Jump to content

Welcome to FutureTimeline.forum
Register now to gain access to all of our features. Once registered and logged in, you will be able to create topics, post replies to existing threads, give reputation to your fellow members, get your own private messenger, post status updates, manage your profile and so much more. If you already have an account, login here - otherwise create an account for free today!
Photo

Beyond Moore's Law


  • Please log in to reply
3 replies to this topic

#1
starspawn0

starspawn0

    Member

  • Members
  • PipPipPipPipPipPipPip
  • 1,233 posts
If you look at the amount of computing performance of a chip per constant dollar that you can buy, you indeed see that it varies along a nice, smooth exponential "Moore's Law" curve (or at least did until recently). People have come up with all sorts of reasons for this; however, there is nothing really preventing a sudden deviation from that curve, through a paradigm-shift in hardware or manufacturing. Here, I want to discuss just one, admittedly far-fetched, way this could happen.

Imagine if, suddenly, the labor required to manufacture chips and run computers became very, very cheap. If you think about it, labor is the main contributor to the cost of chips -- even the energy and raw materials required to manufacture chips comes down to labor (labor to build and maintain solar panels to collect the energy; and labor to extract those raw materials). So, if labor suddenly became cheap, the chips would become cheap, and you will have beaten Moore's Law.

How could this happen? Well, suppose you could build a robot that when you turn it loose on the moon (where property rights won't be an issue), (a) Makes copies of itself, including solar panels to extract energy to operate; and, then, after there are enough copies, (b) Those copies all go to work manufacturing computer chips (using raw materials found in moon dust and rock) that are then sent back to earth. It only will have cost you the effort to build that first robot and ship to send it to the moon; and, in return, you will have billions and billions of essentially free computer chips. Granted, these chips will be current-generation versions; but because they will still be essentially free, you will have beaten Moore's Law, if it happens in the next several decades (any longer, and the difference between "essentially free" and "free" may start to matter).

Another variation: you could also imagine such a robot (and its copies) building a giant computer on the moon, powered by solar panels that it also constructs, acting as "cloud computing" resources for users back on earth. Again, you'd have seemingly limitless compute for just the cost of building and deploying that first robot.

There are many more, much more plausible, possibilities like this out there that we may run into in the next several years, especially as AI and robotics get more and more capable.
  • Casey and Yuli Ban like this

#2
tomasth

tomasth

    Member

  • Members
  • PipPipPipPipPip
  • 228 posts

The exponential-turning-of-space-resources-into-computer-hardware-by-autonomous-system is a well known trope , automateing computer manufacturing from scrach can't be easy , and in space , even more.

 

If you can turn the steroid belt into computer hardware autonomously , one can build settlements out of them.


  • starspawn0 likes this

#3
starspawn0

starspawn0

    Member

  • Members
  • PipPipPipPipPipPipPip
  • 1,233 posts

You're right, it is a common trope.  My specific reason for citing it, however, was not to present a new science fiction scenario, but to illustrate how much cost is due to labor.

 

I actually had a completely different scenario in mind, involving the manufacture of imperfect chips + "translation / multiplexing"; but then thought it too technical, that it would overpower the posting.  So, I resorted to robots on the moon, instead.



#4
Kynareth

Kynareth

    Member

  • Members
  • PipPipPipPip
  • 160 posts

Cost to design a 28nm planar device ranges from $10 million to $35 million, cost of designing a 3nm device will range from $500 million to $1500 million. This is clearly a BIG problem. Costs are going through the roof and I don't know how companies can lower them to acceptable levels. This isn't about if you CAN design and produce a 2nm 3D chip but HOW MUCH will it cost.






0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users