Google’s joint announcement with DWave has hit the airwaves and just as with the three previous generations DWave reports that their quantum annealer achieves significant speedups over classical algorithms. The good news IMO is that the scientific manuscript that underpins the press release is upfront about the main caveat, which is that the speedup all but disappears when the DWave machine is compared to algorithms optimised for the specific problem that the DWave machine solves. All of this is explained in much detail by retired DWave chief critic Scott Aaronson in a blog post and a Q&A here.
I thought this was a good opportunity to show a picture of me holding 12 generations of DWave chips. I took this at a somewhat cringeworthy DWave sales event at the University of Queensland a while ago.
The DWave sales team tried to sell the previous generation machine to our IT department (for a cool 10 million if I remember correctly), with the promise of real-world speedups in all kinds of fantastic applications. Funnily enough they never mentioned the small detail of diminishing returns once apples were compared to apples.