AI summary by TwinMind
Biggest Self-Own in Quantum Computing, Ever
Sabine Hossenfelder
6m 36s
Background on Quantum Computing
- Speaker previously discussed impressive quantum computation, now revisiting due to unexpected development
- Quantum computing marketed as "next big thing" with nations investing heavily
- Speaker's view: quantum computers will remain niche, most funding wasted
- Two fundamental approaches to quantum computation exist
Two Quantum Computing Approaches
- Universal/Gate-Controlled Quantum Computers:
- Works similarly to conventional computers
- Uses qubits instead of bits
- Executes algorithms by acting on qubits
- Pursued by Google, IBM, Microsoft
- Quantum Annealers/Quantum Simulation:
- More like analog computers
- Set up to answer specific questions
- System follows laws of nature
- D-Wave specializes in this approach
D-Wave's Quantum Achievement
- Released calculation 2 months ago
- Calculated interaction of many tiny magnets
- Problem: Ising model (computationally extremely difficult)
- Performance comparison:
- D-Wave: calculation took less than 1 second
- Frontier supercomputer (Oak Ridge, Tennessee): estimated >1 million years
- First quantum claim that hasn't been debunked by conventional computers
Q Control's Challenge
- Sydney-based software company (not hardware)
- New paper on arXiv
- Created algorithm running D-Wave calculation on IBM quantum computer
- Hardware used: IBM Eagle with 127 qubits
- Results: Similar time, better accuracy than D-Wave
- Threatens D-Wave's special position
D-Wave's Response and Contradiction
- Su CH Lanting (Chief Development Officer) responded on LinkedIn
- Claimed "no threat to D-Wave's leading position"
- Critical error: Said new algorithm is "sophisticated classical algorithm," not quantum
- Inadvertently admitted quantum computer unnecessary for their calculation
- Q Control author corrected: calculation is properly quantum
Implications and Analysis
- Universal quantum computers closer to quantum simulators than previously thought
- Q Control claims better performance with right algorithm
- IBM potentially back in competition
- Ongoing cycle: claims of outperformance until someone proves otherwise
- Competition advancing the field
- Key metric suggestion: Dollars per digit, not just speed
- Real question: who can do it cheaper, not faster
Sponsor Content
- Brilliant.org partnership
- Courses in science, computer science, mathematics
- Interactive visualizations, Python scripts
- Special offer: brilliant.org/sabina for 43-day trial + 20% off annual premium
