It's not just spin – boffins give quantum computing a room-temp makeover

Practical quantum computers are still on the horizon, but scientists continue to make improvements in the underlying technology required to make such systems possible. Two research teams have published on alternative approaches to quantum, one on manipulating the spin of electrons at room temperature, the other on using …

[Pascal Monett]

"difficult to implement on today's error-prone machines"

Ho hoo, burn ! Is that why Azure is always falling over ? I thought it was good old fat-fingered misconfiguration.

I wonder what Intel thinks of that description, especially when selling its Xeon Platinum 8280L beasts at almost $37,000.

Or AMD, with its massive €8,700 Threadripper.

Or nVidia, with its $30K A100 supercomputing monster.

There are a lot of people making a ton of money on those error-prone machines. And I don't see that they are all that error-prone. Yes, there's the occasional unexplainable crash every now and then, and yes, Borkzilla has trained us well in the three-fingered salute to recover from Windows bluescreens, but Excel tables are generally quite reliable (when humans have written the right formulas), and CPUs don't generally fish data from the wrong memory cell.

So I think error-prone is a bit of an exaggeration.

Just a bit.

[Roj Blake]

Re: "difficult to implement on today's error-prone machines"

I think it's referring to today's error-prone quanrtum machines...

[Mike 137]

Re: "difficult to implement on today's error-prone machines"

It is indeed referring to quantum machines, and it's extremely probable that "today's" is based on unconscious optimism.

As a basic and immutable principle of physics, the smaller the signal the more susceptible it is to disturbance by outside influences. So at a certain point as we miniaturise (which is essentially what's going on as we go to quantum) the support needed for error correction will exceed the capacity for data processing. Particularly under normal ambient conditions where there's a huge amount of noise, that trade off may well eliminate a lot of the prospective performance benefits. Consequently, I doubt we'll ever have quantum computing on the desktop -- it'll always be a niche for specialised jobs under strictly controlled conditions.