Interview
a16z Podcast | Quantum Computing, Now and Next
- Classical transistor scaling faces fundamental physical and economic limits with diminishing returns, prompting a shift toward specialized ASICs, neuromorphic handsets, and FPGA-based cloud servers.
- Intel targets a 7-nanometer node despite associated manufacturing costs described as "monumental," while a manufacturing "Cambrian explosion" is expected to enable $10 million investments to rival $4 billion supercomputers.
- Neuromorphic chips are slated for deployment in individual handsets "soon," whereas quantum computers will soon address the primary bottleneck of algorithm development over the "next few years."
- Future systems will employ "quantum-classical hybrid algorithms" and a developing full-stack software environment to integrate instruction sets across cloud infrastructure.
- Rigetti Computing plans to expand its quantum software engineering workforce from "approximately zero" to a "meaningful number" within the "next five years or so" to foster a hardware-software flywheel.
- While full-stack quantum companies may gain an edge over incumbents through specific organizational DNA, practical machines capable of running Shor's algorithm for "practically relevant problem sizes" remain "20 to 30 years away."
- A future capability exists for quantum computers to crack RSA encryption, creating potential risks regarding the "shelf life" of currently encrypted secrets.
- Quantum computing is anticipated to disrupt wet chemistry, material science, and drug design, with significant projected impacts on health care, energy generation, and food production.