Suggestions
Stephen Nellis
Reporter at Reuters
Stephen Nellis is a Technology Correspondent at Reuters, based in San Francisco, where he covers technology companies and industry trends. He has a background in business journalism, having graduated from the University of Missouri-Columbia. His work often focuses on significant developments within the tech sector, including topics related to artificial intelligence, semiconductor production, and major corporate strategies.
Nellis actively shares insights and updates on his LinkedIn profile, where he discusses various technology-related topics, such as Nvidia's AI chip demand and the implications of new technologies like Apple's Vision Pro headset. His posts reflect a deep engagement with current events in the tech industry, providing analysis and commentary on market movements and corporate performance.1
Highlights
A huge, huge congratulations to my colleague Hyunjoo Jin, who was part of a team that won one of the two Pulitzer prizes awarded today for her part in a Reuters series about Elon Musk's business empire. If you didn't catch the stories, read them here, and congrats again to Hyunjoo and the entire team that made them possible.
Heads up: Compared to my usual chip and finance posts, this one is more for a niche audience!
Quantum computing could unlock huge advances in science, but one of the roadblocks has been that the fundamental building block of quantum computers called qubits are very delicate. The slightest variance in their conditions can cause noise and data errors.
For years, the usual approach to this problem was to build lots of redundant physical qubits and hope that enough of them got a calculation right to get a reliable result from a smaller subset of logical qubits. There was talk of 1,000 qubit chips and million qubit systems.
But the issue there is that stringing up so many physical qubits is incredibly expensive and, as I understand it, would require some not-yet-invented networking technologies to not lose quantum speed advantage in logical computation to the network overhead of moving/checking data.
So more recently there's been much focus on how to push the error rate of the logical qubits below the error rate of the physical qubits, while keeping the ratio of logical qubits to physical qubits low enough that you don't run into the challenges of needing to string together huge numbers of physical qubits to get 100 or so logical qubits.
Today, Microsoft and Quantinuum say they have made a major breakthrough on this front, getting four logical qubits from just 30 physical qubits. I spoke with Jason Zander and Ilyas Khan, KSG about the news for Reuters -
https://lnkd.in/gCEhYca4
