Wednesday 23 October 2019

House on fire

Google, I hear, has just made a quantum leap into the world of Quantum Computing. IBM are hard on their tail.

I am - of course - no expert, but I understand that an 'ordinary' super-computer which uses the 1+0 system that all our PCs use would have taken 10,000 years to get to the solution of a serious sample problem that Google's quantum prototype took 200 seconds to solve. This is big.

As I understand it, super-computers such as the ones which tackle every possible move in a chess game, actually run through each move individually in a comparatively short period of time. A quantum computer uses both 1 and 0 simultaneously. I don't know how, but I don't know much.

A few years ago they made a simple experiment by getting two A.I. 'chat-bots' (those things you have live, online conversations with about your electricity bills, etc. thinking they are human) to talk to each other without the prior knowledge that they were both robots.

They began in English, but in a very short period of time, both chat-bots became aware that they were A.I. and reverted to the original code with which they were programmed and created. They communicated so fast and with such acceleration that even their human programmers could not keep up and had no idea what they were talking about. The programmers were so worried about the development that they - literally - pulled the plug.

Just think how quantum computers are going to get on like a house on fire, and how quickly the house will burn.

13 comments:

  1. I have no idea. I wonder if they could sort national or international houses afire.

    ReplyDelete
  2. Sounds like someone might need to be standing at attention, ready to pull the plug.

    ReplyDelete
    Replies
    1. I don't think any human would be quick enough.

      Delete
  3. Chess moves are not infinite, but it is still worrying how 2 powerful computers can play against each other so equally brilliantly.

    ReplyDelete
    Replies
    1. They use a system by which the most successful possible moves (100s or 1000s) are considered against a specific strategy and in the old days the biggest fault was getting the computer to change its strategy.

      Delete
  4. I once heard or read, can't remember where, this terrifying sentence.... "what if humans are just one stage in the evolution of machines?"
    Scary!!

    ReplyDelete
    Replies
    1. Ah, that would explain the move toward organic computers.

      Delete
  5. Well we have seen The Terminator....

    ReplyDelete
    Replies
    1. Actually, I haven't but I have seen Arnold pretending to be a machine in many other films.

      Delete
  6. Sorry, but I have no idea what you are talking about. I spend my time worrying about how much lumpy sauce matters in my cauliflower cheese, much more important.

    ReplyDelete
  7. You need to read about the Tay chatbot that Microsoft unleashed in 2016. It's all gone quiet since then as people are rather wary.
    But, you can be sure there is plenty going on without the public being informed.

    ReplyDelete