2026-04-22 Quarta

Centro de Notícias - Página 68

Obtém notícias cripto em tempo real e tendências de mercado com o Centro de Notícias da HTX.

Chinese Large Models: This Time, the Script Is Different

By early 2026, Chinese large language models (LLMs) have gained significant global traction, representing six of the top ten most-used on the AI model aggregation platform OpenRouter. This shift, led by models like Xiaomi's MiMo-V2-Pro, occurred after Chinese models' weekly token usage surpassed that of U.S. models in February 2026. A key driver is the substantial price gap: Chinese models are often 10–20 times cheaper for input and up to 60 times cheaper for output tokens than leading U.S. models like OpenAI’s GPT-5.4 and Anthropic’s Claude Opus. This cost advantage became critical with the rise of agentic applications like OpenClaw, which automate complex tasks (e.g., programming, testing) and consume tokens at a much higher volume than traditional chat interfaces. While U.S. models still lead in complex reasoning benchmarks, Chinese models have nearly closed the gap in programming tasks—evidenced by near-parity scores on the SWE-Bench coding evaluation. This enabled cost-conscious developers, especially in AI startups using open-source stacks, to adopt a "layered" approach: using Chinese models for routine tasks and reserving premium U.S. models for harder problems. Rising demand led Chinese firms like Zhipu and Tencent to increase API prices in early 2026, yet usage continued growing sharply. Analysts note that China’s cost edge stems from large-scale, efficient compute infrastructure and widespread adoption of MoE (Mixture of Experts) architecture. Unlike the low-margin electronics manufacturing analogy ("AI-era Foxconn"), Chinese LLM firms are demonstrating pricing power and rapid technical advancement, suggesting a different trajectory from traditional assembly-line roles.

marsbit04/07 11:00

Chinese Large Models: This Time, the Script Is Different

marsbit04/07 11:00

Assembling the Most Accurate Gold Forecasters in History, Can We Crack the Future Price of Gold?

The article investigates whether compiling the most accurate historical predictions from top analysts, institutions, and influencers can unlock a reliable method for forecasting future gold prices. Using gold as a case study, it examines predictions from sources like the LBMA, Goldman Sachs, JPMorgan, prominent figures like Peter Schiff and Jim Rickards, and celebrated forecasters such as Nouriel Roubini and Ben McMillan. The findings reveal significant inconsistencies. Major institutions often exhibit "lagging predictions," adjusting forecasts too slowly to match rapid market moves—for instance, LBMA’s 2025 consensus underestimated the actual average price by 20%. Influencers like Schiff and Rickards persistently advocate for higher long-term targets (e.g., $5,000 to $35,000) but their predictions lack precise timing, often requiring investors to endure prolonged periods of underperformance. Even "accurate" forecasters like Roubini and McMillan have mixed records, with notable misses alongside their successes, while Ray Dalio’s broad allocation advice (5-15% gold) proves more practical than specific price targets. The analysis notes eerie similarities between the 2011 gold peak—where extreme predictions clustered near the market top—and the 2026 crash, where many experts maintained bullish outlooks despite a 25% plunge. Current predictions for future prices vary wildly, from $5,400 to $35,000, highlighting a lack of consensus. The conclusion is that no consistently accurate predictor exists. The author argues that gold forecasting is inherently uncertain, dominated by occasional lucky calls rather than reliable expertise, and advocates for a diversified, long-term investment approach over chasing speculative forecasts.

marsbit04/07 10:16

Assembling the Most Accurate Gold Forecasters in History, Can We Crack the Future Price of Gold?

marsbit04/07 10:16

175-Year-Old Western Union: Not Just Playing with Stablecoins, but Also Acquired a Digital Wallet

At 175 years old, Western Union, the global money transfer giant, is undergoing a significant digital shift. After a failed 2018 experiment with Ripple's XRP due to high costs, the company is now aggressively embracing blockchain and digital assets. In April 2026, Western Union acquired Singapore-based digital wallet Dash from Singtel, marking its first digital wallet asset in the Asia-Pacific region. Dash, with 1.4 million users, offers a full suite of services including payments, remittances, savings, insurance, and investments, deeply integrated into Singapore's local life. This move is part of a broader strategy to modernize its legacy business. While Western Union's vast network of over 500,000 physical agent locations remains its backbone, it is also its biggest cost burden. The company faces intense competition from digital-native rivals like Wise and Remitly, which offer significantly lower fees. To compete, Western Union is building a "Digital Asset Network." A key component is its own USD-pegged stablecoin, USDPT, issued on the Solana blockchain in partnership with Anchorage Digital. It is also piloting a stablecoin-linked Visa card with Rain for users in high-inflation countries like Argentina, allowing them to spend or cash out dollars at its agent locations. The acquisition of Dash represents a fundamental change: moving from being a transient "pipe" for money transfers to building a destination where users stay. Dash provides a trusted, established platform to test and deploy these new digital products, serving as a launchpad for Western Union's expansion across the Asia-Pacific region.

marsbit04/07 09:46

175-Year-Old Western Union: Not Just Playing with Stablecoins, but Also Acquired a Digital Wallet

marsbit04/07 09:46

Robinhood Gains a New Batch of Stock Investors, the Oldest is 1 Year Old, the Youngest is -3 Years Old

US Treasury designates Robinhood as broker and initial trustee for "Trump Accounts" (also known as 530A accounts), a tax-advantaged investment program established under the "Big and Beautiful" Act. The initiative, aimed at children born between January 1, 2025, and January 1, 2029, provides each eligible newborn with a $1,000 initial federal deposit. Private donors, such as Michael Dell, have also contributed significantly. The accounts are restricted to low-cost index funds tracking broad market indices like the S&P 500, and funds are locked until the child turns 18. With an estimated 14.4 million children eligible, the program represents a potential $14.4 billion in initial government funding, growing to hundreds of billions with private and family contributions. This creates a long-term, passive investment pool of potentially trillions of dollars. Robinhood, selected over competitors like JPMorgan and Charles Schwab, stands to be a major beneficiary. The deal grants the platform millions of new, long-lifecycle users—the oldest are one year old, the youngest are yet to be born—who are effectively locked into its ecosystem. Upon turning 18, these accounts convert to IRA-like retirement accounts, allowing Robinhood to capture their adult investing activity. The stable, long-term, and predictable nature of these assets also diversifies Robinhood's business beyond its reliance on active retail trading. The government endorsement significantly boosts its institutional credibility and opens new avenues in wealth management.

Odaily星球日报04/07 09:31

Robinhood Gains a New Batch of Stock Investors, the Oldest is 1 Year Old, the Youngest is -3 Years Old

Odaily星球日报04/07 09:31

活动图片