In partnership with

Hello AI Lovers!
Today’s Topics Are:

- Ai2’s Tulu 3 405B: A New AI Powerhouse Outperforms DeepSeek
- The Future of AI: Key Takeaways from Microsoft and Meta Earnings

Ai2’s Tulu 3 405B: A New AI Powerhouse Outperforms DeepSeek

Quick Summary Ai2, a Seattle-based nonprofit AI research institute, has unveiled Tulu 3 405B, an advanced AI model that surpasses DeepSeek V3 and even OpenAI’s GPT-4o in certain benchmarks. The model’s open-source nature sets it apart, emphasizing the U.S.’s leadership in AI development beyond big tech corporations.

Key Points

• Ai2’s Tulu 3 405B outperforms DeepSeek V3 and OpenAI’s GPT-4o on specific AI benchmarks.

• It is fully open-source, making all components available for public use.

• Contains 405 billion parameters, trained using 256 GPUs in parallel.

• Utilizes reinforcement learning with verifiable rewards (RLVR) for better accuracy.

• Achieved top scores in knowledge-based and math problem-solving benchmarks.

• Available for testing via Ai2’s chatbot web app and on GitHub and Hugging Face.

Story The AI landscape is evolving rapidly, and Ai2’s latest release, Tulu 3 405B, is a testament to that progress. With an impressive 405 billion parameters, the model’s superior performance is attributed to reinforcement learning with verifiable rewards (RLVR), enhancing its ability to solve complex tasks with verifiable accuracy. This approach helped Tulu 3 405B achieve top scores on PopQA, a knowledge-based benchmark, and GSM8K, which tests grade school-level math proficiency.

Unlike proprietary models such as OpenAI’s GPT-4o, Ai2’s Tulu 3 405B is fully open-source, allowing researchers and developers worldwide to access, modify, and improve upon it. This move reinforces the U.S.’s competitive edge in AI research, demonstrating that innovation doesn’t have to be confined to big tech firms.

Conclusion Ai2’s release of Tulu 3 405B marks a significant milestone in AI development. By outperforming major models like DeepSeek V3 and GPT-4o while remaining open-source, Ai2 showcases the potential for cutting-edge AI to thrive outside corporate walls. With its high benchmark scores and accessibility, Tulu 3 405B sets a new standard for open AI research and development.

The #1 AI Meeting Assistant

Typing manual meeting notes drains your energy. Let AI handle the tedious work so you can focus on the important stuff.

Fellow is the AI meeting assistant that:

✔️ Auto-joins your Zoom, Google Meet, and Teams calls to take notes for you.
✔️ Tracks action items and decisions so nothing falls through the cracks.
✔️ Answers questions about meetings and searches through your transcripts, like ChatGPT

Try Fellow today and get unlimited AI meeting notes for 30 days.

Get Started Free

The Future of AI: Key Takeaways from Microsoft and Meta Earnings

Quick Summary: Microsoft, Meta, and IBM reported strong AI-driven growth in their latest earnings. Despite concerns about competition from Chinese startup DeepSeek, executives reaffirmed their confidence in aggressive AI investments, highlighting the sector's growing economic impact.

Key Points:

• Microsoft and Meta are investing nearly $150 billion in AI infrastructure in 2025.

• Meta AI reached 700 million active users, and its AI-powered marketing service grew 70%.

• Microsoft’s AI revenue hit a $13 billion run rate for the quarter.

• DeepSeek’s cost-efficient AI model caused market concerns but was praised by tech leaders.

• SoftBank is in talks to invest $25 billion in OpenAI, reinforcing U.S. dominance in AI.

Story: Tech giants Microsoft and Meta kicked off 2025 with earnings reports focused on AI, signaling continued commitment to AI-driven growth. Meta’s AI-powered services saw significant user growth, while Microsoft’s AI revenue surpassed expectations. IBM also reported a strong AI performance, booking $2 billion in AI business for the quarter.

Despite Wall Street concerns over high AI infrastructure costs and the emergence of DeepSeek’s efficient AI model, both Microsoft and Meta stood by their investment plans. Microsoft has already spent $30.7 billion in AI infrastructure this fiscal year, while Meta’s investment could total $65 billion. Executives believe these investments will strengthen their market positions over time.

DeepSeek’s disruptive AI model drew market attention, but industry leaders saw it as a positive force rather than a threat. Microsoft CEO Satya Nadella emphasized AI efficiency gains as a long-term trend, while Meta’s Mark Zuckerberg viewed DeepSeek’s innovations as validation of open-source AI development.

Conclusion: Microsoft, Meta, and other major players remain bullish on AI, despite concerns over costs and competition. With continued investments and growing adoption, AI’s role in driving revenue and innovation remains undisputed. As new players like DeepSeek emerge, the race to dominate the AI landscape is far from over.

That was it for this Weeks News, We Hope this was informative and insightful as always!

We Will Start Something Special Within a Few Months.
We Will Tell you more soon!
But for now, Please refer us to other people that would like our content!
This will help us out Big Time!