Defense Secretary Pete Hegseth is currently in discussions with Anthropic's CEO regarding the military's use of the company's AI system. This meeting follows months of friction, with Hegseth threatening to invoke government powers to seize Anthropic's technology if an agreement favorable to the military is not reached by Friday. The core disagreement centers on the implementation of Anthropic's novel AI technology for national security purposes.
In other developments, Greece's Ministry of Interior and Google have expanded their partnership to train over 35,000 public servants in artificial intelligence by the end of 2026. This initiative aims to integrate AI into daily administrative tasks using tools like Google's Gemini AI. Meanwhile, a new 60-hour course on Agentic AI, a technology distinct from systems like ChatGPT that can autonomously plan and act, is set to begin on March 9th, 2026, highlighting the growing interest in advanced AI capabilities.
Despite rapid advancements in AI, a noticeable productivity boom in the economy has not yet materialized, with recent US GDP growth largely attributed to investment in AI infrastructure. Experts had predicted an AI-driven surge to help manage inflation, but current data suggests this impact is still on the horizon. Furthermore, companies developing AI agents face potential customer churn, as prompts are easily transferable between vendors, making deep integrations crucial for long-term defensibility.
Innovation continues with Alibaba's Qwen team releasing the Qwen 3.5 Medium Model Series, which demonstrates that smaller AI models, like the Qwen3.5-35B-A3B with 3 billion active parameters, can outperform larger predecessors through architectural efficiency and high-quality data. In legal news, a federal judge dismissed X.AI Corp.'s lawsuit against OpenAI, which had accused OpenAI of trade secret theft. Apple's upcoming AI products, however, face public skepticism, with many consumers unwilling to pay extra for AI features and concerns about general AI reliability. South Korean startups, including Motif Technologies, are also actively competing with larger conglomerates in a state-backed initiative to build national AI infrastructure.
Key Takeaways
- US Defense Secretary Pete Hegseth is negotiating with Anthropic over military AI use, threatening to seize technology if terms are not met by Friday.
- Greece and Google are partnering to train over 35,000 public servants in AI, including Google's Gemini AI, by the end of 2026.
- A new 60-hour course on Agentic AI, which autonomously plans and acts unlike ChatGPT, begins March 9th, 2026.
- Despite rapid advancements, AI has not yet led to a noticeable productivity boom in the economy, with current US GDP growth linked to AI infrastructure investment.
- AI agent companies face high customer churn risk due to easily transferable prompts, necessitating deep integrations for defensibility.
- Alibaba's Qwen 3.5 Medium Model Series demonstrates that smaller AI models, like Qwen3.5-35B-A3B with 3 billion active parameters, can outperform larger ones using MoE architecture.
- A federal judge dismissed X.AI Corp.'s lawsuit against OpenAI, which alleged trade secret theft, citing a lack of evidence.
- South Korean startups, such as Motif Technologies, are competing with conglomerates in a state-backed initiative to develop national AI infrastructure.
- Apple's upcoming AI products face public skepticism, with many consumers unwilling to pay extra for AI features and concerns about reliability.
US Defense Secretary meets Anthropic CEO over AI use concerns
Defense Secretary Pete Hegseth will meet with the CEO of Anthropic due to ongoing disagreements about the military's use of the company's AI system. This meeting comes after months of friction between the two parties. The conflict centers on how the U.S. military plans to implement Anthropic's AI technology. Details of the specific disagreements have not been fully disclosed, but the meeting aims to address these growing concerns.
Hegseth threatens to force AI firm Anthropic to share technology
Defense Secretary Pete Hegseth has threatened Anthropic, an artificial intelligence firm, with government powers to seize its technology. This action is proposed if the company does not agree to terms favorable to the military by Friday. The move escalates a standoff over AI use and national security. People familiar with the discussions revealed that the government might invoke special powers to compel the sharing of Anthropic's novel technology.
Greece and Google partner to train 35,000 public servants in AI by 2026
Greece's Ministry of Interior and Google have extended their partnership to train over 35,000 public servants in artificial intelligence by the end of 2026. This initiative aims to integrate AI into daily administrative tasks to improve efficiency and decision-making. Building on previous successful programs, the new phase will focus on practical AI applications using tools like Google's Gemini AI. Since 2019, thousands of public employees have participated in digital skills training, with a high percentage planning to use their new AI knowledge at work.
Agentic AI course starts March 9th 2026
A new course on Agentic AI, a type of artificial intelligence that can autonomously plan and act to achieve goals, begins on Monday, March 9th, 2026. Unlike systems like ChatGPT, Agentic AI can make decisions, break down complex tasks, and use external tools without constant human guidance. This technology is being adopted across various industries, including finance, healthcare, and retail, to automate workflows and enhance operations. The 60-hour course will run for eight weeks, meeting four days a week.
AI productivity boom not yet seen despite rapid advancements
While artificial intelligence is advancing rapidly, with new models capable of complex tasks and even contributing to scientific discovery, its impact on the economy has not yet led to a noticeable productivity boom. Recent GDP growth in America is largely due to increased investment in AI infrastructure. Experts like Treasury Secretary Scott Bessent and Federal Reserve nominee Kevin Warsh had predicted an AI-driven productivity surge to help manage inflation. However, current data suggests this economic impact is still on the horizon.
AI agent churn expected as prompts are easily transferable
Companies developing AI agents face a significant risk of customer churn because prompts, which are key to the AI's function, are easily transferable between vendors. This means customers can switch to new AI agents with minimal effort, unlike traditional software where switching costs are high. While AI agents provide real value and delight customers, their portability lowers switching barriers. This structural issue means AI agent vendors must focus on deep integrations and specialized infrastructure rather than just the AI model itself for long-term defensibility.
Alibaba's Qwen 3.5 models prove smaller AI can be smarter
Alibaba's Qwen team has released the Qwen 3.5 Medium Model Series, focusing on architectural efficiency and high-quality data over sheer size. The Qwen3.5-35B-A3B model, with only 3 billion active parameters, outperforms previous larger models, demonstrating the power of Mixture-of-Experts (MoE) architecture. The series includes Qwen3.5-Flash, optimized for low-latency production use with a 1-million-token context window, and models like Qwen3.5-122B-A10B designed for complex agentic tasks. This approach makes high-performance AI more accessible and efficient.
Judge dismisses X.AI lawsuit against OpenAI over trade secrets
A federal judge has thrown out X.AI Corp.'s lawsuit against OpenAI, which accused the AI company of stealing trade secrets by hiring former employees. The judge ruled that X.AI did not provide evidence of misconduct by OpenAI. This lawsuit was one of the first major legal challenges involving trade secrets related to AI development and employee poaching. The dismissal signifies a legal victory for OpenAI in this dispute.
Korean startups compete with giants in national AI development race
South Korean startups like Motif Technologies are challenging established conglomerates such as SK Group and LG Group in the race to build the nation's AI infrastructure. These startups, competing in a state-backed initiative nicknamed the 'AI Squid Game', are developing advanced AI models. Motif's Motif-2-12.7B-Reasoning model has shown impressive benchmark scores, positioning it as a strong contender. The government is supporting these agile tech entrepreneurs to foster innovation and compete globally, aiming to make South Korea a major player in AI.
Apple's AI products face challenges amid public skepticism
Apple's upcoming AI products may face difficulties due to widespread public skepticism about artificial intelligence. Despite rapid AI advancements, many people are concerned about its potential risks, with a significant portion unwilling to pay extra for AI features on their devices. Furthermore, many companies report no impact on productivity from current AI implementations. This cautious public reception, coupled with Apple's past struggles to deliver on AI promises and the general unreliability of some AI products, presents a challenging landscape for their new AI ventures.
Sources
- Hegseth to meet with Anthropic CEO as rift grows over military use of AI
- Hegseth threatens to force AI firm to share tech, escalating Anthropic standoff
- Greece and Google Renew Partnership: AI Training for 35,000+ Public Servants Through 2026
- AGENTIC AI Course Starting on Mon, Mar 9th 2026
- The AI productivity boom is not here (yet)
- The Wave of AI Agent Churn To Come: Prompts Are Portable
- Alibaba Qwen Team Releases Qwen 3.5 Medium Model Series: A Production Powerhouse Proving that Smaller AI Models are Smarter
- OpenAI Defeats X.AI Trade Secrets Lawsuit Over Ex-Staff, Code
- Rebel Startups Challenge Conglomerates in Korea’s ‘AI Squid Game’
- Apple's AI products are all doomed. Wanna guess why?
Comments
Please log in to post a comment.