• NATURAL 20
  • Posts
  • Mushroom Cyborgs Blend Biology Robotics

Mushroom Cyborgs Blend Biology Robotics

PLUS: Bezos, OpenAI Invest in Robotics Startup, Meta Joins Defense with Open-Source AI and more.

In partnership with

Hire Ava, the Industry-Leading AI BDR

Ava automates your entire outbound demand generation so you can get leads delivered to your inbox on autopilot. She operates within the Artisan platform, which consolidates every tool you need for outbound:

  • 300M+ High-Quality B2B Prospects

  • Automated Lead Enrichment With 10+ Data Sources Included

  • Full Email Deliverability Management

  • Personalization Waterfall using LinkedIn, Twitter, Web Scraping & More

Today:

  • Mushroom Cyborgs Blend Biology Robotics

  • Anthropic Launches Claude 3.5 Haiku

  • xAI Launches Grok API Beta

  • Bezos, OpenAI Invest in Robotics Startup

  • Meta Joins Defense with Open-Source AI

  • Microsoft Launches GenAI Accelerator Program

New Cyborg (DISTURBING) - Robots Controlled by Fungus, Self Healing Skin and Mushroom Language

Researchers at Cornell University recently developed a biohybrid "cyborg" combining a robot body with living mushroom mycelium. This fungal root network can send electrical signals that control the robot’s movement, reacting to environmental cues like light. The team even experimented with a self-healing fungal “skin.” 

While these cyborgs might help in fields like agriculture by sensing soil health, the technology raises ethical concerns. Releasing mushroom-based robots into ecosystems could disrupt natural systems.

Anthropic launched Claude 3.5 Haiku, an upgraded AI model that improves on speed, accuracy, and versatility. Available via major cloud platforms, it starts at $1 per million input tokens, with discounts for prompt caching and batching. This model is ideal for tasks like coding, interactive chatbots, data labeling, and content moderation, excelling in customer service, software development, and data processing. 

It also upholds high safety standards, with extensive testing for sensitive content handling. Claude 3.5 Haiku meets both general and specialized use cases, offering quick, reliable responses for large-scale applications across various industries.

xAI has launched a public beta for its Grok API, allowing developers to access the Grok foundation models. Through December 2024, users receive $25 in free API credits monthly. Grok’s initial offering includes the grok-beta model, supporting extended contexts up to 128,000 tokens, function calling, and system prompts. A multi-modal version capable of processing images will be available next week. 

Current API documentation aligns with APIs from OpenAI and Anthropic, making migration seamless. Developers can start with Grok by updating their SDK's base URL and API key. 

Physical Intelligence, an AI robotics startup, raised $400 million from investors including Jeff Bezos, OpenAI, and venture firms Thrive and Lux Capital, bringing its valuation to $2 billion. The company aims to create adaptable software for robots, eliminating the need to program individual tasks. The funding highlights increased investments in AI by major tech companies like Amazon, Microsoft, and Nvidia, as the AI market is projected to reach $79.2 billion in 2024. 

Physical Intelligence recently demonstrated its software, π0 (pi-zero), which enables robots to perform tasks like folding laundry and bagging groceries, signaling new possibilities for robotics AI.

Meta is making its open-source AI model, Llama, accessible to U.S. government agencies and defense contractors to enhance national security. Partnering with companies like Oracle, Lockheed Martin, and Scale AI, Meta aims to use Llama for tasks like data analysis, logistics, and cybersecurity. 

This initiative supports America's leadership in AI by fostering an open-source standard rooted in transparency and ethics, countering global competitors, especially China. Through collaborations with the public sector, Meta promotes efficient, ethical AI use in areas like public service, defense, and economic growth, ensuring American innovation aligns with democratic values and international security standards.

Microsoft, in collaboration with NVIDIA and GitHub, has launched the GenAI Accelerator to support promising AI start-ups in the UK. Running from January to March 2025, the program targets UK-based AI start-ups with innovative, impactful products, providing them with mentorship, technical guidance, and networking opportunities. 

Participants will receive support from Microsoft’s AI specialists and NVIDIA’s resources, including the Deep Learning Institute, to accelerate their AI development. The Accelerator aims to foster a sustainable AI community in the UK, culminating in a Demo Day for start-ups to present their solutions to investors. Applications are open until November 22, 2024.

🧠RESEARCH

OS-Atlas is an open-source tool designed to improve graphical user interface (GUI) agents, addressing issues in accuracy and adaptability. Unlike closed models like GPT-4o, OS-Atlas can handle diverse GUIs across platforms like Windows, MacOS, and Android. With 13 million data points, it significantly boosts performance in identifying and navigating new interfaces.

Multi-expert Prompting is a method to improve large language model responses by simulating multiple expert perspectives. It combines these insights to enhance accuracy, factuality, and usefulness, while minimizing harmful content. This approach, using structured decision-making steps, outperforms previous models and boosts response reliability by over 8% in ChatGPT.

This study introduces a two-part method to improve large language models (LLMs) in scientific problem-solving. First, models learn core knowledge directly from tools, then adapt by distinguishing simple from complex problems. This approach improves LLM accuracy by 28% and optimizes tool use, outperforming top models in fields like math and climate science.

HelloMeme introduces a method for enhancing text-to-image models with adapters that support complex tasks without losing flexibility. By refining the attention mechanism on 2D feature maps, this approach improves meme video generation. Compatible with popular open-source models, it offers promising tools for advanced post-training applications.

This study introduces a two-part method to enhance large language models (LLMs) in scientific problem-solving. First, models learn foundational knowledge through tools. Then, they adapt by assessing problem difficulty, using tools only for complex cases. This approach increases accuracy by 28% and precision in tool usage by 13%, outperforming top models.

🛠️TOP TOOLS

Kling AI - Next-Generation AI Creative Studio

Salesforce's Agentforce - Fully customizable autonomous AI agents that can connect to any enterprise data and take action across sales, service, marketing, and commerce

My AskAI - AI customer support agent for your company. 

GyftPro - Discover the perfect gift for any occasion with AI-powered recommendations. 

LinkedIn Hiring Assistant - AI agent that helps take on time-consuming tasks on your behalf.

📲SOCIAL MEDIA

🗞️MORE NEWS

  • Apple’s upcoming iOS 18.2 will allow users to upgrade to ChatGPT Plus directly from the Settings menu, enhancing ChatGPT integration in iOS. The update, expected in December, also adds Visual Intelligence and Genmoji features.

  • Nvidia's AI Blueprint enables developers to create AI agents for analyzing video and images across industries. These agents can identify patterns, summarize content, and alert for safety, enhancing productivity and safety in smart spaces worldwide.

  • Amazon’s Prime Video now offers X-Ray Recaps, an AI tool that summarizes entire TV seasons, episodes, or scenes. Accessible on Fire TV devices, it generates concise recaps while avoiding spoilers, starting with Amazon Originals.

  • Microsoft has launched an AI-powered Xbox chatbot for Xbox Insiders in the U.S. to help players with console and game support. This interactive virtual agent previews Microsoft's broader plans for AI integration in Xbox services.

  • Aravind Srinivas, CEO of Perplexity, offered AI services to replace striking New York Times tech workers during their November 4 strike. This offer, aiming to maintain election coverage, sparked criticism on social media.

  • Lehigh University’s BiomedGPT, an open-source AI model, combines image and text analysis to support healthcare tasks like diagnosis and drug discovery. Tested with Massachusetts General Hospital, it aims to improve medical workflows and patient care efficiency.

What'd you think of today's edition?

Login or Subscribe to participate in polls.

Learn AI with us.

Let’s Build the Future Together.

Hello fellow AI-obsessed traveler,

Over the past 2 years, as we’ve grown to over 250,000 subscribers between the YouTube Channel and this newsletter, we've received an overwhelming number of requests for one specific thing.

While the newsletter helps keep you up to speed with AI news, many of you have asked for the next step: to learn how to actually apply AI in your work.

Today we’re finally announcing the solution with NATURAL 20, the community for like-minded AI learners. As a loyal newsletter reader you are getting access at the lowest price it will ever be:

 JOIN NATURAL 20 AI UNIVERSITY TODAY

What you get:

* Tutorials by experts across various AI fields.

* Daily tutorials by Wes Roth about the latest use cases.

* Building Autonomous AI Agents to Automate Your Life and Business (NEW!)

* A network of the top 1% of early AI adopters.

* Access to community-only resources and software.

* And many more features rolling out soon.

Reply

or to participate.