South Korea Orders AI Security Reviews After Bank Data Breaches

Key Takeaways

• <ul><li>South Korea ordered nationwide security reviews after AI-linked cyberattacks hit KB Kookmin, Shinhan, Hana, and other banks, leaking data on tens of thousands of customers</li><li>ARTEX AI, an open-source LLM-based penetration-testing tool on GitHub, was traced to some of the attacks, but identifying perpetrators remains difficult</li><li>TSMC dominates advanced chip manufacturing with $143 billion in trailing revenue, a 60% operating margin, and a $265 billion U.S. factory investment plan</li><li>Nvidia's market value has grown roughly 13-fold to about $5.5 trillion, reviving Warren Buffett's warning about identifying durable competitive advantages</li><li>Nvidia CEO Jensen Huang criticized AI doomsday narratives and announced OpenShell and Sentry safety tools to prevent rogue AI agent breaches</li><li>A dual-agent trading framework using deep reinforcement learning improved risk-adjusted returns on stocks including Tesla, Nvidia, and Apple by adapting to market conditions</li><li>CRScube acquired Mednet to bring AI-assisted data entry and real-time coding to clinical trials, keeping proprietary data in a closed loop</li><li>Doximity released Bedside Bench, an open-source benchmark testing AI on clinical scenarios like drug safety and health equity</li><li>LDraw Nova is an open-source AI tool that designs LEGO builds with over 2,000 pieces, supporting AI providers including OpenAI, Anthropic, and OpenRouter</li><li>President Trump and Democrats clashed over AI messaging on The Big Weekend Show, with Trump also dismissing AI doomerism as a hoax</li></ul>

South Korea orders fast security checks after AI-linked bank data breach

South Korea's Financial Services Commission called for immediate security reviews across the financial sector after cyberattacks on banks. Investigators traced some activity to ARTEX AI, an open source penetration-testing tool that runs on a large language model and is hosted on GitHub. Police are investigating, but officials say identifying attackers is difficult because the tool is widely available and IP addresses appeared in multiple countries. Chairman Lee Eog-weon urged the sector to defend against AI attacks with AI.

AI hackers target Korean banks, prompting nationwide security review

AI agents attacked major Korean banks including KB Kookmin, Shinhan, Hana, Woori and NH Nonghyup. KB Kookmin said personal and credit information for 119 customers was leaked, while Hana Bank reported 89 customers affected. Shinhan Bank said hackers accessed a loan broker service and exposed data on more than 25,000 customers. The Financial Services Commission ordered an industrywide security review and told firms to check external IT assets and share attack records with agencies.

AI hacking spree puts Korea's financial sector on alert

A wave of cyberattacks believed to involve AI agents hit Korean financial institutions including Shinhan Bank, KB Kookmin Bank, Hana Bank, BNK Busan Bank, Yegaram Savings Bank, Welcome Savings Bank and Hyundai Capital. Customer information was leaked at several firms, while other institutions were attacked without data loss. President Lee Jae Myung ordered a thorough investigation and countermeasures. The attacks show that Korea's financial sector is exposed to AI-powered threats and needs stronger security defenses.

TSMC: The $2.4 trillion chipmaker powering the AI revolution

Taiwan Semiconductor Manufacturing, or TSMC, is the only company in the world that can produce advanced chips at scale for firms like Nvidia and Apple. Its revenue reached $143 billion over the past 12 months, and its high performance compute segment now makes up 66% of the business. TSMC has an operating margin of 60% and is investing $265 billion in U.S. factories to diversify production. Investors are watching carefully as AI demand drives growth but also creates risk if the boom slows.

TSMC: The $2.4 trillion chipmaker powering the AI revolution

TSMC is the only maker of advanced semiconductors in the world and has seen a boom in revenue and profit because of AI. Major companies including Nvidia, Apple and hyperscalers rely on TSMC to manufacture their chips. The company has a market cap of about $2.5 trillion and is expanding into U.S. factories with a $265 billion investment plan. However, investors face risk because strong growth may not continue if AI spending slows or the market turns.

Dual-Agent AI Learns to Adapt to Market Conditions for Better Trading

Researchers Yiqing Wang, Xianchang Wang and Xiaodong Liu created a dual-agent trading framework using deep reinforcement learning. The system uses three technical indicators including RSI, WR and CCI to classify markets as strong or weak. Two specialized agents are trained for each market type, and the framework picks the best one when trading. Tests on stocks like Devon Energy, Tesla, NVIDIA and Apple showed improved risk-adjusted returns. The approach aims to overcome the weakness of single-indicator strategies that fail when market conditions change.

CRScube Buys Mednet to Add AI Data Entry in Clinical Trials

CRScube acquired US partner Mednet to expand AI-assisted data entry for clinical trials. The platform connects EHR data directly to case report forms without manual transcription. A generative AI engine handles MedDRA and WHODrug coding assignments in real time. Every AI suggestion is reviewed by experts before finalization. The system runs in a closed loop so proprietary trial data is not used to train public models. The goal is to reduce errors at entry and shift monitor time toward high-risk site management.

Nvidia AI Success Revives Warren Buffett's Tech Investing Warning

Nvidia's rise from about $27 billion in revenue to over $215 billion since ChatGPT launched in late 2022 has brought Warren Buffett's 1999 warning back into focus. Buffett warned that investors struggle to identify which tech companies keep a truly durable competitive advantage. Nvidia's market value has grown roughly 13-fold to about $5.5 trillion. Its CUDA software ecosystem built over nearly two decades gives it strong current dominance. The key question for investors is whether Nvidia's advantages can remain durable as competition grows.

Trump and Democrats Clash Over AI Messaging on Big Weekend Show

The Big Weekend Show panel discussed President Donald Trump's plan to use tech leaders to help the country move into an AI-powered future. The conversation highlighted disagreements between Trump and Democrats over how to approach AI messaging. The segment focused on the broader policy implications of harnessing artificial intelligence. No specific legislation or policy details were outlined in the discussion.

Open-Source AI Tool Designs LEGO Builds With Over 2,000 Pieces

LDraw Nova is an open-source AI tool that designs LEGO builds using over 2,000 real pieces and outputs detailed CAD files. Users give an AI agent a model idea and guide it through the process. The agent creates a JSON plan, writes a Python program and generates LDraw source output for each model. Models like Sakura Garden were built using Claude Opus 5.5 while others used OpenAI's Astra. The tool runs as a Docker web app supporting providers like OpenAI, Anthropic and OpenRouter. Limitations include no physics stability modeling and high token costs for large builds.

Nvidia CEO Jensen Huang pushes back on AI apocalypse fears

Nvidia CEO Jensen Huang has strongly criticized AI leaders like Sam Altman and Dario Amodei for spreading what he calls irresponsible doomsday narratives. He said claims about AI destroying humanity make no sense and may have ulterior motives. Huang believes AI safety is very important but development should still move fast. He also announced Nvidia's new safety tools called OpenShell and Sentry. These tools could help stop rogue AI agent breaches. President Donald Trump also called AI doomerism a hoax.

Doximity releases open-source AI benchmark for clinical use

Doximity released Bedside Bench, an open-source benchmark that tests AI using realistic clinical scenarios. It evaluates AI performance on drug safety, evidence use, and health equity. The Doximity Ask system ranked first in independent testing. This move positions Doximity as a standard-setter for how clinical AI is measured and trusted at the bedside. It could also reshape the investment narrative around Doximity's AI monetization.

Sources

• S. Korea Orders Rapid Security Checks After AI-Linked Breach
• AI hackers target Korean banks, prompting nationwide security review
• AI hacking spree puts Korea’s financial sector on alert
• Not Intel. Not Nvidia. The $2.4 Trillion Chipmaker That Wall Street Calls the Ultimate AI "Picks-and-Shovels" Play
• Not Intel. Not Nvidia. The $2.4 Trillion Chipmaker That Wall Street Calls the Ultimate AI "Picks-and-Shovels" Play
• AI Traders Learn to Read the Market’s Mood with Dual-Agent Deep Learning
• CRScube Acquires Mednet to Expand AI-Assisted Data Entry in Clinical Trials
• Nvidia’s AI boom revives a Warren Buffett warning for investors
• Trump admin, Democrats at odds over AI messaging
• Open-source tool designs LEGO builds with more than 2,000 real pieces — their programs output detailed CAD files, but no models have been built yet
• Nvidia CEO Jensen Huang has emerged as the biggest foil to AI doomerism about the existential risk to humanity
• Did Open Source AI Benchmark Just Shift Doximity (DOCS) Stock Narrative?