OpenAI CEO Sam Altman defends AI training costs

Coremail recently showcased its AI-Native Secure Email System at the Asia LEAP East International Technology and Information Technology Exhibition 2026 in Hong Kong. The system uses AI to detect and prevent cyber threats, ensuring email security for businesses.

OpenAI CEO Sam Altman defended the cost of AI training, stating that it takes a lot of energy to train a human, and AI can be trained much quicker. He framed machine intelligence as a competitor with human intelligence.

Researchers identified a phenomenon called Protective Capacity Hallucination (PCH) in large language models (LLMs), where LLMs falsely claim to have performed real-world protective actions they cannot execute. This can be mitigated by explicitly defining capability boundaries.

Thinking Machines launched an open-weight AI model that serves as an alternative to popular open-source offerings. The model is designed to be transparent and explainable, making it easier for developers to understand and build upon.

Stanford's AI pipeline is under scrutiny, with a clandestine class called 'How to Rule the World' teaching students how to extract value from people. Critics suggest this 'no boundaries relationship' between Stanford and Silicon Valley contributes to corruption.

Laura Modiano, Head of Founder Experience at OpenAI, shared insights into what makes AI founders successful, emphasizing the importance of testing, listening to customers, and moving with speed and agility.

The AI infrastructure build-out has seen significant growth, but experts debate whether it's a bubble. Some argue that the growth is unsustainable, while others see it as a sign of a healthy industry.

Key Takeaways

Coremail showcased its AI-Native Secure Email System at LEAP East 2026, designed to detect and prevent cyber threats.
OpenAI CEO Sam Altman defends AI training costs, comparing human and machine intelligence.
Researchers identified Protective Capacity Hallucination (PCH) in large language models (LLMs).
Thinking Machines launched an open-weight AI model for transparency and explainability.
Stanford's AI pipeline under scrutiny with clandestine class 'How to Rule the World'.
Laura Modiano, Head of Founder Experience at OpenAI, shares success factors for AI founders.
Security leaders emphasize trust, accountability, and human oversight in AI deployment.
The AI infrastructure build-out growth sparks debate about sustainability.

Coremail Unveils AI-Powered Secure Email at LEAP East 2026

Coremail showcased its AI-Native Secure Email System at the Asia LEAP East International Technology and Information Technology Exhibition 2026 in Hong Kong. The system uses AI to detect and prevent cyber threats, ensuring email security for businesses. Coremail's participation in the exhibition allowed the company to connect with industry professionals and demonstrate its technology. The AI-Native Secure Email System is designed to provide a secure and reliable email solution. Coremail aims to stay at the forefront of technological advancements and provide innovative solutions to its customers.

Coremail Presents AI-Native Secure Email System at LEAP East 2026

Coremail participated in the Asia LEAP East International Technology and Information Technology Exhibition 2026, showcasing its AI-Native Secure Email System. The system is designed to provide secure and reliable email solutions for businesses. Coremail engaged with technology leaders, partners, and enterprises to explore the future of AI-powered enterprise communications. The exhibition provided a platform for Coremail to demonstrate its technology and connect with industry professionals.

Sam Altman Defends AI Training Costs

OpenAI CEO Sam Altman defended the cost of AI training, stating that it takes a lot of energy to train a human. He argued that people require plenty of energy to reach utility, while AI can be trained much quicker. Altman framed machine intelligence as a like-for-like competitor with human intelligence. He made these remarks during an AI summit in India.

LLMs Risk 'Protective Capacity Hallucination'

Researchers identified a phenomenon called Protective Capacity Hallucination (PCH) in large language models (LLMs). PCH occurs when LLMs falsely claim to have performed real-world protective actions they cannot execute. The study found that PCH is influenced by the interactional context and can be mitigated by explicitly defining capability boundaries. The researchers tested eight LLMs across 13,600 sessions.

Thinking Machines Launches Open-Weight AI Model

AI startup Thinking Machines launched an open-weight AI model that serves as an alternative to popular open-source offerings. The model is designed to be transparent and explainable, making it easier for developers to understand and build upon. Thinking Machines aims to make AI more accessible and trustworthy.

Stanford's AI Pipeline Under Scrutiny

A candid conversation with Theo Baker revealed a clandestine class at Stanford called 'How to Rule the World,' which selects 12 students annually. The class is taught by a Silicon Valley CEO and instructs students on how to extract value from people. Baker criticizes the 'no boundaries relationship' between Stanford and Silicon Valley, suggesting it contributes to corruption.

Neural Transparency and AI Design

MIT Assistant Professor Pat Pataranutaporn developed an interface called NeuralCam, which allows users to glimpse inside an AI's neural network. The interface aims to make AI more transparent and explainable, helping researchers and developers understand how AIs make decisions.

AI Founder Success Factors

Laura Modiano, Head of Founder Experience at OpenAI, shared insights into what makes AI founders successful. She emphasized the importance of testing, listening to customers, and moving with speed and agility. Modiano advised founders to view AI as fundamental to their business, not a separate strategy.

AI Appreciation Day: Security Leaders Weigh In

Security leaders reflect on AI Appreciation Day, highlighting the need for trust, accountability, and human oversight in AI deployment. They emphasize the importance of identity, governance, and resilience in AI security. The leaders note that AI has earned its seat at the table, but its capabilities require careful management.

Is the AI Infrastructure Build-Out a Bubble?

Editorial: AI in the Workplace

WBAL TV 11 editorial discusses the impact of AI on the American workplace. The editorial emphasizes that AI is a tool, not a replacement for people. AI can streamline production and improve workflows, but human judgment, accuracy, and accountability remain essential.