# Artikel Terkait Data Analysis

Pusat Berita HTX menyediakan artikel terbaru dan analisis mendalam mengenai "Data Analysis", mencakup tren pasar, pembaruan proyek, perkembangan teknologi, dan kebijakan regulasi di industri kripto.

AlphaFold Just Disbanded, Claude Enters Protein Design: 14 Out of 15 Targets Secured

Claude demonstrates significant progress in AI-driven protein design, achieving a 35.1% success rate in creating novel "minibinders" (small, engineered binding proteins) for specific targets. In a benchmark of 15 protein targets, it successfully designed binders for 14, with at least six showing high-affinity binding and four matching or surpassing previous human-designed results. The AI autonomously managed the entire design workflow, including generating, optimizing, and screening candidates. In a separate experiment, Claude Opus 5 autonomously analyzed raw experimental data (NMR and LC-MS files), identifying molecular structure and measuring sample purity (96.4%) within minutes, with results closely matching laboratory analysis (96.33%). These experiments indicate AI's potential to reduce the expertise, cost, and time required in life sciences research, from end-to-end protein design to analytical chemistry tasks. However, capabilities remain uneven across different protein targets, and access to the most powerful models for sensitive tasks like drug design is currently restricted under a forthcoming scientist access program.

marsbit08/20 03:01

AlphaFold Just Disbanded, Claude Enters Protein Design: 14 Out of 15 Targets Secured

marsbit08/20 03:01

20% of American Workers Are Offloading Tasks to AI, Where Tasks Are Replaced, Not Jobs

A recent survey by Epoch AI and Ipsos reveals that 20% of US workers report that AI has now fully or mostly taken over at least one task they previously outsourced to colleagues or contractors. The key finding is that AI is currently replacing specific *tasks*, not entire *jobs*. The study examined ten common knowledge-work tasks. While AI usage is widespread—ranging from 25% for maintaining records to 57% for software design—it rarely handles a task completely. In software design, for instance, only 10% of workers reported AI doing most or all of the work. AI's impact on time efficiency is mixed: 53% of tasks where AI does most of the work see reduced time, but about one-sixth of all AI-assisted tasks actually become *more* time-consuming. Furthermore, while 66% of AI outputs are used with little or no modification, this does not necessarily indicate high quality. Researchers note that clearly defined, deliverable tasks—traditionally suited for outsourcing—are most susceptible to AI takeover. This shift pressures task-based contractors more than it eliminates full-time roles. Adoption is also uneven, concentrated among higher-income, college-educated white-collar workers. The report concludes that the core dynamic is a reorganization of work between humans and AI. The critical question for workers is not "Will AI replace me?" but "How many of my job's components can be packaged as discrete, outsourceable tasks?"

marsbit08/17 08:14

20% of American Workers Are Offloading Tasks to AI, Where Tasks Are Replaced, Not Jobs

marsbit08/17 08:14

Analytical platform Santiment reports that the opinion that 'cryptocurrencies are dead' is rapidly spreading among investors! Is this a signal that the market has bottomed? Here...

The analytical platform Santiment has reported a rapid spread of the sentiment that "cryptocurrency is dead" among investors, raising the question of whether this is a signal of a market bottom. Persistent price pressure and low trading volumes continue to negatively impact investor morale. Santiment observed a significant increase in this negative rhetoric on platforms like Reddit and Telegram, noting such phrases typically emerge when retail investor patience wears thin. Historically, periods of extreme pessimism often occur near major market turning points. When public conviction in an upward trend vanishes, markets frequently react contrary to expectations with strong recoveries. Santiment advises contrarian investors to monitor such widespread "crypto is dead" narratives, as they may signal potential opportunities for patient, long-term investors. While fear may exert short-term pressure, past cycles show excessively negative sentiment has coincided with significant lows. Social media sentiment remains a key indicator for assessing market direction.

cryptonews.ru08/14 08:25

Analytical platform Santiment reports that the opinion that 'cryptocurrencies are dead' is rapidly spreading among investors! Is this a signal that the market has bottomed? Here...

cryptonews.ru08/14 08:25

Making a Fortune of $10.32 Million: The World Cup Money-Printing Tactic of a Polymarket Whale

**Earning $10.32 Million: A Polymarket Whale's World Cup Profit Strategy** While teams battle for the World Cup trophy, a hidden whale nicknamed "swisstony" has been quietly making a fortune on the prediction market Polymarket. This account, created around July 2025, boasts total profits of $18.62 million, with $10.33 million earned in the past month alone. With a 52.9% win rate, it has placed over 139,600 predictions, averaging about 380 trades per day—indicating it is likely a high-frequency quantitative bot. The account's signature "trash panda" aptly describes its strategy: sifting through vast market data and tiny price discrepancies to build wealth. Its current holdings are heavily concentrated on the France vs. Spain semi-final, including a roughly $160,000 bet against France. Analysis shows two core tactics driving its success: 1. **High-Volume "Anti-Favorite" Bets:** Placing large wagers (often $400k-$1M) against overvalued favorites like Germany or England, buying "No" shares at favorable prices when market-implied win probability is 46%-64%. It has recorded over 17 individual profits exceeding $1 million using this method. 2. **"Lottery-Ticket" Bets on Extreme Long Shots:** Allocating small amounts (thousands of dollars) to buy shares priced as low as 0.2¢-1.2¢ on outcomes deemed nearly impossible. While most of these bets lose, the occasional win—like those with payouts over 100x—generates significant profits (over $100,000 per hit) that boost overall returns without risking much capital. This dual approach combines consistent, large-scale profit from correcting major market mispricings with opportunistic, high-reward bets on extreme underdogs. The account exemplifies how systematic, high-frequency execution can amplify a small statistical edge into millions in profits on prediction markets.

Foresight News07/13 04:03

Making a Fortune of $10.32 Million: The World Cup Money-Printing Tactic of a Polymarket Whale

Foresight News07/13 04:03

What Kind of People Is Anthropic Actually Hiring? Insights from 1,680 Resumes Provide the Answer

Contrary to the common perception of Anthropic as a PhD-heavy AI research lab, an analysis of 1,680 engineer profiles reveals its core talent is experienced "builders." The engineering team, largely assembled in the past 18 months (median tenure 10 months), hires senior engineers with a median of 12.2 years of prior experience. They primarily come from major engineering-focused companies like Google, Meta, Amazon, Microsoft, Stripe, Databricks, Snowflake, and Palantir, not primarily from other AI labs. Their backgrounds are infrastructure-heavy (40%), focusing on backend, distributed systems, databases, and security, with skills like Python, Java, C++, and AWS being common. Only 13.7% hold PhDs. The dominant job title is "Member of Technical Staff." For early-career hires (<6 years experience), entry is not based on standard work history but on prestigious signals: top internships (e.g., at FAANG or quant trading firms), competition rankings, research publications, or AI safety fellowship experience. The key insight for applicants is to present as infrastructure builders, highlighting experience with large-scale systems, not as pure researchers, reflecting Anthropic's identity as a highly engineered infrastructure company.

marsbit06/15 03:28

What Kind of People Is Anthropic Actually Hiring? Insights from 1,680 Resumes Provide the Answer

marsbit06/15 03:28

Someone Predicts South Korean Stock Market with Hyperliquid, Achieving 74% Accuracy?

A study analyzed whether weekend price movements of four Korean stock perpetual futures contracts (Samsung Electronics, SK Hynix, Hyundai Motor, and EWY) on Hyperliquid could predict their Monday opening directions on their respective primary exchanges (KRX, NYSE). Over 62 weekend observations across the four assets, Hyperliquid correctly predicted the Monday opening direction 45 times (73.8% accuracy). However, performance varied significantly. Samsung Electronics showed the strongest and statistically significant signal, with Hyperliquid's weekend close correctly predicting its KRX Monday open in 15 out of 16 cases (94% accuracy, p-value < 0.001). This signal remained strong (75% accuracy) even when using Saturday's close instead of Sunday's, suggesting genuine price discovery beyond last-minute convergence. Hyundai Motor also showed high accuracy (81%, 13/16 correct), but this was not statistically significant after accounting for a baseline downward bias in its Monday opens. SK Hynix performed marginally better than a coin flip (63%, 10/16). EWY performed the worst (54%, 7/13), underperforming a simple strategy of always predicting a Monday rise. The stark difference between Samsung and EWY is largely attributed to market timing. KRX opens shortly after Hyperliquid's Sunday close, while NYSE opens ~14 hours later, allowing new information to flow in. The results suggest that for assets like Samsung Electronics, where weekend trading on Hyperliquid precedes the primary market open by only minutes, the platform can provide a valuable predictive signal worth monitoring before the Monday auction, despite the currently small weekend trading volumes.

Foresight News06/10 10:08

Someone Predicts South Korean Stock Market with Hyperliquid, Achieving 74% Accuracy?

Foresight News06/10 10:08

Tokens Not Selling? 90% of Crypto Projects Overlook Investor Relations

The article argues that effective Investor Relations (IR) is a critical yet often neglected function for crypto projects, with 90% failing at it and struggling to sell their tokens. Good IR acts as a bridge between a project and the market, broadening the buyer base and improving holder quality. The core of a successful IR strategy is distribution: maximizing the number of target investors who know about the token and converting them into buyers. The two primary buyer types are active crypto funds (requiring clear narratives and data for value reassessment) and large strategic institutions (requiring a long B2B sales cycle). The author emphasizes the necessity of proactively controlling the project's narrative with honesty and context, rather than remaining silent. A major tactical error is poor planning for token unlocks; teams should start 30-50 weeks in advance to manage supply and demand. Data is presented as the best ally for building a compelling story, providing context and comparisons for investors. The author contends that crypto IR should not be a dry, compliance-driven task but an engaging, interactive process similar to modern marketing. To lower the barrier to entry, projects must provide ample public data and research, making it easier for funds to conduct due diligence. Furthermore, the article highlights the power of on-chain data for deep investor analysis and argues that greater transparency, not less, actually expands the market by reducing uncertainty. Success should be measured by improvements in investor base quality and breadth—such as growth in target investors and holder diversification—rather than just token price. The future of IR is envisioned as dynamic, multimedia-rich, and proactive, leveraging the inherent transparency of crypto to build a larger, more engaged investor community.

marsbit03/17 13:39

Tokens Not Selling? 90% of Crypto Projects Overlook Investor Relations

marsbit03/17 13:39

How to Systematically Track High-Win-Rate Addresses on Polymarket?

This article explores methods to systematically identify and track high-success-rate addresses on Polymarket, a blockchain-based prediction market where all transactions are publicly recorded on-chain. It highlights that while data is transparent, the key challenge lies in extracting meaningful signals from vast datasets to detect addresses with potential informational advantages. The piece outlines common characteristics of such addresses: new wallets making large, concentrated bets; specialization in specific verticals; abnormal changes in position size; and exceptionally precise timing, repeatedly entering positions hours before major news breaks. A three-step systematic approach is recommended: First, filter addresses based on sustained profitability (e.g., 30-day positive returns, >55% win rate) using leaderboards like Polymarket Analytics. Second, analyze their holdings in specific event markets, focusing on addresses that are consistently among the top holders before full market pricing. Third, scrutinize their on-chain behavior: entry timing relative to news, position-building patterns (e.g., rapid, concentrated entries), holding periods, and trading focus. Advanced strategies include monitoring exit behavior (e.g., large, unexplained sell-offs), conducting wallet clustering analysis to find linked addresses, tracking unusual volume spikes in low-liquidity markets, and cross-referencing on-chain activity with external real-world data for validation. The goal is to move beyond luck and identify addresses exhibiting repeatable, information-driven advantages.

marsbit03/02 11:35

How to Systematically Track High-Win-Rate Addresses on Polymarket?

marsbit03/02 11:35

Dissecting 290,000 Data Points: We Uncovered 6 Secrets of Polymarket's Liquidity

Based on an analysis of 295,000 markets on Polymarket, this investigation uncovers six key truths about its liquidity. A significant finding is that 67.7% of markets have a lifespan of less than 7 days, with 63.16% of current short-term markets having zero trading volume, resembling the high failure rate of meme coins. These short-term markets, dominated by sports and crypto predictions, suffer from extremely low liquidity, often under $100. In contrast, long-term markets (over 30 days), though fewer in number, attract substantial capital, with an average liquidity of $450,000. U.S. politics is the most capitalized category. The platform exhibits a stark divide: sports markets are either ultra-short-term with high volume or long-term "season bets," with mid-term interest lacking. New, complex markets like U.S. real estate face a "cold start" problem due to high expertise requirements and low volatility, deterring participation. The market is highly polarized; a tiny fraction of high-value contracts (1,000+ with over $10M volume) capture 47% of all trading volume, while the vast majority of markets are illiquid. Finally, the "Geopolitics" category is the fastest-growing, indicating rising user interest. The core insight is that liquidity in prediction markets is not evenly distributed but concentrates around events that offer either instant gratification (sports/crypto) or deep macro bets (politics), transforming Polymarket into a specialized financial tool rather than a universal prediction platform.

比推01/08 08:17

Dissecting 290,000 Data Points: We Uncovered 6 Secrets of Polymarket's Liquidity