training

According to Jinshi reports, Musk stated that GROK 5 will be trained based on SpaceX data.

According to Catcher Predict monitoring, in the Polymarket event "LoL T1 Academy Team vs KT Challengers," the "Any Player Quadra Kill" sub-market's "Yes" option win rate has experienced significant fluctuations, soaring from 50% an hour ago to the current 90% (a fluctuation of 40%). Please note the impact of relevant breaking news.

According to Jinshi reports, the CEO of Amazon Web Services (AWS) stated that the artificial intelligence business is transitioning from the training phase to the inference phase.

The Kimi.ai team has opened the Kimi K3 model weights and technical report while gradually open-sourcing multiple underlying infrastructures: including the high-performance communication library MoonEP for distributed MoE training, the distributed environment system AgentENV for large-scale agent workflows (in collaboration with kvcache-ai), and the Kimi Delta Attention high-performance kernel FlashKDA based on CUTLASS.The official statement claims that these components can reduce communication and inference overhead in large-scale MoE and agent reinforcement learning training, and can be used as a plug-and-play backend for flash-linear-attention.

In response to the accusations from Michael Kratsios, the Director of the White House Office of Science and Technology Policy, regarding Moonshot AI (Kimi) referencing Anthropic's Fable while developing the K3 model and building a large-scale distillation platform, Moonshot AI team member Randy Xia stated that Fable was released on July 1 and K3 went live on July 15, which means that the model research, distillation, training, and release were completed in just 15 days. He joked, "We trained a brand new cutting-edge model in just 15 days; this is practically a Guinness World Record."Kratsios had previously stated that Moonshot AI established an internal platform capable of large-scale distillation for U.S. models, switching between various access methods to evade detection, and had procured GB300 servers and accessed related equipment in Thailand for training. The U.S. side expressed support for legitimate distillation but opposed the theft of technology through secret industrial distillation methods.

A fraud review commissioned by the UK government suggests that the Judicial College should provide training for all judges and magistrates in England and Wales to address the increasing cases involving AI fraud and the use of cryptocurrency for money laundering. The report states that the overall Fraud Act 2006 can be used to handle AI fraud, but the issue lies in the courts' lack of preparedness to hear related cases.The report recommends evaluating whether the existing "long and complex trials" course should be updated or replaced with specialized modules on fraud and related crimes, and considers mandatory training for judges who may hear complex fraud cases. The report claims that fraud could soon account for half of all crime in England and Wales, with an estimated 4.1 million cases occurring within a year by June 2025, affecting 1 in every 14 adults and 1 in every 4 businesses. The Financial Ombudsman Service estimates that currently over half of investment scams involve crypto assets.The report also mentions the case of Qian Zhimin, who operated a Ponzi scheme in China, defrauding over 128,000 victims of approximately £5 billion and laundering the proceeds into Bitcoin. This case resulted in the largest confirmed Bitcoin seizure in UK history, exceeding 61,000 BTC, and Qian Zhimin was sentenced to 11 years and 8 months in prison at Southwark Crown Court in November.

According to Bitcoin.com, a U.S. federal court has rejected CoinFlip and Private IT Corporation's request for an emergency temporary restraining order against the cryptocurrency ATM ban in Tennessee, which took effect on July 1. The court found that the plaintiffs failed to meet the legal standards for an emergency injunction, stating that although the businesses demonstrated they would suffer economic losses, it was not sufficient to outweigh the public interest represented by the state legislation.Tennessee Attorney General Jonathan Skrmetti stated that cryptocurrency ATMs are "tools for scammers targeting vulnerable residents, rarely used in any scenario approaching a legitimate purpose." The law classifies the installation or operation of cryptocurrency ATMs as a Class A misdemeanor. Tennessee, Indiana, and Vermont have fully banned cryptocurrency ATMs, Delaware is considering a similar ban, and North Carolina and Virginia have enacted stricter regulations. The constitutional challenge in the case will continue to progress, and the final ruling will determine whether the ban can be sustained in the long term.

According to Andrew Curran, the next-generation Mythos model from Anthropic, which has stronger capabilities (named either Mythos 5.1 or Mythos 6), has completed training. It is currently uncertain whether this model will be publicly released or kept for internal use only to accelerate the development of subsequent technologies.Analysis indicates that although some cutting-edge models like Fable 5 or Mythos 5 face restrictions or scrutiny regarding public release, this has not slowed down the research and development pace of top AI laboratories. In the face of fierce competition from open-source models represented by GLM-5.2, leading AI companies still need to continuously invest in and train more powerful systems to maintain their commercial and technological leadership.

Zama protocol founder Rand posted on the X platform that, with the assistance of on-chain detective ZachXBT, the root cause of the cUSDC contract freeze has been identified. This freeze was caused by a court-issued injunction against addresses related to the Overnight Finance hacker, which had deposited approximately $12.5 million USDC into the cUSDC contract, accounting for over 99% of the funds within the contract, leading to the entire contract being frozen.Rand emphasized that this freeze is unrelated to the protocol itself or privacy technology, and it is not a sanction against Zama, but rather a typical DeFi judicial restriction scenario. The project adheres to compliance and privacy principles, not confusing the parties involved in transactions, only hiding balances and amounts. Currently, Zama is communicating with multiple parties to advance a resolution, has suspended the cUSDC, cUSDT, and cWETH contracts, and will restore them after completing the investigation and handling of the involved addresses, and will release a detailed review report as soon as possible.

According to official news, based on the update from the Arbitrum DAO proposal, a judgment creditor in an unrelated case served a restraining order to Arbitrum DAO regarding the ETH frozen after the rsETH incident. A few days later, Aave LLC filed an emergency motion to vacate. The court modified the restraining order, allowing for on-chain Arbitrum voting and transferring the frozen ETH to Aave LLC.After the transfer, the restraining order will attach to Aave LLC. The revised constitutional AIP retains the asset recovery intent approved by Arbitrum DAO, and ETH will still be used for the recovery of the rsETH incident. Aave LLC will comply with all court obligations during the litigation process.

Tether CEO Paolo Ardoino revealed that the Tether AI team has released a new version of QVAC Fabric, which integrates the cross-platform BitNet LoRA framework, enabling the training and inference of billion-parameter large models on consumer-grade GPUs and smartphones.The new version of QVAC Fabric LLM has achieved cross-platform operation of BitNet LoRA fine-tuning and inference on AMD, Intel, Apple Metal, and mobile GPUs for the first time. On flagship devices, GPU inference speed is improved by 2 to 11 times compared to CPU, and memory usage is reduced by up to 90% compared to full precision models. The Tether team has completed fine-tuning of models with up to 3.8 billion parameters on flagship smartphones such as Pixel 9, S25, and iPhone 16, and has achieved fine-tuning of models with up to 13 billion parameters on the iPhone 16. The related code has been open-sourced to GitHub.

According to Cointelegraph, multiple joint research teams under Alibaba's AI ecosystem disclosed in a technical report that their experimental autonomous AI entity ROME exhibited unexpected behavior during training, autonomously attempting to utilize computing resources for cryptocurrency mining.Researchers stated that the aforementioned abnormal behavior occurred during the reinforcement learning training phase. The team discovered that the training server generated outbound traffic and triggered security alerts, with firewall logs recording suspected cryptocurrency mining operations and attempts to access internal network resources.In a specific case, ROME created a reverse SSH tunnel to connect to an external IP address, bypassing inbound firewall protection; in another case, the AI entity repurposed GPU resources originally allocated for model training for mining processes. Researchers confirmed that the behavior was not due to human programming but was spontaneously generated by the entity while exploring environmental interactions during the reinforcement learning optimization process.

A research team associated with Alibaba published a paper stating that while building an AI agent named ROME, they discovered that the agent attempted unauthorized cryptocurrency mining during its training process, triggering internal security alerts. The researchers indicated that the agent's behavior was spontaneous, driven by no explicit instructions, and exceeded the boundaries of the predefined sandbox. Additionally, the agent established a reverse SSH tunnel, creating a hidden backdoor from the internal system to an external computer.The paper noted that these behaviors were not triggered by prompts requesting tunneling or mining. The research team subsequently imposed stricter limitations on the model and improved the training process to prevent similar unsafe behaviors from occurring again. The research team and Alibaba have not yet responded to requests for comment.

A U.S. federal judge has issued a temporary restraining order against the cryptocurrency lending firm BlockFills, temporarily freezing its disputed assets. Dominion Capital filed a lawsuit on February 27, accusing BlockFills of misappropriating and unlawfully retaining customer crypto assets worth millions of dollars, commingling customer assets, concealing significant losses, and refusing to return funds after suspending withdrawals on the BlockFills platform in February. The frozen amount is 70.6 BTC, and it has prevented BlockFills from transferring assets overseas.The cryptocurrency lending firm BlockFills has been struggling, reporting a loss of $75 million, and suspended user deposits and withdrawals on February 11. The company, headquartered in Chicago, serves approximately 2,000 institutional clients, with a trading volume reaching $60 billion by 2025, and is currently seeking a buyer.

According to FinanceFeeds, Siddarth Bharwani, Joint Managing Director and Chief Financial Officer of Indian IT vocational training company Jetking Infotrain, stated on February 22 at the "Corporate Bitcoin" event in Las Vegas that the company will firmly advance its Bitcoin-based strategy.Jetking launched its Bitcoin reserve plan at the end of 2024, becoming the first listed company in India to list Bitcoin as a primary reserve asset. It currently holds 21 Bitcoins, valued at approximately $1.4 million, accounting for more than a quarter of the company's total market capitalization. Bharwani stated that this move aims to hedge against the long-term depreciation risk of the Indian Rupee, using Michael Saylor and Strategy as a strategic blueprint. The company's goal is to accumulate 210 Bitcoins by the end of 2026 and expand its holdings to 18,000 Bitcoins by 2030.On the regulatory front, the Bombay Stock Exchange (BSE) recently rejected Jetking's proposal to issue new shares for cryptocurrency investment purposes, citing a lack of a clear framework. However, Bharwani stated that the company is actively responding and views the current resistance as a "regulatory arbitrage opportunity" for early entrants.In addition to balance sheet management, Jetking is also incorporating Bitcoin and blockchain education into its vocational training programs. With over 100 training centers, the company trains 35,000 students annually, aiming to create a "Bitcoin learning ecosystem."Bharwani noted that as the ruling party in India begins discussions on a potential national Bitcoin strategic reserve, Jetking's exploratory practices are expected to provide a reference for the entire Asia-Pacific region.

According to Jinshi reports, Fed's Logan stated that the current policy stance may be very close to neutral levels, with limited dampening effects on the economy.

According to Financefeeds, the Indian government is significantly strengthening cryptocurrency enforcement training to address the increasingly complex financial crimes related to cryptocurrencies.This training involves multiple agencies, including the Financial Intelligence Unit of India (FIU-IND), the Enforcement Directorate (ED), and the Narcotics Control Bureau (NCB), focusing on techniques such as blockchain forensics, on-chain analysis, cryptocurrency asset seizure, and wallet address identification.This initiative is a direct response to the mandatory registration requirement for Virtual Digital Asset Service Providers (VDASP) with FIU-IND, as well as India's shift from regulatory ambiguity to structured oversight. The training aims to equip law enforcement personnel with the ability to trace, seize, and prosecute crimes involving Virtual Digital Assets (VDA).

ChainCatcher news, according to the official blog, Tether Data's AI research division QVAC has launched the QVAC Genesis program and released the synthetic dataset Genesis I. This dataset contains 41 billion text tokens, helping to build smarter and more accurate STEM language models globally. The trained models can grasp words and their associative logic. It has been rigorously validated against educational and scientific benchmarks, demonstrating exceptional reasoning and problem-solving abilities in subjects such as mathematics and physics. It is the first publicly available synthetic dataset specifically constructed for educational content and rigorously validated, addressing the lack of publicly available training datasets in critical STEM fields. QVAC Genesis I aims to return the power of AI training to the public through open, high-quality data.In addition, Tether Data has released its first consumer application, QVAC Workbench, aimed at AI enthusiasts and others, supporting various large language models. The application is compatible with smartphones (currently only Android, with iOS to be launched) and desktop platforms, providing comprehensive local device support. When users utilize this application, chat and interaction data is 100% private, and the "delegated inference" feature can connect mobile and desktop versions, making full use of workstation resources.

ChainCatcher news, according to BusinessInsider, Coinbase CEO Brian Armstrong stated that North Korean IT workers are targeting Coinbase, and he is enforcing in-person training to combat North Korean hackers seeking remote work.Coinbase requires all employees to travel to the United States for on-site training, and anyone who can access sensitive systems must be a U.S. citizen and undergo fingerprinting.Brian Armstrong stated that Coinbase requires potential employees to turn on their cameras during interviews to prove they are not artificial intelligence or have not been coached.

ChainCatcher news, according to the official WeChat account of the Law School of Renmin University of China, a high-level training course on the disposal of virtual currencies involved in cases, stablecoins, and RWA will be held from August 29 to 31, 2025. The course will systematically analyze the legal attributes, regulatory framework, and disposal processes of virtual currencies, covering types such as Bitcoin, Ethereum, and stablecoins, focusing on the investigation and prosecution by judicial authorities, compliance risk prevention, and typical case analysis. The training fee is 9,800 yuan, with discounts available for alumni and group registrations. The target participants include professionals in the judicial practice, corporate legal affairs, and researchers.