It has been a while since we thought about computers and thought about Honeywell. Sure, they had a series of computers they bought from General Electric and Computer Control Company in the 1970s. Even before that they joined with Raytheon and produced vacuum tube computers that later morphed into transistor-based computers. But in recent years, you are more likely to think of Honeywell for thermostats, air filters, and industrial controls. But now, Honeywell has come out of the computer shadows with some impressive quantum computer hardware and they clearly have big plans.
Comparing quantum computers is a bit dicey just as, for example, judging CPUs by instructions per second has its problems. In the past, vendors have jockeyed for the maximum number of qubits, but that’s misleading in some cases. Processing power depends on the number of qubits, their quality, and how they are connected. IBM introduced the idea of quantum volume and Honeywell claims their new machine will hit 64 by that measure, twice that of anyone else’s quantum computer that we know about.
What’s more, is they’ve promised to increase the volume by a factor of ten each year. The company plans to make their computers available via the Microsoft cloud.
According to Honeywell, their use of trapped ion qubits is superior to other computers that use some indirect method which is more prone to noise. Of course, the computer operates in an exotic environment, which Honeywell is used to handling.
Want to know more about quantum computing? Check out our series using (mostly) Quirk. Who knows? You might be able to build your own one day.
By a factor of 10…
A bit into the future: “I want to submit a long term project for Quantum resolvement.”
“OK, just pay the bill over there and I get the results for you now.”
No fair! In observing it you changed the outcome!
I’m curious who gets to see your queries and results apart from Microsoft and Honeywell ?
It may be an appealing way to try before you buy, but a lot of potential customers will not want the ingredients of their secret sauce leaving the premises uncooked.
Honeywell?? Isn’t that a once credible brand-name from decades ago that through MBA led corporate greed drove themselves (along with many excellent Engineers) out of business only to see the trademark resurrected by yet more MBA vultures so they can sell cheap Chinese crap like over-priced sunglasses, flash lights, etc.? Or maybe I’m confusing the Honeywell brand name with retread crooks like Polaroid, Bell and Howell, RCA, etc., etc.
Honeywell has a long history. Albert Butz started the Butz Thermo-Electric Regulator Company in 1885. It was eventually renamed and bought out by W. R. Sweatt — what names!
Mark Honeywell started Honeywell proper in 1906 and it merged with Sweatt’s company in 1927.
There are several divisions, but I don’t think any of them do as seen on TV flashlights. They did offer a kitchen computer back in 1969 for $10,600. It had switches and binary lights and the cook that wanted to use it had to take a two week class to learn how to use it. As far as we know, no one ever bought one.
They’ve had a solid control systems business for decades. If you really want to do what IOT only promises in an industrial/commercial environment, you go to Honeywell.
Honeywell has had some consumer retail exposure, but mostly things like HVAC thermostats and filtration systems and my house alarm system is Honeywell. They were never like GE with consumer goods and I don’t think they’ve sold/licensed their name for consumer retail to the same extent as GE, Polaroid, Bell & Howell, or RCA. They are still quite big in industrial and aerospace.
I don’t recall that they are out of business. Maybe the retread is you.
The claims here are highly overblown. Trapped ion qubits are very, very slow, for a number of reasons. For one thing, the entangling process takes tens of microseconds. For another, the ions have to be *physically moved* between operations. So, barring some incredible breakthrough, ion-based quantum computers will never scale beyond perhaps tens of qubits.
I were going to say that this article seems a bit lacking when it comes to content.
But going through Honeywell’s own statements about their technology is not really giving all that much more information.
The two main things I could gleam is that their system is reconfigurable, and that their q-bits are more defect free than what others supposedly are.
Honeywell also makes rather wide reaching comments about the future implications of quantum computing. Like it will make everything better! Not to mention how Honeywell apparently have 130+ years of experience in relevant areas. (Being founded 130+ years ago is indeed true, working on things even remotely relevant for quantum computers? Nope….)
It is a lot of generally bold claims, but no information of how they intend on doing it.
Quantum computers though seems to follow the same hype train as deep learning still is partly on.
Yes, a new technology can have advantages, but seldom is something a universal cure to all your ills. What quantum computers will actually be good for, that only time will tell. But for now, lets not over-hype about something that few people can even explain how to use for real applications.
Though, my own two cents about the far off future of quantum computers is that it likely will just be another co-processor that one can add to a “standard” PC. Similar to how we have add in cards for graphics, we can have add in cards for quantum computing tasks.
Maybe they mean man-years of experience, aka 520 people working on it for just a bit over 3 months.
now i’m not sayin that quantum computation is bullshit in its entirety… buuuuuuuuuut this guy might be on to something…
https://www.quantamagazine.org/videos/gil-kalai-why-quantum-computers-wont-work/
whole Q-computer industry always seems to be a few years away from getting hello Q world working on real hardware.
A simple reason why quantum computers might be less than you’d expect is the fact that the qubits are destroyed when you read out the results. Each time you run the algorithm, you have to re-set the machine, load up all the variables, prime the qubits, and then start the algorithm.
So where’s the rub? The computation is probabilistic – it doesn’t give you a straight answer. Even if the machine works, you have to run the computation hundreds of times to have a statistical certainty of the answer. This means your computational speed is actually limited by your very classical, very linear I/O operations.
So you literally can’t just dump “everything that Google has” into a quantum computer and *ping* goes the answer to all your questions. It would take days to simply stream the data over, and that’s just one iteration. You still have to run it a thousand times over. It’s the same problem as with any big data computational effort – you’re actually slowed down by your RAM, not your CPU.