PeckShield: в июле ущерб от взломов составил $142 млн

cryptonews.ruPubblicato 2025-08-13Pubblicato ultima volta 2025-08-13

По итогам июля криптоиндустрия потеряла $142 млн в результате 17 хакерских атак. Это на 27% больше, чем в июне — тогда показатель составил $111 млн, сообщает PeckShield.

#PeckShieldAlert In July 2025, ~17 major crypto hacks were recorded, resulting in total losses of $142M—a 27.2% increase (from $111.6M in June). Notably, the #GMX exploiter has returned ~$40.5M worth of cryptos, including 10K ETH and 10.5M $FRAX.#Top5 Hacks in July 2025:… pic.twitter.com/Y5VLUILq5Z
— PeckShieldAlert (@PeckShieldAlert) August 1, 2025

Топ взломов

Крупнейшим инцидентом прошлого месяца был взлом индийской биржи CoinDCX, которая потеряла $44 млн. CEO Cyvers Дедди Лавида считает, что за атакой стоит северокорейская группировка Lazarus Group.

Торговая площадка пообещала полностью компенсировать убытки за счет финансовых резервов.

По данным следствия, хакеры получили доступ к кошелькам компании через ноутбук сотрудника. Его уже арестовала полиция Индии, которая продолжает расследовать инцидент.

На втором месте по сумме потерянных средств — взлом GMX. 9 июля децентрализованная криптобиржа лишилась $42 млн. Однако спустя два дня злоумышленник вернул почти все украденные активы.

16 июля взлому подверглась торговая площадка BigONE. Хакеры скомпрометировали продуктовую сеть и изменили логику работы серверов, которые отвечают за учетные записи и контроль рисков. Платформа пообещала компенсировать все убытки за свой счет.

В июле также пострадали WOO X и Future Protocol, лишившиеся $12 млн и $4,2 млн соответственно.

Напомним, за первые шесть месяцев 2025 года убытки Web3-проектов от взломов и мошенничества составили $3,1 млрд. Сумма уже превысила показатель за весь 2024 год — $2,85 млрд.

«Атаки стали сложнее, атаки стали более продуманными»

Letture associate

OpenAI Researcher Exposes ASI Timeline: Most Have Become Reality

In April 2025, a group of former OpenAI researchers published a 71-page document titled "AI 2027," outlining a timeline for Artificial Superintelligence (ASI). Their predictions, now being tracked by an independent project, show 51% are already confirmed, ahead of schedule, or on track. Notably, alarming predictions are arriving faster than anticipated. The forecast that AI would achieve top-tier human-level capabilities in cyber offense and defense by early 2027 was realized in April 2026, nine months early. Similarly, major Pentagon contracts with leading AI labs were signed 18 months earlier than predicted. The core mechanism for an intelligence explosion—Recursive Self-Improvement (RSI), where AI accelerates its own development—has not yet closed its loop. While AI, like Anthropic's Claude, now writes most new code, the bottleneck has shifted to human review and high-level research direction. A July 2026 study indicates the current AI-driven productivity gain in R&D is about 9%, below the estimated 15% threshold needed for a self-sustaining RSI feedback loop. However, underlying capabilities continue to accelerate rapidly. The "time horizon" metric for AI to autonomously handle tasks is doubling every three months, suggesting monthly-scale autonomous operation could be feasible by early 2027. Consequently, the original authors have revised their median prediction for fully automated AI programming forward to around mid-2028.

marsbit5 min fa

OpenAI Researcher Exposes ASI Timeline: Most Have Become Reality

marsbit5 min fa

Revealing the Flow of $11.2 Billion in Funding Over Six Months: The Most Valuable Asset in Crypto Is Shifting from Code to Licenses

The article reveals that in the first half of 2026, all $11.2 billion in disclosed crypto industry funding flowed exclusively to businesses requiring regulatory licenses to operate, signaling a major shift in institutional valuation. Top-funded sectors were payments/stablecoins ($3.7B), prediction markets ($2B), and exchanges/trading platforms ($1.7B). The core argument is that capital's focus has moved from technological innovation ("what code can do") to compliance capability ("possession of a license"), as licenses represent a slow, costly barrier to entry that code cannot easily replicate. This mirrors the evolution of traditional fintech. Notably, a split is emerging: while institutional capital floods licensed, centralized entities, retail user activity continues to thrive on permissionless DeFi protocols. The article concludes that within the institutional framework, the industry's most valuable asset is being redefined from code to regulatory approval.

marsbit6 min fa

Revealing the Flow of $11.2 Billion in Funding Over Six Months: The Most Valuable Asset in Crypto Is Shifting from Code to Licenses

marsbit6 min fa

Xia Zuoquan, on the Verge of Harvesting a Fifth IPO

China's stock market is poised to welcome another star semiconductor company. High-speed interconnect chipmaker Taoran Semiconductor has filed for an IPO on the ChiNext board in Beijing. Founded in 2015 by Guan Yi, a veteran chip designer with experience at Broadcom, the company focuses on high-performance analog and mixed-signal chips for 5G base stations, AI data centers, and new energy vehicles. Its product portfolio aims to fill gaps in domestic supply chains. Since 2018, Taoran has completed 12 funding rounds, raising over 1 billion RMB. Investors include prominent venture capital firms like Gaorong Ventures, Shenzhen Capital Group, and Qiming Venture Partners, as well as strategic investor BYD and various state-backed funds from Shanghai, Beijing, and Fujian. A notable early investor is Xia Zuoquan, co-founder and early angel investor in BYD, through his firm Zhengxuan Capital. Zhengxuan Capital is now the second-largest institutional shareholder with a 6.7% stake, and Xia holds an additional 5.3% personally. If Taoran successfully lists, this will mark Xia's fifth IPO exit, following BYD, UBTech, Yutai Micro, and Shangshui Intelligent. The company reports its revenue grew at a compound annual rate of over 80% from 2022 to 2024 and has become a supplier to top-three domestic communication equipment manufacturers.

marsbit8 min fa

Xia Zuoquan, on the Verge of Harvesting a Fifth IPO

marsbit8 min fa

A PPT Dismissed as 'Nonsense' by MIT Professors 5 Years Ago Predicted the Core Ideas of OpenAI o1 and o3

In 2020, AI researcher Giambattista Parascandolo presented his vision for neural network reasoning at an MIT faculty interview, only to have the committee dismiss the direction as "nonsense." He later posted the details online. His talk centered on enabling artificial neural networks to generalize and plan beyond their training data, closer to human capabilities. Parascandolo proposed three key research directions. First was "open-ended reasoning," where models could dedicate more computation time to harder problems, continuously refining answers—a precursor to today's compute-adaptive reasoning models. He noted that simply adding steps (e.g., in RNNs) wasn't enough without learning to use them effectively. Second, he advocated using language as a medium for reasoning within reinforcement learning. By leveraging the world knowledge in models like GPT, agents could better describe environments, decompose tasks, and plan—foreshadowing concepts like chain-of-thought and agent workflows. His third direction involved giving AI systems the ability to manipulate their own learning process: resetting to past states, creating counterfactual scenarios, and even editing their own activations or weights to facilitate deliberate practice. Parascandolo, who earned his PhD focusing on out-of-distribution generalization and had internships at Google X and DeepMind, joined OpenAI in 2021. He contributed to GPT-4 and later became integral to the foundational research behind the reasoning models o1 and o3. His early, criticized ideas remarkably charted a course for advanced AI reasoning systems developed years later.

marsbit10 min fa

A PPT Dismissed as 'Nonsense' by MIT Professors 5 Years Ago Predicted the Core Ideas of OpenAI o1 and o3

marsbit10 min fa

US Debt Approaching $40 Trillion, BofA's Hartnett: Going Long on Gold Is the Optimal Solution Now

U.S. national debt is on the verge of reaching $40 trillion. Bank of America's Chief Investment Strategist Michael Hartnett, in his latest "Flow Show" report, identifies this as the core market narrative and argues that **going long on gold is the optimal solution** currently. He views gold as the best hedge against dollar depreciation, bond market collapse, and asset inflation. Key pressures on the bond market include soaring debt interest payments, nearing $1.4 trillion annually and set to become the federal government's largest expenditure. Simultaneously, a surge in corporate bond issuance, particularly for AI/data center projects (up roughly 12 times historical averages), is structurally steepening the yield curve and crowding out buyers for long-term Treasuries. Hartnett reiterates his asset allocation framework for the 2020s: **ABB (Away From Bonds), ABD (Away From Dollar), and AI (All In AI)**. Within the "ABD" theme, he explicitly recommends going long gold. He suggests policymakers have limited tolerance for significantly higher yields, especially with elections approaching. Under the "ABB" theme, he notes that some previously neglected long-duration assets like REITs, biotech (XBI), regional banks (KRE), and small-cap stocks are quietly outperforming as the market prices in a peak in yields. Conversely, within the "AI" theme, he proposes a counterintuitive trade: **shorting AI-related bonds**, given the sector's massive capital expenditure needs and negative free cash flow. Hartnett outlines key upcoming events (Jackson Hole, CPI data, Fed & BOJ meetings, U.S. elections) that will influence markets. His final outlook is politically contingent: a Republican-held Senate could fuel a further AI-driven market bubble into 2027, while a Democratic victory in November could trigger a >10% decline in stocks, the dollar, and bond yields by year-end.

marsbit13 min fa

US Debt Approaching $40 Trillion, BofA's Hartnett: Going Long on Gold Is the Optimal Solution Now

marsbit13 min fa

Trading

Spot