Quantum Supremacy: The PR Race Shames the Google

July 17, 2024

This essay is the work of a dinobaby. Unlike some folks, no smart software improved my native ineptness.

The quantum computing era exists in research labs and a handful of specialized locations. The qubits are small, but the cooling system and control mechanisms are quite large. An environmentalist learning about the power consumption and climate footprint of a quantum computer might die of heart failure. But most of the worriers are thinking about AI’s power demands. Quantum computing is not a big deal. Yet.

But the title of “quantum supremacy champion” is a big deal. Sure the community of those energized by the concept may number in the tens of thousands, but quantum computing is a big deal. Google announced a couple of years ago that it was the quantum supremacy champ. I just read “New Quantum Computer Smashes Quantum Supremacy Record by a Factor of 100 — And It Consumes 30,000 Times Less Power.” The main point of the write up in my opinion is:

Anew quantum computer has broken a world record in “quantum supremacy,” topping the performance of benchmarking set by Google’s Sycamore machine by 100-fold.

Do I believe this? I am on the fence, but in the quantum computing “my super car is faster than your super car” means something to those in the game. What’s interesting to me is that the PR claim is not twice as fast as the Google’s quantum supremacy gizmo. Nor is the claim to be 10 times faster. The assertion is that a company called Quantinuum (the winner of the high-tech company naming contest with three letter “u”s, one “q” and four syllables) outperformed the Googlers by a factor of 100.

Two successful high-tech executives argue fiercely about performance. Thanks, MSFT Copilot. Good enough, and I love the quirky spelling? Is this a new feature of your smart software?

Now does the speedy quantum computer work better than one’s iPhone or Steam console. The article reports:

But in the new study, Quantinuum scientists — in partnership with JPMorgan, Caltech and Argonne National Laboratory — achieved an XEB score of approximately 0.35. This means the H2 quantum computer can produce results without producing an error 35% of the time.

To put this in context, use this system to plot your drive from your home to Texarkana. You will make it there one out of every three multi day drives. Close enough for horse shoes or an MVP (minimum viable product). But it is progress of sorts.

So what does the Google do? Its marketing team goes back to AI software and magically “DeepMind’s PEER Scales Language Models with Millions of Tiny Experts” appears in Venture Beat. Forget that quantum supremacy claim. The Google has “millions of tiny experts.” Millions. The PR piece reports:

DeepMind’s Parameter Efficient Expert Retrieval (PEER) architecture addresses the challenges of scaling MoE [mixture of experts and not to me confused with millions of experts [MOE].

I know this PR story about the Google is not quantum computing related, but it illustrates the “my super car is faster than your super car” mentality.

What can one believe about Google or any other high-technology outfit talking about the performance of its system or software? I don’t believe too much, probably about 10 percent of what I read or hear.

But the constant need to be perceived as the smartest science quick recall team is now routine. Come on, geniuses, be more creative.

Stephen E Arnold, July 17, 2024

Written by Stephen E. Arnold · Filed Under AI, Marketing, News, Technology

Comments

Got something to say?

Search the site
Subscribe to Beyond Search
Feature archive
News archive

Stephen E. Arnold monitors search, content processing, text mining and related topics from his high-tech nerve center in rural Kentucky. He tries to winnow the goose feathers from the giblets. He works with colleagues worldwide to make this Web log useful to those who want to go "beyond search". Contact him at sa [at] arnoldit.com. His Web site with additional information about search is arnoldit.com.