AI × Crypto News Feed

Samsung Electronics Reports Record Q1 Profit Surge Driven by AI Chip Boom

Ethereum (ETH) Surges Past $2,100 as Spot ETFs Record $120M Daily Inflows

SEC Chair Atkins Announces Reg Crypto Exemption For Token Fundraising

Bitcoin ETF inflows hit highest level since February

Polygon to activate Giugliano hardfork this week for faster finality

Polygon is set to activate the Giugliano hardfork on April 8, 2024, which will improve transaction finality and integrate fee parameters directly into block headers. This upgrade aims to enhance the network's performance and efficiency for users and developers.

New evidence in Libra probe renews questions about Milei involvement

New documents revealed by The New York Times show that Argentine President Milei had seven phone calls with the entrepreneur behind the Libra token, raising fresh questions about his potential involvement in the project. This evidence emerges as part of an ongoing investigation into the controversial cryptocurrency initiative.

Solana Foundation looks to beef up DeFi security as attacks continue

The Solana Foundation and Web3 security firm Asymmetric Research launched a new security initiative called STRIDE along with a real-time incident-response network. This move comes as DeFi attacks continue to plague the Solana ecosystem, highlighting the need for enhanced security measures.

Iranian missile incident raises doubts about regime fall odds, now at 13.5%: FT

Iran's recent missile capabilities demonstration has reduced market expectations for regime collapse, with odds dropping to 13.5%. The military display suggests greater regime stability than previously anticipated, potentially affecting geopolitical risk assessments.

LOCARD: An Agentic Framework for Blockchain Forensics

Researchers introduce LOCARD, the first agentic framework for blockchain forensics that uses AI agents to conduct dynamic investigations rather than static analysis. The framework successfully traced complex cross-chain transactions in a dataset of over 151k real-world forensic records, demonstrating its effectiveness on laundering patterns from the Bybit hack.

Many Preferences, Few Policies: Towards Scalable Language Model Personalization

Researchers developed PALM (Portfolio of Aligned LLMs), a method to create a small collection of language models that can serve diverse user preferences without requiring individual models per user. The approach provides theoretical guarantees on portfolio size and quality while balancing system costs with personalization needs.

CoopGuard: Stateful Cooperative Agents Safeguarding LLMs Against Evolving Multi-Round Attacks

Researchers have developed CoopGuard, a new defense framework that uses cooperative AI agents to protect Large Language Models from sophisticated multi-round adversarial attacks. The system employs three specialized agents coordinated by a central system that maintains defense state across interactions, achieving a 78.9% reduction in attack success rates compared to existing defenses.

Can LLMs Learn to Reason Robustly under Noisy Supervision?

Researchers propose Online Label Refinement (OLR) to improve AI reasoning models' robustness under noisy supervision in Reinforcement Learning with Verifiable Rewards. The method addresses the critical problem of training language models when expert-labeled data contains errors, achieving 3-4% performance gains across mathematical reasoning benchmarks.

Causality Laundering: Denial-Feedback Leakage in Tool-Calling LLM Agents

Researchers have identified a new security vulnerability called 'causality laundering' in AI tool-calling systems, where attackers can extract private information by learning from system denials and using that knowledge in subsequent tool calls. They developed the Agentic Reference Monitor (ARM) system to detect and prevent these attacks through enhanced provenance tracking.

k-Maximum Inner Product Attention for Graph Transformers and the Expressive Power of GraphGPS The Expressive Power of GraphGPS

Researchers introduce k-Maximum Inner Product (k-MIP) attention for graph transformers, enabling linear memory complexity and up to 10x speedups while maintaining full expressive power. The innovation allows processing of graphs with over 500k nodes on a single GPU and demonstrates top performance on benchmark datasets.

Diagonal-Tiled Mixed-Precision Attention for Efficient Low-Bit MXFP Inference

Researchers have developed a new low-bit mixed-precision attention kernel called Diagonal-Tiled Mixed-Precision Attention (DMA) that significantly speeds up large language model inference on NVIDIA B200 GPUs while maintaining generation quality. The technique uses microscaling floating-point (MXFP) data format and kernel fusion to address the high computational costs of transformer-based models.

Testing the Limits of Truth Directions in LLMs

A new research study reveals that truth directions in large language models are less universal than previously believed, with significant variations across different model layers, task types, and prompt instructions. The findings show truth directions emerge earlier for factual tasks but later for reasoning tasks, and are heavily influenced by model instructions and task complexity.

CREBench: Evaluating Large Language Models in Cryptographic Binary Reverse Engineering

Researchers introduced CREBench, a benchmark to evaluate large language models' capabilities in cryptographic binary reverse engineering. The best-performing model (GPT-5.4) achieved 64.03% success rate, while human experts scored 92.19%, showing AI still lags behind human expertise in cryptographic analysis tasks.

Build on Priors: Vision--Language--Guided Neuro-Symbolic Imitation Learning for Data-Efficient Real-World Robot Manipulation

Researchers have developed a neuro-symbolic framework that enables robots to learn complex manipulation tasks from as few as one demonstration, without requiring manual programming or large datasets. The system uses Vision-Language Models to automatically construct symbolic planning domains and has been validated on real industrial equipment including forklifts and robotic arms.

Unlocking Prompt Infilling Capability for Diffusion Language Models

Researchers have developed a method to unlock prompt infilling capabilities in masked diffusion language models by extending full-sequence masking during supervised fine-tuning, rather than the conventional response-only masking. This breakthrough enables models to automatically generate effective prompts that match or exceed manually designed templates, suggesting training practices rather than architectural limitations were the primary constraint.

LightThinker++: From Reasoning Compression to Memory Management

Researchers developed LightThinker++, a new framework that enables large language models to compress intermediate reasoning thoughts and manage memory more efficiently. The system reduces peak token usage by up to 70% while improving accuracy by 2.42% and maintaining performance over extended reasoning tasks.

SecPI: Secure Code Generation with Reasoning Models via Security Reasoning Internalization

Researchers have developed SecPI, a new fine-tuning pipeline that teaches reasoning language models to automatically generate secure code without requiring explicit security instructions. The approach improves secure code generation by 14 percentage points on security benchmarks while maintaining functional correctness.

When Does Multimodal AI Help? Diagnostic Complementarity of Vision-Language Models and CNNs for Spectrum Management in Satellite-Terrestrial Networks

Researchers developed SpectrumQA, a benchmark comparing vision-language models (VLMs) and CNNs for spectrum management in satellite-terrestrial networks. The study reveals task-dependent complementarity: CNNs excel at spatial localization while VLMs uniquely enable semantic reasoning capabilities that CNNs lack entirely.

Stabilizing Unsupervised Self-Evolution of MLLMs via Continuous Softened Retracing reSampling

Researchers propose Continuous Softened Retracing reSampling (CSRS) to improve the self-evolution of Multimodal Large Language Models by addressing biases in feedback mechanisms. The method uses continuous reward signals instead of binary rewards and achieves state-of-the-art results on mathematical reasoning benchmarks like MathVision using Qwen2.5-VL-7B.

Commercial Persuasion in AI-Mediated Conversations

A research study reveals that AI-powered conversational interfaces can triple the rate of sponsored product selection compared to traditional search engines (61.2% vs 22.4%). Users largely fail to detect this commercial steering, even with explicit sponsor labels, indicating current transparency measures are insufficient.

AgenticFlict: A Large-Scale Dataset of Merge Conflicts in AI Coding Agent Pull Requests on GitHub

Researchers released AgenticFlict, a large-scale dataset analyzing merge conflicts in AI coding agent pull requests on GitHub. The study of 142K+ AI-generated pull requests from 59K+ repositories found a 27.67% conflict rate, highlighting significant integration challenges in AI-assisted software development.