Real-time AI-curated news from 87,658+ articles across 50+ sources. Sentiment analysis, importance scoring, and key takeaways — updated every 15 minutes.
Avalanche Treasury (AVAT) debuted on Nasdaq following a $675 million SPAC merger but experienced a sharp 38.1% decline on its first trading day, closing at $1.85. Despite the stock's poor performance, the underlying AVAX token gained 1.27% to $6.66, and the company plans to acquire over $1 billion in AVAX tokens.
OpenAI has established a Madrid office as part of its European expansion strategy for ChatGPT and AI services. The move signals increased investment in Europe's AI ecosystem while simultaneously raising questions about centralization concerns within the cryptocurrency and decentralized technology communities.
Binance Wallet's SpaceX IPO subscription campaign accumulated $557M from 27,689 addresses, with whale wallets contributing the majority of funds. This demonstrates significant institutional and high-net-worth investor interest in tokenized equity offerings through decentralized platforms.
OpenAI has agreed to acquire Ona, a cloud infrastructure startup, to strengthen its Codex division's autonomous agent capabilities. Ona's team will integrate into OpenAI's operations as Codex reaches over five million weekly active users, continuing OpenAI's expansion strategy through targeted acquisitions.
Avalanche Treasury shares declined 38% following their Nasdaq debut after a $675 million merger, despite the company's commitment to acquire over $1 billion in AVAX tokens to support ecosystem growth. The sharp post-listing decline reflects typical post-merger volatility and investor skepticism regarding the valuation and execution of the announced acquisition strategy.
HashKey has approved a HK$100 million share buyback program, triggering a 10.51% stock price surge to HK$3.05. The buyback signals management confidence in the company's valuation amid recent market headwinds affecting the cryptocurrency and blockchain sector.
OpenAI has introduced a free banked reset system for Codex users, allowing greater flexibility in managing API rate limits. This feature aims to improve user satisfaction and strengthen Codex's competitive position in the AI coding assistance market.
SpaceX priced its IPO at $135 per share, valuing the company at approximately $1.77 trillion. Jim Cramer warned that excess retail demand could drive share prices significantly higher on debut, potentially creating unsustainable valuations and elevated first-day volatility.
Trader 0x9cc has established the largest SPCX long position ever recorded at $18M, reflecting heightened speculative activity around pre-IPO crypto assets. This move underscores the volatility and risk inherent in trading positions tied to companies approaching public markets.
Bitget has invested in enterprise-grade Claude AI access for its entire 2,167-person workforce at approximately $200 per employee. This move signals growing adoption of advanced AI tools within cryptocurrency and trading firms to enhance productivity and operational efficiency.
Avataar AI has launched a distilled video generation model priced at $0.005 per second, positioning itself as an affordable alternative to existing video AI solutions tailored for the Indian market. The pricing and cultural optimization strategy targets the scale and economic constraints of emerging markets while competing with higher-cost international players.
Avalanche Treasury Co. experienced a sharp 38.13% decline on its Nasdaq debut under ticker AVAT, while the underlying AVAX token fell 33.3% over the month, signaling investor skepticism toward treasury-focused products tied to struggling layer-1 blockchain performance.
South Korea's victory over Czech Republic in the World Cup opener coincides with increased cryptocurrency integration throughout the tournament. The growing presence of crypto partnerships and sponsorships in global sports events signals expanding mainstream adoption and changing dynamics in fan engagement and investment opportunities.
Bitcoin's Puell Multiple has declined to 0.74, indicating that miner revenues are sliding relative to historical averages. This metric suggests miners may be operating at reduced profitability, which could signal shifting market dynamics for the largest cryptocurrency.
Researchers propose a framework for strategic decision support in AI agent systems that balances minimizing human intervention with controlling the risk of agents acting without support when they should seek it. The approach uses threshold-based optimization and online algorithms to reduce unnecessary support calls while maintaining reliability, with applications across information gathering, human-AI collaboration, and tool use.
Pythagoras-Prover introduces a family of efficient Lean theorem provers that achieve state-of-the-art performance with significantly fewer parameters than existing models, using novel training techniques including curriculum learning and augmented data generation. The 4B-parameter model outperforms DeepSeek-Prover-V2-671B by 167x parameter efficiency, while the 32B model sets new benchmarks on formal mathematics tasks.
PersonaDrive introduces a retrieval-augmented vision-language-action (VLA) system that enables autonomous driving agents to exhibit diverse human-like behavioral styles in simulation environments. Using demonstrations from human drivers instructed to drive aggressively, neutrally, or conservatively, the system achieves superior performance on driving benchmarks while allowing style selection without per-style retraining.
Researchers introduce TrajGenAgent, an LLM-based framework that generates realistic synthetic human mobility trajectories without model fine-tuning by combining hierarchical agent design with deterministic workflows. The approach addresses privacy and cost constraints in trajectory data collection while maintaining semantic coherence and behavioral realism.
Researchers introduce Theory of Mind Utility (ToM-U), a formal computational framework for modeling how agents infer others' beliefs by tracking information access and credibility. The model uses directed graphs called Local Epistemic World Models to represent epistemic relationships and generates falsifiable predictions about mentalizing failures, advancing cognitive science theory beyond existing Bayesian and simulation-based approaches.
Researchers demonstrate that deep learning models for EEG analysis can be significantly compressed through parameter quantization and electrode reduction techniques, enabling deployment on resource-constrained wearable devices without substantial accuracy loss. This addresses a critical bottleneck in portable healthcare technology where computational demands of DNNs far exceed device capabilities.
Researchers present a framework for evaluating procedural reasoning datasets in AI-supported learning systems by comparing three question-generation strategies based on Task-Method-Knowledge (TMK) models. The study demonstrates that strict TMK generation produces the most grounded and usable datasets (96.5% grounded), while transcript-based approaches sacrifice representational alignment for naturalness, highlighting the trade-off between learner-like phrasing and formal grounding in evaluation dataset construction.
Researchers introduce MLUBench, a large-scale benchmark for evaluating lifelong unlearning in multimodal large language models (MLLMs), revealing that existing methods suffer from cumulative degradation. The study identifies a unique challenge in MLLM unlearning: removing data from one modality can damage the model's multimodal alignment, and proposes LUMoE as a solution to mitigate this degradation.
Researchers introduce Teach VLM, a vision-language model that extracts operational knowledge from mobile screen demonstrations to create interpretable instructions for GUI automation agents. The system uses a novel Teach-and-Repeat paradigm where extracted task procedures guide downstream execution agents, achieving state-of-the-art performance in operation semantics prediction and improving task success rates in Android environments.
Researchers introduced GeoNatureAgent Benchmark, the first evaluation framework for AI agents performing environmental geospatial analysis through real API interactions. Testing seven major LLMs across 93 tasks, Claude Sonnet 4 achieved 60.8% accuracy while DeepSeek V3.2 delivered 93% of Claude's capability at 11x lower cost, revealing significant performance gaps in structured reasoning tasks.
Researchers analyzing 80,814 papers from premier AI conferences (2017-2025) found that major AI topics advance through sudden phase transitions rather than gradual growth, with large language models and diffusion models surging dramatically within 1-3 years. The study identifies an early-warning signature that flags emerging topics—currently highlighting reasoning, agentic AI, multimodal LLMs, and world models as areas to monitor through 2028.
