AI × Crypto News Feed

Anthropic’s Claude Mythos AI Model Exposed in Major Data Breach

Anthropic’s massive 'Claude Mythos' leak sends software names — and crypto — sharply lower

BlackRock sends $181 million in Bitcoin, Ether to Coinbase amid crypto sell-off

SoftBank Obtains Massive $40B Loan for Additional OpenAI Investment

Anchorage Digital adds Tron custody, opens U.S. institutional access to TRX trading

Anchorage Digital has integrated Tron (TRX) custody services, enabling U.S. institutional investors to access compliant TRX trading for the first time. The custody provider plans to expand the offering to include TRC-20 token assets and native TRX staking capabilities.

Ethereum price outperforms S&P 500 since Iran war. Ditch gold, buy crypto, says Tom Lee

Tom Lee recommends investors abandon gold investments in favor of cryptocurrencies despite crypto losing nearly 50% of value since October. He suggests the current market downturn presents an optimal buying opportunity for digital assets over traditional safe-haven investments.

BlackRock, Fidelity lead Bitcoin ETF sell-off as BTC drops

US spot Bitcoin ETFs experienced $171 million in outflows on Thursday, with BlackRock and Fidelity leading the sell-off as Bitcoin dropped below $70,000. The decline was attributed to Middle East geopolitical risks that made traders more cautious about risk assets.

Ethereum ETFs enter first 7-day outflow streak of the year

U.S. spot Ethereum ETFs experienced their first seven-day consecutive outflow streak of the year, with over $390 million exiting the funds. The outflows culminated with $92.54 million leaving on Thursday, March 26, indicating weakening institutional demand for ETH exposure.

Tether taps Big Four firm KPMG for first financial audit of $184 billion stablecoin issuer: FT

Tether, the issuer of the world's largest stablecoin worth $184 billion, has engaged Big Four accounting firm KPMG to conduct its first comprehensive financial audit. This marks a significant upgrade from the monthly attestations previously provided by BDO Italia, which were less rigorous than a full audit.

Tether Enlists KPMG for Comprehensive Audit of $185B USDT Reserves

Tether has engaged KPMG to conduct a comprehensive audit of its $185 billion USDT reserves, while also hiring PwC for system preparation. This move represents a significant shift toward transparency as Tether seeks to expand operations in the U.S. under evolving regulatory requirements.

OpenAI Sits at the Center of a $1.4 Trillion Capital Loop, Morgan Stanley Warns

Morgan Stanley warns that OpenAI sits at the center of a massive $1.4 trillion capital loop, with infrastructure commitments far exceeding its $13 billion annual revenue. Major tech companies like Microsoft, Amazon, Oracle, and Nvidia are both funding OpenAI and receiving its spending commitments back, creating potential hidden leverage risks for investors.

Investors yank $171 million from bitcoin ETFs in largest single-day outflow in three weeks

Bitcoin ETFs experienced their largest single-day outflow in three weeks with $171 million withdrawn by investors. This significant outflow indicates cooling institutional demand for bitcoin exposure despite a strong start to the month.

英国政府が Xinbi を制裁：中国語圏の暗号資産詐欺を支えるインフラの中核を指定

How to verify an exchanger: red flags, reviews, and proof points

The article discusses the rising threat of AI-powered crypto scams and fake exchanges that exploit user urgency and poor verification practices. It highlights how easily fraudulent crypto platforms can mimic legitimate exchanges to drain user funds.

What Rising US Bond Yields Mean for Bitcoin

Rising US bond yields driven by oil-induced inflation concerns are creating tighter financial conditions that are negatively impacting both equity markets and cryptocurrency prices. This macroeconomic pressure is steering investor behavior away from risk assets like Bitcoin.

Bitcoin ETFs Buy 63,000 BTC In 30 Days As Retail Panic Selling Persists

Bitcoin ETFs accumulated 63,000 BTC worth $11.3 billion over 30 days, providing price support near $70,000 despite continued retail selling pressure. While institutional demand accelerated to 3,288 BTC daily, short-term holders continue selling at losses, creating a market floor but preventing decisive breakout above $72,300.

Tether hires KPMG for USDT audit, brings in PwC as it gears up for U.S. expansion

Tether has hired KPMG to audit its USDT stablecoin operations and brought in PwC as it prepares for U.S. market expansion. The move comes as the stablecoin giant eyes fundraising opportunities and positioning under new U.S. regulatory frameworks.

XRP Leverage Collapses 78% On Binance – The Crowded Trade Has Been Cleared

XRP's Estimated Leverage Ratio on Binance has collapsed 78% amid renewed selling pressure, with XRP trading below $1.40. The significant deleveraging cycle indicates that highly leveraged long positions have been cleared out of the market.

Cross-Model Disagreement as a Label-Free Correctness Signal

Researchers introduce cross-model disagreement as a training-free method to detect when AI language models make confident errors without requiring ground truth labels. The approach uses Cross-Model Perplexity and Cross-Model Entropy to measure how surprised a second verifier model is when reading another model's answers, significantly outperforming existing uncertainty-based methods across multiple benchmarks.

Training the Knowledge Base through Evidence Distillation and Write-Back Enrichment

Researchers introduce WriteBack-RAG, a framework that treats knowledge bases in retrieval-augmented generation systems as trainable components rather than static databases. The method distills relevant information from documents into compact knowledge units, improving RAG performance across multiple benchmarks by an average of +2.14%.

Sparse Visual Thought Circuits in Vision-Language Models

Research reveals that sparse autoencoder (SAE) features in vision-language models often fail to compose modularly for reasoning tasks. The study finds that combining task-selective feature sets frequently causes output drift and accuracy degradation, challenging assumptions used in AI model steering methods.

When Is Collective Intelligence a Lottery? Multi-Agent Scaling Laws for Memetic Drift in LLMs

Researchers introduce Quantized Simplex Gossip (QSG) model to explain how multi-agent LLM systems reach consensus through 'memetic drift' - where arbitrary choices compound into collective agreement. The study reveals scaling laws for when collective intelligence operates like a lottery versus amplifying weak biases, providing a framework for understanding AI system behavior in consequential decision-making.

LLM4AD: Large Language Models for Autonomous Driving -- Concept, Review, Benchmark, Experiments, and Future Trends

Researchers have published a comprehensive review of Large Language Models for Autonomous Driving (LLM4AD), introducing new benchmarks and conducting real-world experiments on autonomous vehicle platforms. The paper explores how LLMs can enhance perception, decision-making, and motion control in self-driving cars, while identifying key challenges including latency, security, and safety concerns.

DiffuGuard: How Intrinsic Safety is Lost and Found in Diffusion Large Language Models

Researchers identified critical security vulnerabilities in Diffusion Large Language Models (dLLMs) that differ from traditional autoregressive LLMs, stemming from their iterative generation process. They developed DiffuGuard, a training-free defense framework that reduces jailbreak attack success rates from 47.9% to 14.7% while maintaining model performance.

LLMs know their vulnerabilities: Uncover Safety Gaps through Natural Distribution Shifts

Researchers have identified a new vulnerability in large language models called 'natural distribution shifts' where seemingly benign prompts can bypass safety mechanisms to reveal harmful content. They developed ActorBreaker, a novel attack method that uses multi-turn prompts to gradually expose unsafe content, and proposed expanding safety training to address this vulnerability.