Google Replacing Assistant with Gemini 🎙️, Sesame Releases AI Model Behind Maya 🚀, Anthropic API Updates 💾

Start learning AI in 2025

Keeping up with AI is hard – we get it!

That’s why over 1M professionals read Superhuman AI to stay ahead.

• Get daily AI news, tools, and tutorials

• Learn new AI skills you can use at work in 3 mins a day

• Become 10X more productive

Sign up and start mastering AI

💬 OpenAI Developments

OpenAI declares AI race “over” if training on copyrighted works isn’t fair use

OpenAI has warned that without clear policies allowing AI models to train on copyrighted material under fair use, the U.S. risks losing its AI leadership to China. The company has urged President Trump to address this issue in the upcoming AI Action Plan to maintain national security and technological competitiveness.

OpenAI calls DeepSeek ‘state-controlled,’ calls for bans on ‘PRC-produced’ models

In a recent letter to the U.S. government, OpenAI criticized Chinese competitor DeepSeek, labeling it as state-controlled and expressing concerns over potential security risks due to possible manipulation by the Chinese government. OpenAI urged the government to consider banning AI models produced in the People's Republic of China to safeguard national security and maintain technological competitiveness.

OpenAI’s proposals for the U.S. AI Action Plan

OpenAI's proposals for the U.S. AI Action Plan focus on maintaining AI leadership by balancing innovation with regulation, securing export controls, and clarifying copyright policies to support AI development. The plan emphasizes infrastructure investment, AI adoption in government, and fostering international AI cooperation while safeguarding national security.

🚀 Tech Industry Moves

Nvidia: Sheikh Tahnoon bin Zayed Al Nahyan, the UAE's national security adviser and brother of the president, plans to visit Washington to seek approval for purchasing additional Nvidia AI chips. This visit aims to ease access to advanced U.S. technology, essential for the UAE's ambitions to become a regional AI leader. The discussions will also address U.S. export controls that currently limit the sale of such high-tech components to the UAE.

Nvidia: Samsung Electronics is collaborating with NVIDIA to enhance AI capabilities in mobile networks by integrating NVIDIA's accelerated computing platform with Samsung's virtualized Radio Access Network (vRAN). This partnership aims to facilitate the seamless adoption of AI in mobile networks, optimizing performance across diverse deployment environments, from rural to dense urban areas.

Google is replacing Assistant with Gemini – Google announced that it will phase out Google Assistant on mobile devices later this year, replacing it with Gemini. The transition will extend to tablets, cars, smartwatches, headphones, and home devices like speakers and TVs. Google is improving Gemini with features like music playback, timers, and lock screen actions to ensure a smooth shift. This move aligns with Gemini’s role as the default assistant on the latest Pixel devices, offering more advanced AI-powered capabilities.

Google has joined OpenAI in urging President Trump to relax AI copyright restrictions, emphasizing that current limitations hinder AI development and competitiveness. Both companies argue that allowing AI models to train on copyrighted material under fair use is essential for maintaining U.S. leadership in artificial intelligence.

Anthropic Optimizes API for Token Efficiency – Anthropic has introduced token-saving updates for Claude 3.7 Sonnet, including cache-aware rate limits, improved prompt caching, and token-efficient tool use, reducing costs by up to 90% and latency by 85%. The new text_editor tool enables targeted text edits, enhancing collaboration. These updates are now available on the Anthropic API, Amazon Bedrock, and Google Cloud's Vertex AI, helping developers boost throughput and efficiency with minimal code changes.

DeepSeek Several Chinese banks, including the Industrial and Commercial Bank of China (ICBC) and the Postal Savings Bank of China, have integrated DeepSeek AI models into various operations such as customer service and credit approval. This integration aims to enhance operational efficiency and service quality, though experts caution about potential data security risks and emphasize the need for compliance with relevant regulations.

💡 Emerging AI Technologies and Companies

Sesame, the startup behind the viral virtual assistant Maya, releases its base AI model

Sesame Releases AI Model Behind Maya – AI startup Sesame has open-sourced CSM-1B, the base model powering its viral voice assistant, Maya. The 1-billion-parameter model, licensed under Apache 2.0, generates realistic AI voices but lacks safeguards against misuse. Co-founded by Oculus creator Brendan Iribe, Sesame is also working on AI-powered glasses for all-day wear.

Big tech’s AI models come alive in humanoid robots

Tech giants are integrating advanced AI models into humanoid robots, enhancing their adaptability and functionality. Google's DeepMind unveiled Gemini Robotics, developed in collaboration with Texas-based Apptronik, showcasing robots capable of tasks like filling lunchboxes and folding paper. Similarly, OpenAI partnered with Figure to release the Figure 01 robot, though this collaboration ended in early February. Meanwhile, Elon Musk introduced Tesla's Optimus Gen 2 at the "We, Robot" event in October, where the robot performed social activities such as serving drinks to attendees.

Evolve programme gives students with skills for future careers in robotics, AI, and nuclear energy

The Evolve programme has engaged over 100 students from 11 West Cumbria schools, including home-educated learners, in a five-day immersive experience tackling real-world challenges in robotics, AI, and nuclear energy. Participants collaborate to devise technological solutions for monitoring Sellafield's extensive pipework, enhancing skills in problem-solving, creativity, and teamwork, while gaining insights into diverse career paths within the nuclear and energy sectors.

🌍 Global AI Race

Tech leaders at Anthropic, IBM, and Meta warn that AI is coming for software developer jobs

Advancements in artificial intelligence are rapidly transforming software development, with AI systems increasingly capable of handling complex coding tasks. Dario Amodei, CEO of Anthropic, predicts that within 3 to 6 months, AI will write 90% of software code, potentially reaching full automation within a year. Similarly, JPMorgan Chase reports a 10% to 20% boost in software engineers' efficiency through the use of AI coding assistants. However, industry leaders from companies like Meta and IBM express differing views on the timeline and extent of AI's impact on software development roles.

If you want more updates related to AI, subscribe to our Newsletter