of the legitimate issues with 5G (eg, very precise tracking of each user, the fact that the cost of 5G infrastructure might be better spent on fixed wired infrastructure, potential backdoors built in by the manufacturer…) then you get tarred as a loony antivaxer.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.