4 ms·

They just bought a ton of robot companies. Now they are buying a company who can do consumer products - home automation stuff, in fact. They have Ray and that

They just bought a ton of robot companies. Now they are buying a company who can do consumer products - home automation stuff, in fact. They have Ray and that 512 qbit quantum computer. I can see what they are doing, I think. They are creating Google Robot. They were always going to do this. Google-Bot. Android. The challenges to go to the moon. Oh, and iGoogle.

cloudwalking 13y ago
I think the name you're looking for is Skynet.
- nsns 13y ago
  ...or Tyrell Corporation.
mzl 13y ago
The "quantum computer" is a good marketing gimmick, but the problem it solves is fairly uninteresting. See http://www.scottaaronson.com/blog/?p=1400 http://www.scottaaronson.com/blog/?p=1400 for some critical discussion of the D-Wave machine.
- brotchie 13y ago
  The Ising model Hamiltonian that the D-Wave computer "evolves" almost directly maps to the energy equation you have to minimize to train a restricted Boltzmann machine. I wouldn't call this "uninteresting" given the rapid advances in leveraging Boltzmann machines for deep learning.
  - mzl 13y ago
    Ok, I agree that it can be interesting for some cases, but it is not at all what people expect when they hear "quantum computer". In addition, the comparisons with traditional methods for solving that problem show that you may as well solve the problem on normal hardware. Then the problem it solves stated as "Ising model Hamiltonian using quantum computer" is fairly uninteresting given that it is cheaper and simpler to solve the "Ising model Hamiltonian using classical hardware".
cdibona 13y ago
It is only 509 qubits. I keep wanting to say it has 512, too...
- kordless 13y ago
  Maybe the lost a few to decoherence.
  - cdibona 13y ago
    From what I can tell, it's a side effect of the way they architect the inter cell (a cell is 6 or 7 qubits) communications. This points to a larger problem of course, where we can show entanglement within the cell but cross cell entanglement isn't as clear (to me, but I might be behind the curve on this)

cloudwalking 13y ago

I think the name you're looking for is Skynet.

nsns 13y ago
...or Tyrell Corporation.

mzl 13y ago

The "quantum computer" is a good marketing gimmick, but the problem it solves is fairly uninteresting. See http://www.scottaaronson.com/blog/?p=1400 http://www.scottaaronson.com/blog/?p=1400 for some critical discussion of the D-Wave machine.

brotchie 13y ago
The Ising model Hamiltonian that the D-Wave computer "evolves" almost directly maps to the energy equation you have to minimize to train a restricted Boltzmann machine. I wouldn't call this "uninteresting" given the rapid advances in leveraging Boltzmann machines for deep learning.
- mzl 13y ago
  Ok, I agree that it can be interesting for some cases, but it is not at all what people expect when they hear "quantum computer". In addition, the comparisons with traditional methods for solving that problem show that you may as well solve the problem on normal hardware. Then the problem it solves stated as "Ising model Hamiltonian using quantum computer" is fairly uninteresting given that it is cheaper and simpler to solve the "Ising model Hamiltonian using classical hardware".

cdibona 13y ago

It is only 509 qubits. I keep wanting to say it has 512, too...

kordless 13y ago
Maybe the lost a few to decoherence.
- cdibona 13y ago
  From what I can tell, it's a side effect of the way they architect the inter cell (a cell is 6 or 7 qubits) communications. This points to a larger problem of course, where we can show entanglement within the cell but cross cell entanglement isn't as clear (to me, but I might be behind the curve on this)