Ask Hackaday: What Can Only A Computer Do?

It is easy to apply computers to improve things we already understand. For example, instead of a piano today, you might buy a synthesizer. It looks and works — sometimes — as a piano. But it can also do lots of other things like play horns, or accompany you with a rhythm track or record and playback your music. There’s plenty of examples of this: word processors instead of typewriters, MP3 players instead of tape decks, and PDF files instead of printed material. But what about something totally new? I was thinking of this while looking at Sonic Pi, a musical instrument you play by coding.

But back to the keyboard, the word processor, and the MP3 player. Those things aren’t so much revolutionary as they are evolutionary. Even something like digital photography isn’t all that revolutionary. Sure, most of us couldn’t do all the magic you can do in PhotoShop in a dark room, but some wizards could. Most of us couldn’t lay out a camera-ready brochure either, but people did it every day without the benefit of computers. So what are the things that we are using computers for that are totally new? What can you do with the help of a computer that you absolutely couldn’t without?

I’m not sure there is a crisp answer to this. It is more like a spectrum. Sonic Pi — you can see a video below — uses coding to create music, but it is still music. Maybe to be totally revolutionary it should be directly doing brain entrainment. Still, I think the idea of creating via programming in real time is a bigger leap and one that not everyone appreciates. For example, some people love OpenSCAD and some want to use a mouse to draw 3D models. Other people use FreeCAD and switch back and forth since FreeCAD can use OpenSCAD.

So it seems like most of what computers have done is made things faster or easier. You don’t need a Sears catalog anymore because you shop on a Web site. Even Sears has one. So does the couple next door who makes and sells custom baby blankets. When you order one, you don’t have to send a check in the mail and the small business owner doesn’t have to go to the post office. Everything can be done via connected computers.

But what are the things that are uniquely due to the computer? From what I understand, modern jet aircraft, for example, are only flyable because of computers. You could know how to handle the negative stability, but you’d never be able to make adjustments fast enough. This might be the key. After all, at the base level, computers don’t do anything magical. But they can do what they do faster than we could ever hope to do it. But given enough time you could do the math for, say, tomography or X-ray crystallography by hand. It was just be a lot of work.

Making music by coding probably qualifies as do most other coding tasks although there were player pianos and jacquard looms with no computers. Again, something like Sonic Pi that does it in real time is probably more the province of computers. Sure, I guess you could posit some kind of machine that isn’t a traditional computer accepting the code, but then you are really just reinventing the computer if you make that argument.

So the question is: what are things that you can only do with a computer? Seems easy until you start thinking about it. Of course, some things are infinitely more practical or made much better. Handmade 3D prints would suck. It was very expensive to create camera-ready copy, math tables, or 3D animation before computers, yet it was all done by someone. Where are the new frontiers enabled solely by computers?

128 thoughts on “Ask Hackaday: What Can Only A Computer Do?

  1. well, modern airplanes are impossible without computer control, but certainly, there were airplanes before silicon chips, even if they were lesser machines.

    I’d start with space travel. You simply can’t get an enormously unstable rocket into space and have it do anything useful without real-time computer control, it’s just too alien to human senses and reaction times. Even the simple ballistic missiles of the 40’s and 50’s that didn’t do much more than go up, over and down, did their thing under the control of mechanical computers.

    Also, modern medical imaging. X rays are about as far as you can go without the ability to do Fourier transfers, and while it’s *possible* to do those on paper, waiting 100 years for the results of your CT scan is not optimal.

    Oh, and cryptocurrencies, although I think the jury is till out on whether that represents “progress”

    1. Many modern aircraft should still fly just fine without the computer, in recent years the Maxi shows that the computer is actually dangerous rather than helpful even (obviously being currently a fly by wire rather than manual connection it does have a computer in the loop – but your typical passenger and cargo plane arrangements are more than stable enough a human could handle it fine)… The only aircraft you really can’t fly without a computer are those that are really unstable by nature, which as far as I know right now is entirely military aircraft…

      I’d say space is still perfectly doable ‘manually’ – you might need to account for lower fuel efficiency, but nothing about flying in space is actually you MUST have a computer. It just makes everything so much easier – with paper, orbital maps and slide rule it would probably take you most of a day to really figure out the burn to do a rough orbit shift, intercept with some planet etc and then most of the coast time to figure out how much adjustment you need… Add in a calculator probably down to a few hours, but the computer makes that nearly instant – in many ways KSP’s drag and plot style burn engine shows how easy the work is for a computer…

      I’d also say stuff you can call Cryptocurrency has existed for a very very long time before computers, just now its convenient enough to be basically cash, rather than traded secrets and IOU that only have value to the right people…

      I don’t think so far there is anything a computer does that could not be done by a person, perhaps with some design changes, unless you count functioning with such vast uptimes doing the same tedious job, with good reliability…

      1. Civilian planes are definitely flyable without a computer, though a computer can help keeping them away from unsafe attitudes *. As an example, the A320 that ditched in the Hudson was essentially kept from stalling by the flight computer, even though it was in a degraded state. But relaxed stability military planes are completely unflyable without a computer, and sometimes going for a stable design is not an option. For example the B2 relies on its pitch instability to reduce control surface deflection and keep the radar cross section low, while air superiority fighters are unstable in pitch to increase their agility and immediate pitch response. Maybe it would be possible to dampen their pitch response with rate gyros mechanically mixed with the pilot inputs, but the workload would be enormous, close to manually hovering an helicopter. Definitely something you wouldn’t do for hours at a time.

        * If the software is not written in a rush, without too much scrutiny and testing, without redundancy in inputs, to patch up an intrinsic aerodynamic issue …

      1. Are you sure that’s what you meant to say? I’ve seen lots and lots and lots of computers do Fourier transforms. But I’ve also seen humans calculating Fourier transforms, so while it’s tedious, it’s definitely doable for man or machine.

          1. I guess, the word “transfer”. I assumed you meant “Fourier transform” because there’s no such thing as a “Fourier transfer”, and
            1) Fourier transforms are almost universally performed by computers, and
            2) Now I see that you were just mocking the previous comment, thereby violating the “Please be kind” plea at the bottom of the page.

            So yeah, “transfer” was the part of “Fourier transform” that eluded me. Fixed.

    2. Every pseudo random number of an Computer is more random as an random number from an human.
      The human brain has a preference for dear-numbers an unlike clustering of random numbers. The generate wrong random numbers.
      Only Computer with ein Analog seed generator kann generate real random. No human can make this.

  2. Without computers we would be waiting a very long time for COVID-19 vaccines, probably decades.

    Without computers we would not be able to manage complex engineering tasks like the big dig.

    Without computers we would not have retail stores filled with the merchandise that we want to buy today.

    Without computers we would not be able to track and react to outbreaks of disease or poisoning.

    Without computers we would not be able to forecast the weather, terrible storms would hit with no warning.

        1. And complex engineering tasks were done before computers even at the level of the Big Dig, retail stores were filled with goods we wanted (just not computerized gadgets), we did track and react to outbreaks of disease and poisoning (see e.g. the cholera epidemics of 1850s)…

      1. Simple remote control signal wouldn’t do. You can’t gain signal to noise ratio without DSP and forward error correction codes.
        Therfore, the answer is “without (embedded) computers you can’t have any autonomous robots, nor communicate across wast stretches of interplanetary space”.

    1. In the spirit of “You could do this without a computer, as long as you’re willing to put in a lot more time and energy”, the no-computer option for Mars rovers would be to send a group of humans there to drive the rovers around.

    2. Define “a computer”.

      You could build a BEAM bot like analog system that rovers around and beams up passive telemetry. You just have to build it robust enough that it survives bumping into things as it goes.

      1. Oh noes. Somebody left a roomba on the rocket.
        (Or… if your remember a long time ago, there were these Radio Shack cars that would run forward, and if something hit their front bumper, they would go in reverse but the wheel would angle so that it would steer in a curve on the reverse… until it hit the rear bumper, then it would run forward again)

  3. I would nominate FDM 3D printers. I don’t believe that anyone had the steadiness of hand needed to lay hundreds of hair fine lines of molten plastic one on top of the other for hours at a time to build complex parts. That is without considering the difficulties of doing the job of the slicer software to generate the instructions for the layers. Machine tools existed before CNC, but I am unaware of any FDM type processes before the Stratasys patent in 1989.

    1. We had those art’s’crafts folks that did all manner of such-
      The hot-glue-popsicle-stick people
      The multilayered candlestick people
      The glassblowers
      Those “clay rope” pottery people
      Those weirdos who used their mig welders to make things
      Those jewelers who made those “wire trees” (or for that matter, the “string art” people)

      But… they were unique and very time-consuming works of art… and were universally “expensive for what it was.”
      And… usually required a lot of finishing to make something acceptable (all hail the die grinder, all hail the sander).

      But now we’ve automated production… and I wonder about those Trunk Show people… I’ll be looking for them at this year’s state fair…

    2. I think you’re basically that without a computer we couldn’t make a part “using this method”, but the part could almost certainly be made using other methods. It’s a bit like saying that without a computer you couldn’t print using a laser printer. You could still print using any of a number of non-computer methods, so the computer isn’t necessary to accomplish the real task, which is to print something.

    3. By-hand additive manufacturing is difficult but possible.
      One thing I thought about was multi axis milling of stuff like titanium or 304 steel.
      You have to cut deep enough so the material doesn’t work harden, so you can’t take your time, and if you need to do it accurately in three or four (or realistically even two) axes, it’s not manageable to accurately juggle that many variables at once.
      Like, cutting jet turbine blades.
      You’d have to move to a different process, abrasive, to do that kind of work by hand, and that leaves a different and maybe unacceptable surface finish and residual strain.

        1. Interesting link.

          I also read the second story, “The Assassin”

          The gist – aliens come to Earth purporting to give peace and knowledge, one man is suspicious, thinks humans are getting weak, reveals said aliens to be robots, humanity is so revolted it rejects aliens and is ‘saved’

          I found it interesting that the author assumes that 1950 humanity would be absolutely freaked out and repulsed by finding the aliens had robotic form, but as I read it in 2021, I thought to myself “oh yeah, a spacefaring species that projects itself as robots – that makes total sense”

  4. Math.
    Sure, you could do the math by hand — and some people did, but you need a computer to get the results you desire:

    [Perspectives and pointillism] => Raytracing
    [Film holograms] => Computer generated holograms
    [Roman ciphers] => Cryptography => Blockchain => Digital Fungible Tokens [using crypto to make certificates of authenticity]
    [Morse coding] => High-speed encoding (from radar to telecom to space signals)
    Iterated function systems and fractals and Laplacian things
    [Vaccines from antibodies] => Creating a vaccine from the virus (no survivors needed)…
    [Tape loops and squink boxes] => Digital audio workstations
    [Craftsmen of all kinds] => CNC machinery (arts’n’crafts bits => 3d printing)

    I mean, they’re all something similar to what existed before, but…
    Then again, they all seem to be related to “Machine Thinking”

    How to I stop “thinking” about “the stool” and start thinking about “make a machine to make stools” to… “make a company to make machines” [eg Bridgeport] to… “make a machine to make factories” [the Toyota way] to… “make a system that makes companies” [VCs and incubators] to…

    Sooner or later we’re going to wonder what it is to be human again…

    1. Math was my answer as well. I thought there was an XKCD comic that used a picture of a body of water (the ocean) to describe where people fell re: capability to understand mathematical concepts and the depth of where that concept fell. At the very bottom there was a section labeled that only computers could perform the math. I’ll have to look for that again now….

  5. ” Where are the new frontiers enabled solely by computers?”

    I’m sure microcontrollers have changed things even more, both in the marriage of digital and analog, as well as moving what was formerly done analog into a digital way.

    1. ”Where are the new frontiers enabled solely by computers?”

      I’m sure anyone who is walking around today because of a predictive algorithm once run by some anonymous microcontroller in an airbag controller has all sorts of thoughts on new horizons.

      That reminds me, if I’m ever in that particular situation, I should remember to salvage that control unit from the car, put it in a comfortable place of honor, and buy it a drink every year.

      1. “That reminds me, if I’m ever in that particular situation, I should remember to salvage that control unit from the car, put it in a comfortable place of honor, and buy it a drink every year.”

        Interesting thought…
        I wonder if the insurance company would allow you to have it.

  6. The Global Positioning System would be impossible without computers. The satellites couldn’t have been launched, their orbits couldn’t be monitored in real-time with sufficient precision, the almanac and ephermeris data couldn’t be calculated and transmitted to the receiver in a timely fashion and with sufficient precision, the receiver couldn’t convert the timing deltas into latitude and longitude, etc.

    Quadcopters couldn’t be flown in a controlled way without computers to translate the desired motion into speed control for each of the four motors. GPS-stabilized quadcopters would be impossible for even more reasons.

    3-d animated movies of the type done by Pixar and similar couldn’t be done without computers. Perhaps in theory the ray-tracing could be done, but it certainly wasn’t economically feasible to do a full-length feature film at cinema-worthy resolution.

    1. While GPS as it stands you are very correct is impossible, a ‘GPS’ system that has no computer wouldn’t be impossible. Probably even get down the same precision as GPS if you really wanted, it would just take a very long time to calculate, and lots of very tedious direction finding…

      You also can fly multi rotors without computers, simple analogue feedback mechanisms could be used – so your RC inputs bias the drive systems fractionally – perhaps even self balanced by tilt switch, pendulum on a pot type systems – so no computer or even digital process required… It would just be fiendishly tricky to actually tune the initial build to be at all close to stable, and no doubt require quick reflexes to keep it running, but it could be done… In the same way Furze’s hoverbike is just about flyable…

      I’d also argue that the magicians of the hand drawn Disney etc animated movies could have done a more Pixar 3-D style movie just fine, don’t think they would gain anything by it – its just not needed to tell the story, and the animated flatter (though pseudo 3D) art style really looks good – the 3d elements are highlighted and drawn in only on the bits that really need it which makes it pop nicely…

      I do agree all three are made massively simpler and faster to get running by computers, but I’d say none of them are really impossible.

      1. Protractor+telescope+plumb bob.
        GPS is still just providing a more convenient, if convoluted, way of navigation.

        At the end of the day, it’s not computers that can’t do anything new, it’s humanity and it’s imagination that’s the limitation.

        1. I challenge anyone to use a protractor, telescope, and plumb bob to keep a quadcopter hovering stationary (within a few meters) in gusty wind over the middle of the ocean, a thousand miles from land. That’s a piece of cake for a GPS stabilized consumer drone with its on-board computer.

          GPS works faster than older surveying methods, and it works on moving vehicles, without landmarks, anywhere on earth, to precision higher than celestial navigation or dead reckoning.

          Computing speed matters, at least sometimes.

          1. In fact, the long integrating time for GPS means that the PIGA or some other dead reckoning system is necessary to achieve the necessary response speed to keep a quadcopter stationary in winds. A modern equivalent is simply to use a MEMS accelerometer.

            Most GPS receivers give position updates at 1 Hz but those are based on a number of previous measurements by prediction (Kalman filtering) because each individual fix can have random noise up to 50 meters. A gust of wind can blow the copter off a hundred meters before the GPS notices anything.

          2. >Those were called “analog computers”.

            Again, define “computer”. A pen dropped from your fingers is an analog computer that computes acceleration by earth’s gravity against aerodynamic drag.

          3. Okay, let me explain what “analog” means. It does not mean continuous as opposed to “digital” having discrete steps; it means that quantities in a simulated world are represented by different quantities. For example, the position of a physical item may be represented by voltage on a node in an electrical circuit. So a falling pen is just a falling pen – it is not an analog computer unless you’re using the pen to represent something else.

            I agree that analog computers stretch the definition of “computer”, and that’s a result of fundamental differences between analog and digital computation. What constitutes a digital “computer” comes from a more formalized definition for digital computers. Digital computers today are defined by their Turing-completeness, and by the fact that anything that can be done with any Turing-complete computer can be done by ANY Turing-complete computer. No digital machine is considered a computer if it is not Turing-complete. This does not transfer to analog computers, because analog computers are limited by the precision and repeatability of their components, so they are not interchangeable; two integrators may have wildly differing capabilities, and it matters how long a chain of modules a continuous (“analog”) signal goes through, while an AND gate is an AND gate no matter how it is implemented or how many of them you string together. I don’t know of an analog counterpart to Turing-completeness that applies to analog computers, and this is why the integerator and multiplier mechanism used in mechanical gasoline pump meters is considered a computer, while the direct digital equivalent would not be.

        1. But also in the context of the article, synthesizing instruments that don’t exist in acoustic form doesn’t require a computer. Analog oscillators, voltage-controlled amps and filters, and the other useful modules, fall far short of qualifying as computers.

          1. To be fair, there were analog computers prior to digital computers. I saw one in school back in the day. Lots of op-amps and jumpers. I think they were used a lot by the military to calculate projectile flight and so on.

            Not sure if this qualifies for the above discussion, but definitely a computer.

          2. Heh… I think we found a Moog man.
            I mean, we’ve had speakers on slip-rings that rotate, tape loops that were 20 feet long, and spring-loaded reverb for ages.
            Hell, a “flange” box used to be a speaker and a contact mic on a special shaped bit of metal (the flange)… electronics not necessary if you rigged it to the fx loop on your board (so you could get the mix/levels right).

            Granted, a BBD device (is that considered a computer?) is a whole lot easier than a 25 foot tape loop.
            And making a whole DAW out of it (or an iPad controlled mixer)… is easier with a computer.

  7. I don’t understand the point of the question as asked. It seems like it has been defined down into irrelevancy. “What can professional athletes do that I can’t do?” Well, if you don’t care that my performance is incredibly different as to be nearly unrecognizable compared to that of a professional athlete, than sure they are basically the same thing…but that isn’t really true or a useful way of defining the problem. The way a professional athlete approaches the problems, trains, reacts, and their results are completely on another level. Similarly, Computer AI, while in some ways mimicking human intelligence (and failing in many ways), completely outstrips us in pattern recognition capabilities in other areas. The patterns that humans would almost certainly never be able to see are easy for AI. What about sending terabytes of data across the world in seconds to minutes? Sure a human can do that in a longer time span, but calling those things similar is a huge stretch, due to the multiple orders of magnitudes in difference and also in the tools/techniques that are used.

    1. This!

      Essentially everthing a computer does a human can do manually. It will just take ages.

      The argument of the Author that some military aircraft can’t be flown without computers is simply nonsense. If there were no computers there would have been aircraft that are not that imbalanced and actually can be flown with a manual stick. It is just another example of increasing precision.

      I have worked on a project in 2001 where we used high speed (single line) cameras (5000+fps) and would search for air bubles in molten glass pipes that are shooting out of a high pressure chamber at 5 meters per second. The pipes would be automatically cut every 2 meters and the system will sort out pipes with bubles. It is impossible to do this by eye, while the pipe is still molten and passing by at this high speed. The job was done none the less by humans, but after the pipes were cooled down and investing a whole lot of time to inspect a single pipe. The added benefit of computers is simply speed. (Pipes were being used to create automotive lightbulbs. The more air bubles are in there the shorter the lifespan.) A single machine would be as effective as a team of 76 workers. All of whom got new positions at the plant.

      Same applies for fingerprint matching in a database. Can be done by humans when you compare one fingerfrint to maybe 10 thousand. But when you need to compare one fingerprint to 500 million, humans fail at the job.

      1. But as you can see numerous examples of in these comments, the essential thing is that computers can do things way faster than humans, and therefore there is a limit to things that people can control in real time, creating a whole class of things that computers CAN do that humans can’t. This is why the “remote control of Mars landers” and “videogames” examples are exactly right. There is no number of humans you can throw at a simulation that will make a videogame work.

          1. In any practical application, yes. They can be fundamentally different when you have a task that simply can’t be done slowly. If your task is to catch a flying ball, not being able to compute and move quickly enough makes the task impossible for you. Not difficult, but impossible.

    1. So anything where speed is necessary that’s where computer will win. Stock trading, simulating 1000’s of structural designs, reactive armor on tanks, etc. If speed is key, computers are king.

      1. Yeah, I was thinking high frequency trading, which is by its nature inhuman.
        Real-time video and audio processing. I remember how impressed I was with an early Amiga demo where they had a camera hooked to it and had a fake fly flying around the person’s head as the video on the screen was mirroring what the person was doing.
        A lot of services where the utility in the service is that it’s real-time. Moving maps. Sure you can make a map by hand or whatever, but you can’t accurately say “you are exactly here” continuously for a significant number of objects, like air traffic control does. Back in the day when ATC was done by hand, their position update time was in the one minute range, which wasn’t great for collision avoidance with jets flying head-on to each other.

    1. What would your preferred time scale be? Someone mentioned COVID vaccines. However, even with computers, it’s still taken multiple human lifetimes to develop them because multiple people are working together. Putting things on an individual human’s time scale doesn’t seem fair either.

      Before electronic computers, there were human computers. Rooms of people who each had just a couple of jobs, but worked together to achieve a result faster than if one person was doing all of the necessary calculations.

      I think it’s an intrinsic property of computers that they can’t do anything that couldn’t be done without them. They’re just faster, smaller, and more efficient. That’s no small feat. But they can’t intrinsically make something possible that was previously impossible, in the strictest sense of the word.

        1. I’m thinking I’ll have a baloon to the edge and then a rocket for the rest. Or perhaps I’ll just find a way to afford to build a machine that can stay still and let the earth fall out from under me.
          Maybe a centrifugal syphon into high orbit, insert a capsule at the bottom and let the water take you to the top. Hang out for a while and let the moon catch up with you. :D

    1. … Not really; current computers are not Three Law compliant:

      A computer can, by action or inaction, harm a human. (you have to define ‘harm’ pretty broadly, though- like draining someone’s bank account through a mistake in coding, or a batch job that runs repeatedly to take money out of the account, etc.)
      A computer does not always obey orders given to it by human beings. (I’m struggling with that one presently trying to configure a cloud service to allow documents to be encrypted, but allow sharing with others outside my organization)
      A computer cannot protect its own existence. (In fact some of them are designed with obsolescence in mind, or intentionally with parts designed to fail after a given time period.)

      It’s like the old set of jokes: “Computers are devices to speed and automate errors.” (author unknown, but from the Jargon File) and “To err is human, but to really foul things up you need a computer.”- Paul R. Ehrlich

  8. Okay, I’m going to go sideways. The thing that requires a computer is a compiler. If no computer exists, then there is no need for a compiler, so even though a human could theoretically do the task, they just never would. Don’t throw Lady Ada Lovelace at me either, since what she was doing was machine code assembly. Compilers only became a thing when computer users found out how tedious it was to program in machine code.

    1. It´s not because they never would that they cannot. The question is what can ONLY a computer do ?
      If humans can – albeit slowly and tediously – compile source code, then not only computers can.

      1. I’m staying within the spirit of the article, which is looking for things that are done by computers that are NOT imitations of things that previously had been done by people or other animals.

      2. i was thinking the same as BrightBlueJim, and i find his argument compelling. if you can’t think to do something, then you can’t do it. it’s possible to get a mistaken impression about the word “possible.” if it isn’t possible to imagine doing something, then it isn’t possible to do it. our imagination flows sequentially from our environment (our experiences). we’ve grown up with computers so we can imagine compilers. people did start thinking about compilers before computers existed, but so far as i know no one thought about a compiler before they thought about a computer.

        so i guess maybe it’s more strictly correct to say, compilers are the thing that only thinking about a computer can do.

    2. “Compilers only became a thing when computer users found out how tedious it was to program in machine code.”

      A hundred years ago, a “computer” was a person who computed. It is not a far stretch to say that a person who compiles information (such as stocking a library -brick and mortar- not .lib) would be considered a “compiler”.


      1. Indeed… I think you’d call all men of industry “compilers” — we’ll take the arcane one-off-instructions on how to build a car,
        break it down into serializable steps,
        forge out the tasks to dedicated components
        and produce a set of instructions that use the inputs (blueprints) to make mass quantities of products.

  9. Reduce the voltage potential of electrons from one level to a lower level, continuously, without deteriorating, sustainably, at higher currents than a human. All the while producing consistent results. :)

    1. “in Illinois, the retailer’s home state”

      There are still around 300 stores left, including Kmart, which I didn’t realize had merged. (The last Kmart in my area closed over 20 years ago.)

  10. The answer is “the internet”. And the obvious counterexample (“what doesn’t require a computer?”) is easy: Hiroshima.

    I’d like to see an article about the Mercury program and computers. Challenge accepted?

    1. You may be right about Hiroshima – that was an enriched uranium gun bomb, which the physicists were confident would work. If you had said “Hiroshima and Nagasaki”, you would be wrong, because the Fat Man bomb was much more complex, the design of which did require computers to do simulations.

      Regarding the Mercury program, while the spacecraft included no computers, it certainly depended on the ones on the ground to calculate firing times and durations.

        1. No, they weren’t. They started to try to do simulations using a sort of manual DSP, but Richard Feynman tells about how the arrival of their first electronic computer made things possible that just weren’t going to work with the human computers, ever. Never mind that work practically stopped for a couple of weeks because the physicists were playing with the new machine.

  11. It’s not that we CAN’T do x without a computer. It’s just that we wouldn’t want to do x without a computer because the cost in time, money, labor, energy, etc would be too high. For some applications, hard constraints on any of those are enough to require a computer for all practical purposes.

    1. Wasn’t there some technology for not carrying cash before EFTPOS? I recall there something called .. cheque? It’s not electronic (the first letter in EFTPOS), but it is still fund transfer at point of sale.

  12. My computer can make other people believe I’m awake and attentive 24/7, even if I’m actually asleep or not present.

    This extrapolates into all areas where I wrote a script to be my “transistor man” in my absence :)

  13. Backup and restore.
    Cheat death.

    Build a robot and teach it to walk, talk, and do something useful.
    Take a backup before sending it into a dangerous situation, then after it dies, restore it to an identical or upgraded machine.

  14. Waste huge amounts of energy, releasing tonnes of CO2 into the atmosphere, to generate massive blocks of inherently useless cryptographic hashes and numbers, so that they can be sold to people who think their value will continue to increase forever.

  15. There s nothing a computer can do a human cannot, they are merely “force multipliers” just like a shovel or a wheelbarrow. If we hadn’t had those in the past we could still dig a ditch or build a dam it would have taken much longer, more laborers but it would have gotten done. Computer were created by humans using a slide rule and a pencil, also force multipliers.

    1. More importantly, they are force accelerators. The dramatic difference between human-mind-calculation and digital-computer-calculation just makes it impossible for human minds to do many of the things that computers do. If you have a force multiplier, you can substitute the parallel application of multiple people. Therefore, if you get enough people together you can do the work of a D9 Caterpillar. But as they say, you can’t produce a baby in one month by setting nine women to the task.

      1. Not quite, but stage them properly, and you can get to 1 baby a month as an ongoing concern.
        Welcome to supply chain logistics…

        Or you can “on average” have 1 baby a month even without staging. Welcome to statistics and forecasting.

    1. Thank you, that is the correct answer. Some of the earliest video games (e.g., Pong) didn’t really need a computer – the arcade version was originally just a state machine and some analog electronics, not Turing-complete. But that could be considered a hybrid computer, an analog computer with some digital circuits to make the state machine. In general, video games are an art form that simply did not exist and could not exist before computers, and video games were among the first applications developed on them. Case in point, Space Wars on the DEC PDP-1. There are even earlier “space war” style games that were text based, and I am convinced that the PDP-1 was equipped with an X-Y DAC-driven CRT display specifically to be able to make cooler video games.

  16. A lot of things, while technically possible, were too slow to be practical due to the math overhead. So I nominate the field of DSP.
    Technical analysis of trades on the stock market was possible before computers, so I refine my nomination to: audio (and RF) DSP.

    Analog filters can be designed, but component values and tolerances put limits on what can be built. The performance of a well designed DSP filter can have a response impossible to match using available capacitors and inductors.

  17. IMHO this question cannot be answered without defining “computer” first. Less than 100 years ago, “computer” was a profession, not a machine. They computed math tables and similar stuff. And then there are mechanical, electrical, electronic and quantum machines, coming in analog, binary, whatever-base and all-at-once flavours, with or without program or data storage. Does a simple lever apply? It can compute y=c*x faster than any human. Or feedback loops, that can do higher and time-related math?

  18. My vote would be on particle colliders. Even not the acceleration part, that could be done with analog system, but the collection and analysis of sensor data. The sensors would probably be different, lower resolution and more manual work, but some of the particles would be hard (or impossible) to find.

  19. We’d be able to do almost everything we can do now, just orders of magnitude slower, less reliable, power hungry, and more expensive – which of course would rule out a lot of current use-cases and 99% of users.

  20. Computers are a tool. Tools are things that are used to enhance human abilities. All of the things that computers do are thought up by people, to enhance some ability that they lacked. It’s inherent to a tool that it does something that a human could do, only better.

    To have a computer do something that no human could do, I think will need an imagination that goes beyond the human imagination.

    The thing with imagination in humans is that it works by combining the millions of things that we learned during our life. It’s impossible for a brain to imagine something that it never had at the least a clue about. A brain can only work if it has input, and only knows about what was put into it.

    I guess it’s also the inverse: there is nothing that a human can do, that a computer can’t do.

  21. Computers can provide an unbiased response to a set of input variables that will be unrelated to external influence. Similar to the earlier comment on a computer always failing at the same point and never learning. It is merely responding to a set of input variables with no external influence to bias a result (failure).

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.