• NATURAL 20
  • Posts
  • Alibaba Launches Open-Source AI Models

Alibaba Launches Open-Source AI Models

PLUS: AI Phone Powered by Perplexity Assistant, Opera Launches AI Agent Within Browserand more.

In partnership with

Cold Email Setup Offer

We started sending 10,000 cold emails per day, and scaled a brand new B2B offer to $108k MRR in 90 days. Now, you can have the same system set up (completely done-for-you) inside your own business - WITHOUT going to spam, spending thousands of dollars, or any manual input. Close your next 20 clients easily. We’ll set up the tech, write your scripts, give you the leads, give you the inboxes, and the sending tool - all starting at $500/mo.

Today:

  • Alibaba Launches Open-Source AI Models

  • Microsoft Launches Dragon Copilot for Healthcare

  • Anthropic Reaches $61.5B Valuation with Series E

  • Google Launches Free Gemini Data Agent

  • AI Phone Powered by Perplexity Assistant

  • Opera Launches AI Agent Within Browser

Deepseek R2 and Wan 2.1 | Open Source DESTROYS everyone

Alibaba has launched its latest AI video model, One 2.1, which includes text-to-video and image-to-video capabilities. It's open-source and runs on consumer-grade GPUs. Meanwhile, DeepSeek is accelerating the release of its R2 model, which promises better coding and multilingual reasoning. 

The disruptive nature of DeepSeek’s cost-effective, open-source models is challenging larger companies like OpenAI. These developments are reshaping AI access, sparking global competition and innovation.

Microsoft has launched Dragon Copilot, an AI assistant designed to streamline clinical workflows. It combines voice dictation from Dragon Medical One and ambient listening from DAX to improve documentation, automate tasks, and enhance clinician well-being. The tool aims to reduce burnout, save time, and improve patient experiences. Dragon Copilot will be available in the U.S. and Canada in May 2025, with further global expansion planned. It aligns with Microsoft’s commitment to responsible AI in healthcare.

Anthropic has raised $3.5 billion in a funding round, reaching a $61.5 billion post-money valuation. Led by Lightspeed Venture Partners, with participation from major investors like Fidelity and Salesforce Ventures, the funds will be used to advance AI development, expand compute capacity, and further research in mechanistic interpretability and alignment. The company’s Claude AI, including the new Claude 3.7 Sonnet, is already being used by businesses like Zoom, Pfizer, and Replit to enhance productivity.

Google has launched a free Gemini 2.0-powered Data Science Agent on its Colab platform. This AI assistant automates data analysis by generating fully-functional Jupyter notebooks based on natural language inputs. It helps users quickly analyze datasets, clean data, and visualize trends without needing manual coding. Currently available to users aged 18+ in select regions, it streamlines workflows for researchers and developers, accelerating scientific and industrial research while saving significant time.

Deutsche Telekom, the parent company of T-Mobile, is developing a new "AI Phone" powered by the Perplexity Assistant. The device will be controlled primarily by voice and run on "Magenta AI," offering tools like Perplexity, Google Cloud AI, ElevenLabs, and Picsart. The AI phone will allow users to perform tasks such as booking flights and making restaurant reservations without switching apps. It will launch later this year, with AI features also available on other devices via the MeinMagenta app.

Opera has introduced a new AI agent called Browser Operator, which operates directly within the Opera browser. The agent can complete tasks on various websites, such as shopping for items, booking tickets, and finding flights. It works securely on-device, unlike cloud-based alternatives. Users can see the agent's actions and take control at any point. The feature will be available soon through Opera's Feature Drop program, joining a competitive field of AI-powered browser agents.

🧠RESEARCH

SolutionRAG is a system designed to improve the design of complex engineering solutions. It uses tree-based exploration and bi-point thinking to generate reliable solutions. The authors present SolutionBench, a benchmark to evaluate solution generation, demonstrating that SolutionRAG outperforms previous methods in solving engineering challenges.

Chain of Draft (CoD) is a new approach for Large Language Models (LLMs) that mimics human cognitive efficiency. CoD generates concise, informative intermediate reasoning, reducing verbosity while maintaining accuracy. It outperforms Chain-of-Thought (CoT) in accuracy and significantly reduces token usage, cost, and latency for reasoning tasks.

muCode is a simple approach for multi-turn code generation using single-step rewards. Unlike complex methods, muCode trains a generator and verifier to optimize code solutions based on multi-turn execution feedback. The approach outperforms existing methods, providing significant improvements in generating accurate code efficiently.

🛠️TOP TOOLS

Ecrett Music - AI-powered music composition platform designed specifically for content creators, including video producers, game developers, podcasters, and other digital content makers.

Hubble AI - Designed to make artificial intelligence (AI) accessible and integrable into various business operations without the need for coding expertise.

GhostCut - AI-driven video authoring tool designed to enhance localized video marketing and simplify professional video production.

Neeva AI - Search engine that stood out for its commitment to user privacy and an ad-free experience.

Echo Labs - AI-powered tool designed to transform the way digital content is created, particularly in the realm of video production.

📲SOCIAL MEDIA

🗞️MORE NEWS

  • Unitree Robotics has open-sourced its G1 humanoid robot’s algorithms and hardware, enabling rapid development. The robot now performs martial arts moves, showcasing significant advancements in China’s robotics race following its debut at the Spring Festival Gala.

  • Stability AI partnered with Arm to bring generative audio to smartphones, enabling high-quality, on-device audio generation without an internet connection. This innovation reduces generation time by 30x and enhances content creation on mobile devices.

  • Podcastle has launched its AI text-to-speech model, Asyncflow v1.0, featuring over 450 voices. The platform offers an API for developers to integrate this technology, providing cost-effective, high-quality audio generation for diverse use cases like marketing and education.

  • TWG Global combines assets from Walter's interests in sports and Guggenheim with Tull's investments in biotechnology, defence start-ups, and media. The partnership aims to implement technology enhancements, identify unique opportunities, and build broad AI capabilities across portfolio companies.

  • Micron unveiled its first 1y DDR5 memory chip samples, enhancing AI processing. The company also launched LPDDR5X and UFS 4.1 mobile storage solutions, offering improved performance and power efficiency for AI-driven smartphones and mobile apps.

  • Google will soon introduce screen-sharing and live video capabilities for Gemini Advanced subscribers on Android through the Google One AI Premium plan. These features are part of Project Astra, enabling users to ask Gemini questions about their device screen.

  • A new study by Copyleaks reveals that 74.2% of DeepSeek's written text resembles ChatGPT's style, suggesting potential use of OpenAI's outputs without authorization. This raises concerns about AI training transparency, intellectual property, and regulation.

What'd you think of today's edition?

Login or Subscribe to participate in polls.

Reply

or to participate.