Anthropic risks contract as OpenAI discusses changes

Key Takeaways

• The Pentagon threatens to end a $200 million contract with Anthropic due to its refusal to remove AI safeguards against lethal autonomous weapons and mass surveillance.
• Anthropic's Claude model was the first AI in the Pentagon's classified networks and was reportedly used with Palantir in an operation to capture Nicolas Maduro.
• OpenAI, Google, and xAI show more willingness than Anthropic to remove AI usage restrictions for military applications.
• AI significantly boosts coding productivity, with GitHub reporting 35% growth in code pushes in the US and 30% in the UK in 2026, driven by tools like GitHub Copilot and ChatGPT.
• Axios's CTO, Dan Cox, reported that AI agent teams helped reduce their tech team from 63 to 43 people while doubling output, with one engineer completing a three-week project in 37 minutes.
• Anthropic continues to hire software engineers, emphasizing their role in strategic thinking, customer empathy, and technical leadership beyond just writing code.
• Nineninesix.ai released Kani-TTS-2, an open-source text-to-speech AI model with 400 million parameters, capable of zero-shot voice cloning on consumer-grade GPUs with 3GB of VRAM.
• HUUCH launched its AI full-body motion gaming platform, HUUCH Box, at CES 2026, offering controller-free, family-friendly games for 1-4 players.
• Congress is pursuing bipartisan legislation to allow individuals, including figures like Paris Hilton, to sue AI companies for nonconsensual deepfake images.
• Justin Sun, founder of TRON, introduced AINFT at Consensus Hong Kong 2026, enabling blockchain payments for accessing AI models from OpenAI, Anthropic, and Google.

Pentagon Threatens Anthropic Over AI Military Limits

The Pentagon may end its partnership with AI company Anthropic due to disagreements over how the US military uses its technology. Anthropic has safety features that stop its AI from being used for things like lethal autonomous weapons systems. The Defense Department warned Anthropic it might cut off access to its cloud computing environment if these safeguards are not removed. This dispute highlights the ongoing tension between fast AI development and ethical concerns in warfare.

Pentagon Pressures Anthropic on AI Military Access

The Pentagon is pushing AI firm Anthropic to allow its tools for all military uses, including weapons development and intelligence. Anthropic insists on limits, especially against mass surveillance of Americans and fully autonomous weapons. This disagreement has led to threats from the Pentagon to end their $200 million contract. Anthropic's Claude model was the first AI in the Pentagon's classified networks. Other AI companies like OpenAI, Google, and xAI have shown more willingness to remove similar safeguards.

Pentagon May Cut Anthropic Over AI Usage Rules

The Pentagon is considering ending its relationship with AI company Anthropic, according to Axios. Anthropic wants to keep restrictions on how the US military uses its AI models. The Pentagon, however, wants to use AI tools for all lawful purposes, including weapons development and battlefield operations. OpenAI, Google, and xAI are also part of these discussions. Anthropic's AI model Claude was reportedly used in an operation to capture Nicolas Maduro through a partnership with Palantir.

Pentagon Warns Anthropic on AI Safeguards

The Pentagon is thinking about ending its work with AI company Anthropic. This is due to Anthropic's insistence on keeping some limits on how the US military uses its AI models. The Pentagon wants its AI tools available for "all lawful purposes," including in weapons development and intelligence. Anthropic's spokesperson stated their talks focus on strict limits against fully autonomous weapons and mass domestic surveillance. Other AI companies like OpenAI, Google, and xAI are also involved in these discussions.

Justin Sun Discusses AI and Blockchain at Hong Kong Event

Justin Sun, founder of TRON, spoke at Consensus Hong Kong 2026 about AI and strong blockchain systems. He introduced AINFT, a new project that lets users access AI models from OpenAI, Anthropic, and Google with payments on the blockchain. Sun believes blockchain provides the stable financial base AI needs for global commerce and agent economies. He also discussed how AI agents will use blockchain for identity and transactions. TRON DAO sponsored the event and highlighted its leadership in AI and stablecoin infrastructure.

Congress Pushes Lawsuits Against Deepfake AI Companies

There is a bipartisan effort in Congress to allow people to sue AI companies. This push is about nonconsensual AI-generated deepfake images. Figures like Paris Hilton and AOC support this move. The goal is to create legal ways to address harm caused by these fake images.

HUUCH Launches AI Full-Body Motion Gaming Platform

AI startup HUUCH has finished developing its new full-body motion gaming platform and products. Their main product, HUUCH Box, is a home gaming device that uses special 3D tracking technology. Players control games with natural body movements, without needing controllers or wearables. The platform, which debuted at CES 2026, offers family-friendly games like adventure, sports, and dance. HUUCH Box supports one to four players and aims to bring families together through active, shared play. It is set to launch officially in the first half of the year with monthly content updates.

Axios CTO Shares AI Impact on Tech Work

Dan Cox, the Chief Technology Officer at Axios, describes how AI is changing tech work. He shared that an engineer completed a three-week project in just 37 minutes using AI-based agent teams. Axios's tech team shrunk from 63 to 43 people but doubled its output in January, expecting to double it again. Cox noted that AI has eliminated "technical debt," allowing engineers to focus on building new things. He also mentioned that software contracts are becoming shorter because AI can build tools at near-zero cost.

New Kani-TTS-2 AI Model Offers Fast Voice Cloning

Nineninesix.ai released Kani-TTS-2, a new open-source text-to-speech AI model. This model, with 400 million parameters, can run on consumer-grade GPUs with only 3GB of VRAM. It uses an "Audio-as-Language" approach with LiquidAI's LFM2 and NVIDIA NanoCodec to create human-like speech. Kani-TTS-2 was trained on 10,000 hours of data in just six hours and offers zero-shot voice cloning. This means it can instantly replicate a voice from a short audio clip, making it a fast and accessible tool for developers.

Amazon Layoff Becomes Fresh Start for Tech Pro

Hemant Virmani, a 47-year-old tech professional, was laid off from Amazon after 11 and a half years in October 2025. He views this as a refreshing change, using the time to learn new AI skills and focus on his health. His teenage daughter inspired him to stay positive and look forward. After also losing his father, he took time to reflect on his career goals. Virmani has received supportive messages and job leads from a LinkedIn post about his layoff.

Wargaming Essential for Modern AI Era Conflicts

Wargaming is crucial for managing national interests in today's complex hybrid conflicts and the age of AI. It provides a safe way for leaders to test strategies and understand what works or fails. Modern conflicts involve military, economic, cyber, and information pressures, making detailed scenario research vital. Historically, the Western Approaches Tactical Unit used wargaming during World War II to develop anti-submarine tactics that saved many ships. Today, computer-based wargaming uses real and conceptual weapon systems to simulate performance, helping to develop better defense systems.

AI Boosts Coding Productivity Significantly

New data from GitHub shows that AI has greatly increased coding productivity. In 2026, GitHub code pushes grew by about 35% in the US and 30% in the UK, after being mostly flat in previous years. This surge aligns with the widespread use of AI coding tools like GitHub Copilot and ChatGPT. These AI assistants help developers by generating code, suggesting fixes, and even designing entire features. This allows individual developers and small teams to create products much faster, giving companies a competitive edge through quicker development and market entry.

Anthropic Hires Engineers Despite AI Coding Prowess

Boris Cherny, the creator of Claude Code at Anthropic, explains why the company still hires software engineers even though AI writes most of the code. He says that engineering work goes beyond just writing code. Engineers are needed to prompt AI tools, talk to customers, coordinate with other teams, and decide what products to build next. This shift means engineers now focus more on strategic thinking, customer empathy, and technical leadership. Anthropic believes skilled engineers who can use AI tools effectively are more valuable than ever.

Sources

• Pentagon threatens to cut off Anthropic in AI safeguards dispute, Axios reports
• Exclusive: Pentagon threatens to cut off Anthropic in AI safeguards dispute
• Pentagon threatens to cut off Anthropic in AI safeguards dispute, Axios reports
• Pentagon threatens to cut off Anthropic in AI safeguards dispute, Axios reports
• Justin Sun Highlights AI and Resilient Blockchain Infrastructure at Consensus Hong Kong 2026
• Bipartisan push to sue AI companies over deepfake images
• AI Startup HUUCH Completes Foundation of AI Full-Body Motion Products and Game Content Platform
• Axios CTO is living your coming AI reality
• Meet ‘Kani-TTS-2’: A 400M Param Open Source Text-to-Speech Model that Runs in 3GB VRAM with Voice Cloning Support
• I got laid off from Amazon after 11 years. It's a refreshing change.
• Why Wargaming Matters In An Age Of Hybrid Conflict And Artificial Intelligence
• GitHub, Website And App Data Shows How AI Has Led To An Explosion In Coding Productivity
• Claude Code Creator Boris Cherny Explains Why Anthropic Is Still Hiring Software Engineers In Spite Of Nearly All Code Being Written By AI