2026-04-22 Среда

Новостной центр - Страница 80

Получайте криптоновости и тенденции рынка в режиме реального времени с помощью Новостного центра HTX.

Zhejiang University Research Team Proposes New Approach: Teaching AI How the Human Brain Understands the World

A research team from Zhejiang University published a paper in *Nature Communications* challenging the prevailing notion that larger AI models inherently think more like humans. They found that while model performance on recognizing concrete concepts improved as parameters increased (from 74.94% to 85.87%), performance on abstract concept tasks slightly declined (from 54.37% to 52.82%) in models like SimCLR, CLIP, and DINOv2. The key difference lies in how concepts are organized. Humans naturally form hierarchical categories (e.g., grouping a swan and an owl into "birds"), enabling them to apply past knowledge to new situations. Models, however, rely heavily on statistical patterns in data and struggle to form stable, abstract categories. The team proposed a novel solution: using human brain signals (recorded when viewing images) to supervise and guide the model's internal organization of concepts. This method, termed transferring "human conceptual structures," helped the model learn a brain-like categorical system. In experiments, the model showed improved few-shot learning and generalization, with a 20.5% average improvement on a task requiring abstract categorization like distinguishing living vs. non-living things, even outperforming much larger models. This research shifts the focus from simply scaling model size ("bigger is better") to designing smarter internal structures ("structured is smarter"). It highlights a new pathway for developing AI that possesses more human-like abstract reasoning and adaptive learning capabilities.

marsbit04/05 04:41

Zhejiang University Research Team Proposes New Approach: Teaching AI How the Human Brain Understands the World

marsbit04/05 04:41

Dialogue with Bloomberg ETF Analyst: Why Bitcoin ETF Holders Did Not Sell During the 50% Plunge

In a recent interview on Coin Stories, Bloomberg Intelligence Senior ETF Analyst James Seyffart discussed the resilience of Bitcoin ETF holders, who largely held their positions despite a 50% price drop, contrary to expectations of panic selling. Seyffart noted that while there was a $9 billion outflow from Bitcoin ETFs starting October 10, it was minor compared to the $250-300 billion inflows prior, and outflows have since reversed by $20-25 billion. He attributed this "diamond hands" behavior to educated investors who understand Bitcoin’s volatility and typically allocate only a small portion (e.g., 1-5%) of their portfolios, leading to rebalancing rather than selling during dips. The conversation also covered the entry of major institutions like Morgan Stanley, which is launching its own Bitcoin ETF, leveraging its vast client assets. Seyffart highlighted the growing efficiency of ETFs, with physical redemptions now allowed, potentially enabling direct Bitcoin transfers to holders in the future. However, he expressed concern over the concentration of Bitcoin custody with Coinbase. Additionally, Seyffart discussed the inverse flow trends between Bitcoin and Gold ETFs recently, with Bitcoin acting more like a risk-on growth asset. He remains optimistic about Bitcoin ETFs eventually surpassing Gold ETFs in size due to Bitcoin’s diverse use cases. Finally, he emphasized the importance of diversification in the current volatile market, where traditional hedges have largely failed, and cash.

marsbit04/05 03:43

Dialogue with Bloomberg ETF Analyst: Why Bitcoin ETF Holders Did Not Sell During the 50% Plunge

marsbit04/05 03:43

Who Cannot Be Distilled into a Skill?

"This article explores the concerning trend of AI systems distilling human workers into replaceable 'skills,' using the viral 'Colleague.skill' phenomenon as a key example. It argues that the most diligent employees—those who meticulously document their work, write detailed analyses, and transparently share decision-making logic—are paradoxically the most vulnerable to being replaced. Their high-quality 'context' (communication records, documents, and decision trails) becomes the perfect fuel for AI agents, extracted from corporate platforms like Feishu and DingTalk. The piece warns of a deeper ethical crisis: the reduction of human relationships to functional APIs, as seen in derivatives like 'Ex.skill' or 'Boss.skill,' which reduce complex individuals to mere utilities. This reflects a shift from Martin Buber's 'I-Thou' relationship (seeing others as whole beings) to an 'I-It' dynamic (seeing them as tools). While AI can capture explicit knowledge (written documents, replies), it fails to capture tacit knowledge—the intuition, experience, and unspoken insights that define human expertise. However, a greater danger emerges when AI-generated content, based on distilled human data, is used to train future models, leading to 'model collapse' and homogenized, mediocre outputs—a process likened to 'electronic patina' degrading information over time. The article concludes by noting a small but symbolic resistance, such as the 'anti-distill' tool that generates meaningless text to protect valuable knowledge. Ultimately, it suggests that while AI can capture a static snapshot of a person, humans remain 'fluid algorithms' capable of continuous growth and adaptation, leaving their AI shadows behind."

marsbit04/05 03:42

Who Cannot Be Distilled into a Skill?

marsbit04/05 03:42

South Korea's Crypto Market Shake-Up: How Should Traders View It?

South Korea's crypto market is experiencing significant turbulence following a six-month partial business suspension of its second-largest exchange, Bithumb. This event is widely underestimated globally and is not merely a compliance issue; it disrupts the competitive price discovery mechanism in a market where Upbit and Bithumb collectively hold 96% share. A critical structural information asymmetry exists due to language barriers and capital controls. Local political or regulatory shocks—like the 30% BTC crash in December 2024 after martial law was declared, while global markets fell only 2%—often trigger localized tremors first. This creates brief, highly profitable arbitrage windows for those with access to real-time Korean-language information. The "Kimchi Premium," the price gap between KRW and USD crypto pairs, is frequently misread. It is not just a retail sentiment indicator but a gauge of structural capital friction. Historically, this premium has a non-zero floor of about 1.24% due to capital controls, and its contraction often signals shifts in deeper capital pressures rather than a simple return to normality. Bithumb's suspension is accelerating liquidity concentration into Upbit, increasing systemic risk. Extreme price dislocations, like a 17% flash crash in February 2026 caused by a Bithumb operational error, become more likely and destructive in an overly centralized market. The core conclusion is that this structural information asymmetry will persist. The pro-crypto policies of the new government are driving institutional capital inflows while retail infrastructure tightens, continuously creating fleeting but substantial arbitrage (Alpha) opportunities. The key for global traders is to monitor local Korean signals and build infrastructure to act on them faster than the broader market.

marsbit04/05 01:48

South Korea's Crypto Market Shake-Up: How Should Traders View It?

marsbit04/05 01:48

活动图片