2
  • I have to point out, it is an attempt to make this on-topic, all thanks to @JBH, so tell us in the comments how well we did. the old one is history now.

50 years ago my grandfather created humanity's most impressive computer

When grandpa dreamed, he dreamed big! he took the best tech of his day, from early solid-state hard drives to the latest derivative of the Motorola 68000 CPU, from the most amazing radioisotope thermoelectric generators to solar panels just like those on Skylab! Look, I'm not going to bore you with the details about how this thing was built — that's really not at all important to my question — but I will explain one more thing that is.

It's HUGE!

OK, two things...

It's IN SPACE!

Specifically, my grandfather's big honking supercomputer is floating out there, right now, 2021, in a Lagrange point. I can Telnet to it, I can play games on it... But here's my problem: what do I do with it?

You see, I have a few gambling debts...

Who knows what grandpa did with the thing, but it's up there, running just fine, and when he passed away he left it to me. Now I need this thing to start generating some cash! Don't worry why, that's not really important either... but what is important is this: its computational prowess is something on the order of 1.9 zettaFLOPS. I'll let you work that out in your head for a moment. (This easily blows away most of the top 500 computing systems in 2021 combined.)

But it's worth remembering, 50 year old tech! That means the machine works through massive parallelization. I really meant it when I said the individual processors are like Motorola 68000s. I know the OS will have some impact on your answers, so please believe me that it's a somewhat souped-up version of SunOS 1. Yeah, 1980s... where do you think they got their ideas? Hah! Grandpa! I'm kinda hoping that your answers won't depend on the OS, just so you know.

Now, if you really want to know the details, I'll tell you, but honestly, it's completely unimportant. How that monster got up there, what satellite network we use to connect to it, what operating system it uses, how it manages its power and maintains itself... it's all unimportant. The only thing you need to know at this point is that I'm clueless about how to generate some cash with this thing. I'm not at all interested in selling it (I'm sentimental!). I want to put it to some use!

So here's the question, folks...

I know this is a bit vague, but frankly, I think the sheer scope of this thing limits its potential uses and it's not like there's an infinite number of business models. So...

What business model can I apply to start generating some cash?

  1. Remember, I'm not going to sell it.
  2. Ignore costs of operation. Yes, they're sizeable, but if you're going to suggest something low-brow like online gambling... honestly, 1.9 zettaFLOPS... you didn't look up that number if you think something like that is an effective use of this machine.
  3. You can believe that I have the resources to ensure international-grade encryption (don't ask how we did it with SunOS 1 — it's a family secret!), so don't worry about how data is moved, how fast it's moved, or how secure that transmission is. This question is all about the computer.
  4. Look around, it's 2021 and it doesn't matter where you're living in the world. If you just learned about grandpa's legacy, then you haven't been reading newspapers for the last 50 years! What I have is a resource and I need your help figuring out how to put that resource to work. But remember... you need to use the resource the way it was intended, as a computer. No suggestions about blackmailing countries or I drop the thing on their capital. If it helps, assume I can't or won't move it and I certainly won't damage or destroy it! Look, I'll entertain, shall we say... shady... business models, but I'm no terrorist!
  5. If you're really caught up in how that thing got there or can do what it does... assume aliens helped grandpa do it. It won't bother me and it's as good an excuse as any.
  6. Finally, if you have any hope of being selected as my business partner, you'll explain why the business model you pitch is better than any other.

Some technicality

Connection to the thing is quite modern, even if on the side of the system it is just old fashion solar panels accepting laser beams, at multiple points with let's say 64kbit per receiver with a total capacity 100Gbit, and retrofit 100Gbit downlink. Good catch from @Alexander.

So yes, it not for big data be shuffled up and down, if we take just ram 16MB per each node thrn it takes 770 thousand years to upload download ram snapshot from all the nodes using that 100 Gbit link.

So mode of operation sending relatively small data - GB's, TB's, PB's sizes, expand it with calculations filter out those which aren't useful collect useful data which fits the bandwidth.

quite big limitation actually, which has to be accounted for. Which creates a significant difference to that another question.

Location seems to be important piece of information as well, thanks to @notovny and @Goodies

Location is lagrange point L1 of Sun-Earth, so it is the sunny side, it is how the thing gets its energy from and what makes flashing with leds for advertisement to be pointless, besides that it is 1.5 million km's away.

MolbOrg
  • 6,470
  • 1
  • 18
  • 60
  • Strictly speaking this is an attempt to circumvent the system by reposting a closed question instead of editing the old one – L.Dutch Aug 06 '21 at 06:44
  • 1
    @L.Dutch consensus on meta was, 5 votes, that it has a problem to be too story dependant. posting a new seems reasonable as this angle\edit may invalidate existing answers. Be a judge, if it is the right course of action in this case. But yeah, a little bit of mess created, srry ups – MolbOrg Aug 06 '21 at 06:54
  • Which Lagrange point is it at? If it's at Earth/Sun L4 or L5, all communications with it are going to be about 16 minutes, round-trip. – notovny Aug 06 '21 at 09:43
  • @notovny Sun-earth L1. but in general, ping is one of the problems which limits the use. 10 seconds better than 16 minutes, buuuut not exactly a realtime – MolbOrg Aug 06 '21 at 11:41
  • 2
    It is the same question, just different packaging. I never thought the original question was too story-based, but if it was this one is no different. Switching Mr X to grandpa's inheritance does not change the essence of the question. The main difference between the two is that the original question lacked some details needed for cost-benefit analysis. However, this version asks to exclude operational costs, which is a big no for any proper business plan, especially the one about a monstrous old-tech computer in orbit. Are you looking to generate cash or to put this thing to some (any) use? – Otkin Aug 06 '21 at 18:37
  • 1
  • 1
    "I can Telnet to it, I can play games on it" - but what kind of bandwidth can you get? With 50 years old technology, you will be pretty much limited to telnet speeds, massive parallelism or not. – Alexander Aug 06 '21 at 18:58
  • @Otkin, unfortunately, no, it does not answer my q, but also you may look more carefully at the first line of the q, there is a hint, so as comments at both of those (yes, technically they are dupes, but the q you refer was closed as off-topic and if this isn't off-topic then there is a significant difference between those, I hope - it is an effort refine off-topic q to be an on-topic q - if it fundamentally possible, fingers crossed - problem is real, the answer is required, if possible) – MolbOrg Aug 06 '21 at 19:28
  • @Otkin I'm interested in specific use cases which I do not know about. I can make a list for use as-is or free use cases, have no problems with that, I do not know where are money for it can be, and I have no experience which helps me in that regard, and there are few limitations which cut off your typical general stuff, hosting vps VPN and all that low level typical datacenters stuff. Cost benefit analysis - I can do it myself, profitability never was a question, direction where money are for the thing - that is the q and I was surprised to discover I had no good suggestions, that why the q. – MolbOrg Aug 06 '21 at 19:39
  • @Alexander true, thank for the catch, i really totally missed the limitation and it turns out to be huge. totally forgot about that problem, updated the q with relevant info. – MolbOrg Aug 06 '21 at 20:15
  • 1
    @MolbOrg The main difference between this and the previous version is gift wrapping. The previous version includes more irrelevant information that people have a hard time disregarding. This version explicitly states which information is irrelevant. However, this does not make this version substantially different from the original. If some people think that this version is fine, but the other version is off-topic or too story-based, it speaks more about their reading comprehension ability and their interpretation of your question than the actual merits of your question. – Otkin Aug 06 '21 at 21:32
  • @MolbOrg Take for example 2 questions: 1) 2+2=? ; 2) One day Mary went for a walk. She met a couple of her friends and had a nice chat with them. On her way to a shopping district, she saw an apple tree and grabbed two low hanging ripe apples. She visited a tailor, a butcher, a grocer, a chemist and got everything that was on the shopping list her Mum gave her. On her way home she met John. John is a handsome boy Mary fancies. They spent 2 hours chatting. Then Mary returned home. On a credenza, she saw a note from her Mum saying that they are going to have guests in the evening and [cont.] – Otkin Aug 06 '21 at 21:42
  • [...] suggesting to use 2 last apples in the pantry to make some sweets. Mary nodded and started to prepare baking ingredients. How many apples can Mary use for her sweets? |||| Questions 1) and 2) are the same question, just different wrapping. – Otkin Aug 06 '21 at 21:43
  • 1
    @Otkin But if people already wrote answers based on Mary and John's story, changing the earlier question to get rid of unnecessary backstory would invalidate that - and edits that invalidate answers are illegal. So if this question is a duplicate then you are basically preventing MolbOrg from ever asking this question in any shape or fashion. – KeizerHarm Aug 06 '21 at 22:24
  • @KeizerHarm The backstory for this question was changed because someone did not like the idea of building the described computer today. Just this signifies that the backstory is completely irrelevant. I would recommend closing this question and editing the original to tidy it up a bit and include necessary information while explicitly mentioning which parts of the backstory are irrelevant (answers can be safely merged). One could also argue that if the backstory is of such paramount importance then this question is also 'too story based'. – Otkin Aug 07 '21 at 01:34
  • 1
    @KeizerHarm I do not like questions being closed. And I often find the duplicate rationale for closing questions to be questionable. However, in this case, I agree with L.Dutch. This is not a new question. It is the same question reposted instead of the old one being edited. Most of the content of this edit is superficial and the original question can be improved without incorporating it. – Otkin Aug 07 '21 at 01:39
  • @Otkin we can discuss problems associated with the question on meta I'm very interested in doin so and looking forward to your takes. Atm q is sufficiently different, due connection limitation, which was not present or implied in another one. – MolbOrg Aug 07 '21 at 06:28
  • @MolbOrg Connection limitation information makes this question outside of my area of expertise and my understanding of the question is no longer sufficient. I will retract my vote since I do not have the qualifications to judge the question in its current form. As for meta, I will try to write a brief summary of my thoughts, but I am a bit busy so it may take some time before I get to it. – Otkin Aug 07 '21 at 18:43

7 Answers7

5

Save the world, getting a substantial income from every nation on Earth.

Assuming a generous 0.12 floating point operations per instruction cycle (generous, as the 68000 needed a math co-processor to efficiently perform) and a needed 1.9 Zeta FLOPS then a distributed total of 1.58*10^22 instruction cycles per second are required. According to Ray Kurzweil et al., in 1971 approximately 6 watts were needed per Mega instruction cycle:

Graph showing how much processing was power intensive in the 1970s

Open access Wikipedia 2021

Which would indicate that a 6*1.58*10^16 watt or 94.8 Peta Watt power supply was up there in space ready for use.

The world's current energy usage is about 15 terawatts continuous, so this is substantially more than would be required to supply the world's energy needs for the foreseeable future and clean-up the planet in the process.

Your father was a misguided visionary, and you are a true hero. Now the tricky bit, getting the power down to the people without accidentally roasting them (I leave to the writer) but it should be worth 20 Billion Dollars at current prices. Presumably, interplanetary propulsion by laser would be at your disposal too, so the true space-race begins. Elon Musk weeps into his Dom Perignon.

As to the computer hardware, you have the first orbiting "Museum of Curiosity".

MolbOrg
  • 6,470
  • 1
  • 18
  • 60
Escaped dental patient.
  • 24,651
  • 6
  • 57
  • 115
  • You guessed it right, power consumption of the thing is about 40e16 W, this is a K1 device. Fortunatly or not, but in a market framework nothing runs on good wishes, and before thinking how to get that energy to the planet or do something useful otherwise - it needs cash to hire people to do R&D or buy useful IP's and this is one of the why's the question. And for that one has the Thing. One of the options sell it to gov or maybe UN, loose control over it and hope they will do good things with it, at least there is a chance, lol. – MolbOrg Aug 07 '21 at 07:07
  • @MolbOrg I didn't want to duplicate answers to this question by explaining any one of the suggested methods - in fact I'm thinking about coming up with an alternative question which doesn't involve roasting the planet (adding to the pre-existing temperature issues by beaming more energy down), but still solves the energy need issue. – Escaped dental patient. Aug 10 '21 at 21:52
  • the thing is in form of a ring, cut in a center for not to affect radiation coming to earth, but some sort of umbrella that can scale radiation down, maybe tracking some regions as the earth rotates or cut some portion of the spectrum infrared one - not impossible, when you do one thing so you can build another one for that purpose. on the other hand, fusion efficiency can be 99% (thermodynamics-wise) subtract that with umbrella - can be very convenient - energy produced where it is consumed. For specific projects, energy can be beamed - rectennas and phased arrays may be a good option. – MolbOrg Aug 10 '21 at 22:39
  • Better technologies and implementation of active supporting structures may lift, to the point almost removing the limits of energy-producing/consuming so as to be the way to deliver energy, replacing most of the other ways. But essentially living in space is a better deal, sure it has some price but it offers more, just much more. So that energy delivery makes sense in different times/circumstances but it has no universal solutions and destiny to be replaced. The same problem as with my q - there is no universal cash cow for the situation, same with energy delivery - no universal solution – MolbOrg Aug 10 '21 at 22:39
  • @MolbOrg I was thinking along the same lines, it might just work best as a stepping stone to a space-based civ. or off-planet civ. - no single solution is ideal. – Escaped dental patient. Aug 10 '21 at 22:46
  • The best stepping stone is the moon as a pile of materials, bootstrapping industry in space, make and test what we need for a living there, teleoperate what necessarily until you get there, and lit a patch in pacific with microwaves and rectennas to get electricity - to build means to lift people(billions, if that is the plan) and get required energy - again some combination of things. energy isn't a big factor here, because when one gets to the point of needing it there already will be possibilities to solve that problem in a number of ways. i mean good luck making the q in a good way, lol. – MolbOrg Aug 10 '21 at 23:04
3

Hugely parallel architecture? No concerns around encryption or connectivity? All maintenance done by magic at zero cost??!?

Start mining!!

Specifically: Cryptocurrency mining.

Cryptocurrencies are built on the back of complex maths. Every Bitcoin or Etherium nugget or insert own trendy name here requires that a shedload of computational work be done to prove it is it’s own unique thing and that it can be used as the basis of a shared, trust free ledger (blockchain or some other proprietary algorithm. The need for work is much the same). The more successful (valuable) a currency is the harder it is to do the work, and the more computational power you need to do it.

Good news is it can mostly be done in a distributed fashion. Get a good GPU (hardware built to do the same thing a billion times in parallel) and you can mine cryptocurrency, turning power into value that you own! Then you sell that cryptocurrency for government backed currency and you’re on your way.

But there’s a catch.

You might have noticed that at the moment high end GPUs are very expensive. This isn’t because they’ve gotten harder to make: it’s because they’re being bought by cryptocurrency miners. Why?

Because GPUs burn out. They get old. They overheat. They break. This is a major source of cost for a mining operation, along with the obvious costs for power, cooling and space to store the racks of components.

But your grandpa solved that with Space Magic. So you (using your family secrets) can mine cryptocurrency for free, with no overheads. Then sell it. Choose whichever cryptocurrency is riding high at the moment and go to town.

And pray you don’t crash the market with a poorly thought out tweet...

Joe Bloggs
  • 66,279
  • 25
  • 166
  • 273
  • 2
    I seem to remember that you can control what counts as valid transactions in the Bitcoin blockchain if you control more than half of the network's computing power. Would this computer be enough for that? Because if so, you need to keep it super secret, or the cryptocurrency loses all value immediately. – Schmuddi Aug 06 '21 at 08:44
  • @Schmuddi: Depends on how many nodes are controlling the currency, but probably not (plus that would fall foul of the no nefarious shenanigans requirement in the OP). The main advantages are the lack of overheads in running the mining operation and the fact that it must already be a hugely distributed architecture. – Joe Bloggs Aug 06 '21 at 09:05
  • Re: nefarious shenanigans: You wouldn't have to actually exert your control over the network. I'm pretty convinced that if it was disclosed to the public that the power to validate transactions on the blockchain rested in the hands of a single individual, that knowledge alone would cause the value of the Bitcoin to drop to virtually nothing. – Schmuddi Aug 06 '21 at 09:12
  • Schmuddi's comment is on the nose: cryptocurrencies which use "proof of work" algorithms (not all do) do so to put a limit on the contributions one user can make to the consensus ledger; the ledger is trusted only as long as everyone believes this is true. Mine too many blocks in a row, and you have the biggest crash imaginable as everyone cashes out of a ledger which no longer has that trust basis. Everyone not caught up in the bubble will thank you for giving their electricity and GPUs back, but you won't end up rich. – IMSoP Aug 06 '21 at 09:12
  • @IMSoP: The fix for that is to limit the contributions you make to any one network and diversify. Doesn’t matter how many currencies you’re mining if you’re in space. – Joe Bloggs Aug 06 '21 at 09:31
  • 2
    @Schmuddi: Then short the currency, make your existence known, reap the rewards, and feel virtuous because you singlehandedly stopped cryptocurrency from killing the planet. Use your vast computing power to run a minecraft server or something. – Joe Bloggs Aug 06 '21 at 09:34
  • 1
    @Schmuddi no, that computer won't be enough, by a mile (surprised me, when I calculated it) if we talk bitcoin, not sure about others – MolbOrg Aug 06 '21 at 11:46
  • 2
    @MolbOrg you can get some decent profit mining Ethereum, but nowhere close to controlling the net with only 1.9 zettaFLOPS – Alexander Aug 06 '21 at 19:00
  • @Alexander yep. Wasn't commenting on the answer as it belongs to first ideas which comes to mind when we talk about significant computation power, and sure some one had to mention it, but in essence it not a business but using existing opportunity to participate in speculative asset, this or another way. This q could be more substanfial, intersting, if OP is crypto entusiast and if it easy to do for him - how much satoshis do we expect from that zettalion flops, how big is the opportunity, lol. Otherwise it fine answer, and sure crypto world dominance is out of q because lack of power, lol – MolbOrg Aug 07 '21 at 06:44
3

Sell computational power

While I agree with Joe Bloggs that the best use (at least at the moment) for this computer is crypto-mining, a secundary possibility is to lend your huge computational power to companies

A hugely parallel architecture seems very well suited to solve problems in modeling complex systems (using the finite element method).
I think a lot of pharmaceutic, automotive and aeronautic companies (without even mentioning dinosaur-cloning bilionaires) will be very happy to buy computational power from you in order run the most complicated simulations.

Probably also mathematicians will be happy to have access to your architecture (see this question on math stackexchange about problems that can't be solved at the moment because of lack of computational power)

McTroopers
  • 7,459
  • 1
  • 13
  • 26
  • hmm, interesting link, do they really have money, however, a bit of a shameful activity(IMHO) to rob them for money, they should have some percentage for free, what if aliens indeed ask for R(6, 6), we have to be prepared, not talking that they are pushing unknown away, and no one besides them can do that. Those who train ai's yep those have to pay - overlord tax. Soo back to protein folding and drugs, chemistry, material science calculations, okay, got it. – MolbOrg Aug 06 '21 at 16:09
2

Make it shine

These 80s Super computers were indeed very big and they had huge led panels ! Find a sponsor, upload a program to show his logo on the led panel, flashing through the skies world wide..

Goodies
  • 14,909
  • 1
  • 8
  • 43
  • Unfortunatly, led were not included, because of bill of materials limitations. Owner has idiosyncrasy against advertising, so retrofiting leds not in the plans for foreseeable future. But yah, q nèds to be updated by where the system is located, it is in L1 point, so it always sunny there, it how the thing get's its energy, so leds won't do at any rate. – MolbOrg Aug 07 '21 at 07:37
2

Contract with the NSA to brute force break encryption keys. That project would have low communications load but huge parallel processing load.

There would be two advantages for this. The NSA has an unlimited budget for projects like this. And they would be able to offer "protection" against those who are coming after you for the gambling debts.

David R
  • 5,653
  • 1
  • 7
  • 18
  • Yes, it can be an option, but it somewhere down on the list. I would easily comeup with the reasons why, but let's say I would preffer business oriented activities distanced from both sides crimianl and gov(enter 1 buck, exit a 100), just regular B2B or B2C activity. But yes, objectively it is a possibility. – MolbOrg Aug 07 '21 at 15:50
1

Recycle the hardware for the gold in it.

When you have 50 year old technology in a hard radiation environment, most of it won't work. The Hubble Space Telescope electronics are failing after 30 years and some have already been replaced with newer. My Mac Plus (68000) from 1986 will turn on, the floppy disk still spins, but other parts don't work anymore. Over 50 years, metals start migrating and internal chip connections fail because of metal migration. Long term space spec. memory of that day is iron core which is vulnerable to "flipped bits" (see https://retrocomputing.stackexchange.com/questions/4647/core-memory-stability)

I really doubt that you have any computational power left. And if you did, trying to find which processors still worked, what memory cells still would hold values, which local disks still spun at the correct velocity, and what cables still worked would be a huge challenge that only someone in slavery might undertake. You would be needing to physically restructure the system - perhaps daily. Look at how data centers of today swap out equipment after a very short operational life (such as 18 months) - because the failure rates start adding up.

The most value that you have there is in the precious metals and space spec. hardware will have more than usual (and stuff made back then used more than today).

(If you have the technology to physically visit this system and return, you could use that technology to visit the moon and bring back both samples and souvenirs from the Apollo missions. Now, there is a way to make some serious money. Moon rocks and souvenirs from the Apollo missions where you can prove that it was on the moon bring huge bucks.)

David R
  • 5,653
  • 1
  • 7
  • 18
  • 1
    What makes you think that this answer helps anyone and serves any purpose other than being a "gotcha?" The premise of the question was a working system and the asker wants to know what they could do with it, and you're here saying "actually, it doesn't work. Just recycle it."? – Henry Shao Aug 07 '21 at 00:11
  • Thanks for your concerns, this was ommited in the q, to not clog it with unnecessarly details. But every piece of hardware is spanking new, less than a year old. The system has full production stack/manufacturing capabilities for the technologies used, and it recycles brocken and reaching end of life components and those are replace with new ones. – MolbOrg Aug 07 '21 at 07:25
  • It is thanks to the stack it is how it got there in the first place, from the moon btw, so send your coordinates and how much moon rocks do you need, will deliver it to outpost(ISS) for you to pickup, or directly. There is some business activity on the moon, including rent time to drive rovers at perimeter of appolo and sovets and china programs without direct drive over the places (for safety reasons, do not wana them nuke me, lol) but moon activity is a different topic. So yes, there are plenty of stuff already, behind the scene, but atm I'm interested in this specific q. – MolbOrg Aug 07 '21 at 07:27
  • @HenryShao No, the question was how to extract the most value out of such a system. That is a business question, not a how to operate it question. Business questions need to look at all the factors. Now, the comments add information that change the business value of the system from how it was originally described. – David R Aug 07 '21 at 14:18
  • Spirit of the question, not the letter of the question. – Henry Shao Aug 07 '21 at 18:49
  • Hmm, the small story of the q is not the story, for reason that anything placed here is CC-BY license, and ownership is of SE, that why it relatively pointless to makeup stories as in q's so as in answers, so q story is just a wrapper. So there is a real story behind the scene, which explains how the device come to existance and whole chain of events starting from a small startup. Stepping outside a scope of a q and assuming too much it means stepping on a territory where things may be already set, not always, but this q is the case. Those days it more a history of old wb, just so u know. – MolbOrg Aug 07 '21 at 20:02
  • It just has no gold in it, it uses aluminium for contacts, a trick taken from modern days and in a vacuum it is as good as gold. However it is true that as scrap it has a value, but delivering something from the moon in million of tons is not a problem(story wise), the problem is that current space market is relatively small and very rigid, there are some opportunities, may be even for a billion a year, but this thing is huuge - can it make a little bit more, like a 10 or a 50B, could be helpfull. Provided those comments just to satisfy your curiosity(if there was some), no other purposes. – MolbOrg Aug 07 '21 at 20:15
1

Hold other satellites hostage

As per Rogue Ant's answer, this computer can generate a few orders of magnitude more power than our current civilization uses. That much power, if used at once, would vaporize the satellite. If it runs for one second at max overclock, the total power spent is ~2.61 $\times$ 1013 Wh (26.1 TWh). According to what is probably the most quoted table in this site, that is about half the yield of the original Tsar Bomba.

The satellite would become a shower of particles that, hours later, might mess up a few satellites - maybe the ISS as well. That would be too expensive for the world to risk. So you can charge the world leaders the sum of one million dollars in order to not do that.

The Square-Cube Law
  • 141,440
  • 29
  • 264
  • 586
  • Aren't you stretching the definition of shady a litle bit? Lol. Haven't seen this movie btw, but I guess it probably again a case of fantasy being inferior for potencial reality(seen clips so I'm aware it is a comedy(?)). The thing can burry (almost) any continent with half meter blanket of trash/dust from the sky. I think Mr. Evil would be eligible to ask for fort knox and get what's there in such situation. Hardly even an option, but I upvote anyone who answers regardles, but you know, lol – MolbOrg Aug 07 '21 at 17:27