Sushi

Golden Ten Data reported on March 31 that the Swiss National Bank sold a small amount of foreign currency at the end of last year, slightly shrinking its balance sheet while risking a stronger franc. The Swiss National Bank stated on Tuesday that it reduced its foreign exchange reserves by 6 million Swiss francs (approximately 7.5 million US dollars) in the fourth quarter. According to the institution's full-year statistics, the size of its balance sheet remained virtually unchanged. As investors looked for safe havens to hedge against the impact of the Iran war, the franc surged to a ten-year high this month. The Swiss National Bank previously indicated that it would reinforce its determination to curb excessive appreciation of the franc through intervention measures.

According to Golden Ten Data on March 31, Malaysian Transport Minister Anthony Loke stated that Malaysian oil tankers permitted to pass through the Strait of Hormuz will not be required to pay the transit fees levied by Iran. Loke said at an event on Tuesday: “We are a friendly party and maintain good diplomatic relations with the Iranian government.” Last Saturday, Malaysia's Foreign Minister said Iran had approved seven Malaysian oil tankers to cross the Strait of Hormuz, which has become a conflict hotspot in the Middle East war. The Malaysian vessels currently detained in this key waterway include ships belonging to energy giant Petronas, shipping company MISC, and Sapura Energy.

According to 1M AI News monitoring, on the same day Google released its Quantum AI Whitepaper, the neutral atom quantum computing startup Oratomic published a paper on arXiv stating that only about 10,000 reconfigurable atomic qubits are needed to run the Shor algorithm at a cryptography-relevant scale. This paper directly takes Google's optimized low-depth Shor circuit as input and further optimizes another layer of the quantum computing stack: while Google's optimization focused on reducing the required number of logical qubits (from thousands to about 1200), Oratomic's optimization focused on reducing the number of physical qubits needed per logical qubit. The combination of these two optimizations has significantly reduced the hardware scale required for breaking encryption to an unprecedented low point.Key to Oratomic's approach is the use of high-rate qLDPC codes instead of traditional surface codes. Surface codes are the current mainstream quantum error correction scheme used in Google's superconducting approach, but they have low encoding efficiency, requiring about 400 physical qubits per logical qubit, totaling around 500,000 qubits. With a qLDPC code encoding rate of about 30%, significantly fewer physical qubits are needed to protect the same number of logical qubits, compressing the total requirement by about two orders of magnitude.The paper presents several architecture scenarios (assuming a stabilizer measurement cycle of 1 millisecond):1. Around 10,000 physical qubits can run the Shor algorithm to break 256-bit elliptic curve encryption (used by Bitcoin and Ethereum), with runtime depending on parallelism2. With approximately 26,000 physical qubits configured, the runtime for breaking elliptic curve encryption is about 10 days3. With around 102,000 physical qubits configured, the runtime for breaking RSA-2048 is about 97 daysThe trade-off is speed: neutral atoms have a much lower clock frequency compared to superconducting approaches, extending the decryption time from minutes to days. However, this does not mean the threat is diminished. Google's superconducting approach (500,000 qubits, 9 minutes) is suitable for intercepting real-time broadcasted transactions; Oratomic's neutral atom approach (10-26,000 qubits, multiple days) is suitable for attacking dormant wallets with exposed public keys, which do not require a sense of urgency. Google's whitepaper estimates that about 6.9 million bitcoins fall into this category.The hardware gap is narrowing. The paper points out that neutral atom experiments have demonstrated a physical trap array with over 6,100 qubits, though these arrays have not achieved quantum computation yet; fault-tolerant neutral atom systems currently have around 500 qubits. From 500 to the required 10,000 in the paper, the gap is about 20 times, much smaller than the approximately 5,000 times for Google's superconducting approach (currently about 100 qubits vs. the necessary 500,000). The paper's authors are from Oratomic, affiliated with Caltech, including quantum computing authorities John Preskill and Manuel Endres, with Dolev Bluvstein as the corresponding author. The paper concludes that future hardware acceleration and error correction improvements are expected to further reduce the runtime by an additional order of magnitude, potentially down to hours or even minutes.