Page 1 of 1

qubits

Posted: Wed Oct 23, 2019 2:41 pm
by EdLee
One of our local Go people cameo'd in the group ( ~03:46 ~ 03:52 ), but the person is too busy qubitting to drop by the coffee shop more than once over many months. :)
Look forward to q-Go-bots.

Re: qubits

Posted: Thu Oct 24, 2019 8:05 am
by Gomoto
While I do not think go is solveable in a classic way in our time&space, I suppose the children of this chip might be able to achieve this feat.

Perhaps I even live to see ...

Posted: Thu Oct 24, 2019 9:09 am
by EdLee
Hi Gomoto,

Maybe chess will be (q-)solved first. :)

SA/Nature

New Atlas: googl v.ibm

NPR: googl v. ibm

Local: the independent

Re: qubits

Posted: Thu Oct 24, 2019 7:56 pm
by xela
It's an exciting milestone to be sure, but the name is a bit inflated. Quantum computers have just arrived at the starting line; they're not ready to launch into their victory lap yet.

So now there's a quantum computer big enough that classical computers can't literally simulate what happens inside it. If you accept the name "quantum supremacy" for this milestone, then we still have "human supremacy" over computers for chess and go. A computer can't simulate what my brain does when I play a game of go, but that doesn't stop it beating me pretty easily.

The last speaker in the video captured it well: "...find something cool to do with it". That's the bit I'm really looking forward to (and also am slightly scared of).

Posted: Thu Oct 24, 2019 9:46 pm
by EdLee
we still have "human supremacy" over computers
You know Skynet is just around the corner, right? :mrgreen:

Re: qubits

Posted: Fri Oct 25, 2019 2:11 am
by Gomoto
Solving go seems pretty cool to me.

It is not related to simulating a human brain. A quantum computer would just simultaneously play out all possible games and one would look at the maximum of the endstates. Simple, isn't it ;-)

Re: qubits

Posted: Fri Oct 25, 2019 4:22 pm
by xela
Gomoto wrote:A quantum computer would just simultaneously play out all possible games and one would look at the maximum of the endstates. Simple, isn't it ;-)
And this is the sort of misconception that just feeds the hype. Thanks for a perfect illustration!

Yes, a qubit is a superposition of states -- but observing it will change the system. You can't just peek at all the states and pick your favourite. There are two big challenges in quantum computing. One is to physically build the computers. This is where the recent news really is a significant breakthrough: jumping from computers with only 1 or 2 bits up to 50+ is a big deal. And then the second challenge is to learn how to program the things, which is much trickier than it might first look.

(Disclaimer: so far I've only made it about a third of the way through Nielsen and Chuang's excellent book. So I can't say I'm an expert here. Just slightly more informed than one might be based on the media coverage. Hoping to finish the book one day, but playing around with go AIs is currently more fun and a bit easier!)

Re:

Posted: Fri Oct 25, 2019 4:24 pm
by xela
EdLee wrote:
we still have "human supremacy" over computers
You know Skynet is just around the corner, right? :mrgreen:
My point exactly. The word "supremacy" does not mean what they think it means.

Re: qubits

Posted: Fri Oct 25, 2019 5:05 pm
by Gomoto
humankind will never put their step on the moon

Posted: Fri Oct 25, 2019 6:18 pm
by EdLee
A very nice 'chapter zero of Q.C. for dummies' :)