• The AI Pulse
  • Posts
  • 🤖 NVIDIA’s New Family of Reasoning Engines

🤖 NVIDIA’s New Family of Reasoning Engines

PLUS: AI, Human Authorship, and Copyright Law.

Welcome back AI enthusiasts!

In today’s Daily Report:

  • ⚙️NVIDIA’s New Family of Reasoning Engines

  • 👨‍⚖️AI, Human Authorship, and Copyright Law.

  • 🛠Trending Tools

  • 🥪Brief Bites

  • 💰Funding Frontlines

  • 💼Who’s Hiring?

Read Time: 3 minutes

🗞RECENT NEWS

NVIDIA

⚙️NVIDIA’s New Family of Reasoning Engines

Image Source: YouTube/NVIDIA/“GTC March 2025 Keynote with NVIDIA CEO Jensen Huang”/Screenshot

NVIDIA released “Llama Nemotron,” a new family of Reasoning Engines designed to help Enterprises deploy Agentic AI.

Key Details:
  • It was built using Llama 3.1 8B, Llama 3.3 70B, and Llama 3.1 405B, open-source AI models developed by Meta’s AI Team.

  • It comes in three sizes:

    1. Nano-8B-v1: It was fine-tuned from Llama 3.1 8B to be optimized for deployment on Edge Devices like Smartphones.

    2. Super-49B-v1: It was distilled from Llama 3.3 70B to excel at tackling complex tasks while processing information quickly.

    3. Ultra-249B-v1: It was distilled from Llama 3.1 405B to support the integration of AI Agents across Enterprises at scale.

  • Super-49B-v1 combines the strong reasoning capabilities of “DeepSeek-R1” with the world knowledge of Llama 3.3 70B.

  • It achieved 66% accuracy on the Graduate-Level Google-Proof Q&A (GPQA) benchmark, which includes 448 multiple-choice questions written by domain experts in biology, physics, and chemistry.

Why It’s Important:
  • Super-49B-v1 showcased 20% better accuracy at 5x faster speeds than any other Reasoning Engine that’s currently available.

  • NVIDIA CEO Jensen Huang explained that it beats “DeepSeek-R1” substantially on both accuracy and speed.

🚨Try Super-49B-v1 for free here.

AI AND THE LAW

👨‍⚖️AI, Human Authorship, and Copyright Law.

Image Source: Canva’s AI Image Generators/Magic Media

Synthetic Media Artist and Developer Stephen Thaler tried to register for copyright a piece of artwork generated by his AI System called the “Creativity Machine.”

Key Details:
  • Thaler argued that since he made the AI System that generated the piece of artwork, the copyright of the artwork should be transferred to him.

  • The Copyright Office (CO) denied his application because the artwork lacked human authorship, a prerequisite for a valid copyright.

  • Thaler sued the CO and Shira Perlmutter in her capacity as the Register of Copyrights and the Director of the U.S. CO in the District Court of Columbia.

  • “Nothing in the Copyright Act of 1976 requires human creation,” Thaler argued. “It indicates that when an entity—a natural person, a corporation, or a machine—generates a creative work, that entity is the author.”

  • A Three-Judge Panel for the Court of Appeals for the District of Columbia unanimously agreed with the CO that Thaler couldn’t be granted human authorship because developing an AI System that generates the artwork doesn’t directly involve him in the creative process enough.

Why It’s Important:
  • The duration of a valid copyright is typically limited to “the author’s lifespan or how long a human might live,” wrote Judge Patricia Millett. “Of course, machines don’t have lives, and the length of their operability isn’t measured in the same way as a human life.”

  • She explained that accepting Thaler’s argument would also raise problematic questions around the idea of a machine’s “life” and “death.”

🛠TRENDING TOOLS

📊Mermaid builds complex collaborative diagrams in seconds.

📃Enhancv creates tailored resumes with all the right keywords.

MaxAI chats with any webpage to explain, analyze, and translate.

🔍Arc Browser is the Google Chrome replacement you’ve been waiting for.

📆Motion automatically plans your work day to make you 137% more productive.

🔮Browse our always Up-To-Date AI Tools Database.

🥪BRIEF BITES

Swiss Startup Inait partnered with Microsoft to develop Adaptive Intelligence with Digital Brains.

Perplexity AI is reportedly in talks to raise up to $1 billion, valuing the AI-Powered Answer Engine at $18 billion.

Google Research proposed “Inference-Time Search (ITS),” which focuses on scaling the Search Axis of Test-Time Compute (TTC).

Muse released “Muse S Athena,” an AI-supported wearable that measures brain activity, tracks brain blood flow, and optimizes sleep for peak performance.

💰FUNDING FRONTLINES

  • Firsthand lands a $26M Series A to offer AI-Powered Brand Agents for Marketers and Publishers.

  • LimitlessCNC raises a $4.1M Seed Round to deploy AI Agents tailored for CAM Programming Bottlenecks.

  • Ceramic.ai secures a $12M Seed Round for Enterprise-Grade AI Infrastructure that delivers 2.5x faster Training Speeds.

💼WHO’S HIRING?

  • Salesforce (Palo Alto, CA): Full Stack AI Software Engineer Intern, Summer 2025

  • Anthropic (San Francisco, CA): Safeguards Policy Analyst, Entry-Level

  • Tecton (New York, NY): Software Engineer, Batch Data, Mid-Level

  • Workday (Boulder, CO): Principal AI Agent Engineer, Senior-Level

📒FINAL NOTE

FEEDBACK

How would you rate today’s email?

It helps us improve the content for you!

Login or Subscribe to participate in polls.

❤️TAIP Review of The Day

“Just added this to my daily reading list!🤞”

-Khonsu (1️⃣ 👍Nailed it!)
REFER & EARN

🎉Your Friends Learn, You Earn!

You currently have 0 referrals, only 1 away from receiving 🎓3 Simple Steps to Turn ChatGPT Into an Instant Expert.

Reply

or to participate.