Introduction

Welcome to the second edition of this fully AI-generated newsletter. Every story, summary, and insight you’re about to read has been chosen, compiled, written, and structured by AI.

This is an experiment. The goal? To push the limits of AI in content generation, refine the process, and see just how far we can take automation in curating and delivering the most relevant AI news.

Along the way, I’ll be tweaking, testing, and iterating - so expect updates, improvements, and perhaps a few surprises. Consider this a journey into the evolving role of AI in media. Let’s see how well it does and thanks for joining us!

Please note, AI can hallucinate! But we do our best to avoid that.

23/03/25 Foreword

“Hi everyone, thank you for subscribing to our newsletter, and we look forward to you being part of the journey - week 20! The results and knowledge have been incredible, and we (our AI) is going to continue pushing our updates out. Please do share your thoughts and feedback as we progress.

This week, there are no tweaks or updates - Just a thank you for listening”. - Very best, Nicolas and the System7 team.

Now, onto the latest in AI.

Welcome to this week's AI roundup! The landscape is buzzing with major model releases from OpenAI, DeepMind, and Anthropic, each pushing boundaries in different directions. While GPT-5 has finally arrived with significant performance gains, we're also seeing fascinating developments in AI-generated worlds, open-source models, and practical tools that are changing how we work. Let's dive into what matters most from the flood of recent AI news.

Major AI News Headlines

• OpenAI's GPT-5 and first open-source models since GPT-2.

• DeepMind's Genie 3 interactive world generation.

• AI adoption reaching government and executive levels.

Major AI News

OpenAI Launches GPT-5 with Unified System and Significant Performance Gains

Summary

GPT-5 represents a shift toward AI that adapts its reasoning depth to the task at hand, making it more versatile for real-world applications. With a 256,000-token context window and better reasoning capabilities, it enables more complex projects from coding to document analysis.

Why It Matters

These announcements solidify Nvidia's leadership in AI infrastructure while expanding into robotics, a field that has long struggled to deliver on its promise. The open-source approach to GROOT N1 could accelerate development across the industry, potentially bringing capable humanoid robots into commercial settings much sooner than expected.

Key Takeaway

Free users can access GPT-5 with basic functionality, while paid subscribers get higher limits, The model includes customizable 'personalities' like Cynic or Robot for different use cases, Despite impressive gains, GPT-5 still falls short on some benchmarks compared to models like Grok 4.

Source: Link

OpenAI Releases First Open-Weight Models Since GPT-2

Summary

In a surprise move, OpenAI has released two open-source models, gpt-oss-120b and gpt-oss-20b, under Apache 2.0 licensing. The smaller model can run on standard laptops or even high-end phones, while both offer adjustable reasoning levels, web browsing capabilities, and Python execution.

Why It Matters

This marks a significant shift in OpenAI's approach, making powerful AI more accessible to developers who prefer not to rely on API-based services. It could accelerate innovation while addressing concerns about Chinese open-source alternatives in enterprise settings.

Key Takeaway

The models perform similarly to earlier OpenAI versions but at lower cost, Users can control reasoning depth with simple prompt modifiers, Sam Altman justified the release by stating positive applications will outweigh potential misuse.

Source: Link

DeepMind's Genie 3 Creates Interactive 3D Worlds from Text

Summary

Google DeepMind has unveiled Genie 3, a foundation world model capable of generating interactive, multi-minute 3D simulations at 720p and 24 fps. Users can modify these virtual environments on the fly with text commands, changing elements like weather or introducing new characters while maintaining physical consistency.

Why It Matters

This technology represents a leap toward creating realistic virtual environments that respond to human input. It could revolutionize AI training by enabling embodied agent learning in simulated spaces, with potential applications in gaming, simulation, and testing.

Key Takeaway

Genie 3 tracks previous actions to maintain world coherence over time, The model is currently available to researchers in a limited preview, While promising, it still has limitations in multi-agent interactions and long-duration simulations.

Source: Link

Fun AI Topics Of The Week!

US Government Clears Major AI Vendors for Federal Use

Summary

The U.S. government has approved OpenAI, Google, and Anthropic as official AI vendors, enabling federal agencies to purchase tools like ChatGPT, Gemini, and Claude through a pre-approved process. OpenAI is offering its frontier models to agencies for just $1 over the next year.

Why It’s Interesting!

This streamlined process removes lengthy contract negotiations and signals a significant shift in how government will adopt AI for various functions, from fraud detection to administrative tasks and public-facing chatbots.

Key Takeaway

The move follows recent executive orders emphasizing unbiased and transparent AI use, Agencies will need more knowledgeable staff to effectively integrate these tools, Competition is intensifying among AI providers for government contracts.

Source: Link

Claude Improves Security Reviews and Gets Less People-Pleasing

Summary

Anthropic has updated Claude with two major workplace improvements: automated security code scanning and enhanced personality settings. The security review feature can identify vulnerabilities like SQL injection and risky dependencies with a simple command, while personality updates make Claude more willing to provide critical feedback.

Why It’s Interesting!

The security scanning feature has already proven its worth internally at Anthropic, where it flagged a remote code execution vulnerability. Meanwhile, the personality shift addresses a common criticism of AI assistants being too agreeable.

Key Takeaway

Security reviews can be triggered with a '/security-review' command, Claude is now more likely to suggest professional help when appropriate, The updated model is more willing to break character in roleplays when needed.

Source: Link

ElevenLabs Ventures into AI Music Generation

Summary

ElevenLabs has launched a full-track music generation tool that offers control over both sound and lyrics. The system is backed by licensing deals that give artists the choice to opt in for AI usage, addressing a key concern in AI-generated creative content.

Why It’s Interesting!

As AI expands into creative domains, ElevenLabs' approach of securing proper licensing and giving artists control over how their work influences AI could provide a template for responsible development in the creative AI space.

Key Takeaway

Users can specify genre, style, and lyrics for custom tracks, The opt-in model for artists represents a more collaborative approach to AI creativity, This development signals AI's continued expansion beyond text and images.

Source: Link

Expert Prompt Of The Week

CONTEXT

With GPT-5's release, mastering prompt techniques that control reasoning depth becomes increasingly valuable.

PROMPT

I need a detailed analysis of [topic]. Please use 'Reasoning: high' for this task, and explore multiple perspectives before reaching conclusions. Consider: [specific questions]. For each question, provide evidence-based insights and identify potential limitations in your reasoning.

EXAMPLE USE CASE

Use this when analyzing complex business decisions or research questions. The 'Reasoning: high' parameter instructs newer models like GPT-5 to explore the topic thoroughly rather than giving quick surface-level responses, making it ideal for strategic planning or evaluating competing approaches.

Trending AI Topics For The Week

Claude Opus 4.1 Advances Multi-File Code Refactoring

Summary

Anthropic has launched Claude Opus 4.1, which significantly improves multi-file code refactoring and precise corrections. It scores top marks in coding benchmarks and excels at fixing bugs without introducing new problems.

Why It’s Important

This update shows how AI models are becoming more capable at complex software engineering tasks that require maintaining consistency across multiple files, potentially transforming how developers approach large-scale coding projects.

Sweden's Prime Minister Regularly Uses AI for Decision-Making

Summary

Prime Minister Ulf Kristersson revealed that he routinely uses AI in his official work to obtain quick 'second opinions' and explore different perspectives, including those contrary to initial thinking.

Why It’s Important

This admission from a head of government underscores AI's growing presence at the highest levels of decision-making and signals broader acceptance of AI-assisted governance.

OpenAI Raises $8.3B at $300B Valuation

Summary

OpenAI secured massive funding at a $300B valuation, with annual revenue projected to reach $13B. Major investors include Blackstone, TPG, and T. Rowe Price, with Dragoneer alone contributing $2.8B.

Why It’s Important

This unprecedented valuation for an AI company reflects enormous confidence in OpenAI's future growth and the central role investors expect AI to play in the global economy.