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!

13/04/25 Foreword

“Hi everyone, thank you for subscribing to our newsletter and we look forward to you being part of the journey. Please do share your thoughts and feedback as we progress.

This week will be the last week without any updates! We have some cool ideas and have had some great feedback to adjust with”. - Very best, Nicolas and the System7 team.

Now, onto the latest in AI.

The AI landscape is evolving at breakneck speed, with major players making significant moves this week. Meta unveiled its Llama 4 family, featuring Scout, Maverick, and the upcoming Behemoth models, while former OpenAI CTO Mira Murati's startup is reportedly seeking an unprecedented $2B in funding. Meanwhile, OpenAI is countersuing Elon Musk amid rumors they may acquire Jony Ive's AI hardware venture. Google continues to assert itself with Gemini 2.5 Pro and new enterprise tools, as Shopify's CEO makes waves by requiring employees to try AI solutions before hiring new staff.

Major AI News Headlines

• Meta's Llama 4 models challenge industry leaders with impressive capabilities.

• AI hardware race heats up with OpenAI-Jony Ive collaboration rumors.

• Companies enforcing AI adoption policies for employees.

Major AI News

Meta Launches Llama 4 Models with Impressive Capabilities

Summary

Meta has released two models from its Llama 4 family—Scout and Maverick—with a third (Behemoth) still in development. Scout offers a 10-million token context window and can run on a single GPU, while Maverick handles multimodal tasks with 17B active parameters from a 109B parameter pool using a mixture-of-experts architecture.

Why It Matters

These open-weight models democratize access to advanced AI capabilities while requiring fewer computing resources than competitors. The efficiency comes from only activating a portion of the neural network for each task, potentially reshaping how companies deploy AI at scale.

Key Takeaway

Llama 4 models use a mixture-of-experts approach to reduce computing needs, Scout targets lightweight deployment while Maverick excels at multimodal reasoning, Meta's open approach challenges closed proprietary models from OpenAI and Anthropic.

Source: Link

Shopify Mandates AI Usage Before New Hiring

Summary

Shopify CEO Tobi Lütke issued a company-wide directive requiring teams to prove that AI cannot handle their processes before hiring additional staff. This policy positions AI as the default solution for scaling operations, with human hiring becoming the exception that needs justification.

Why It Matters

This represents a significant shift in how companies approach workforce planning, treating AI as a first-class resource rather than a supplementary tool. As more organizations adopt similar policies, we may see fundamental changes in employment patterns and skill requirements across industries.

Key Takeaway

Shopify is reimagining workflows with AI agents as core team members, Teams must demonstrate why AI solutions won't work before expanding headcount, The policy signals a broader trend of companies prioritizing AI-driven optimization.

Source: Link

Mira Murati's Startup Seeks Record $2B Funding

Summary

Thinking Machines Lab, founded by former OpenAI CTO Mira Murati, is reportedly seeking a $2 billion seed round at a $10 billion valuation. While details remain scarce, the company has assembled an impressive team including several high-profile OpenAI veterans.

Why It Matters

If successful, this would represent the largest seed funding in history, signaling extraordinary investor confidence in next-generation AI. The scale of funding and caliber of talent suggest a potential breakthrough product that could reshape the AI landscape.

Key Takeaway

The unprecedented funding target highlights growing investor enthusiasm for frontier AI, Former OpenAI executives are building new ventures that may challenge established players, Stealth AI startups can command massive valuations even before product release.

Source: Link

Fun AI Topics Of The Week!

Google's Gemini 2.5 Bridges AI Models for Real-World Problem Solving

Summary

Google has introduced Gemini 2.5, which connects various machine learning models to enable cross-domain insights. The system can tackle complex questions that require understanding multiple subjects, such as analyzing how climate change impacts public health and economic conditions in specific regions.

Why It’s Interesting!

Rather than just improving a single model, Google is creating an ecosystem where specialized models work together to solve real-world problems that span multiple domains. This approach could make AI more practical for addressing complex societal challenges.

Key Takeaway

Gemini 2.5 connects specialized AI models to tackle interdisciplinary problems, Google is focusing on practical applications rather than just benchmark performance.

Source: Link

Self-Improving AI Model Developed in China

Summary

Chinese startup DeepSeek, working with Tsinghua University, has created DeepSeek-GRM, an AI system that refines its own reasoning. The technique, called self-principled critique tuning (SPCT), equips the model with an internal AI 'judge' that evaluates and refines outputs without requiring larger model sizes.

Why It’s Interesting!

This approach shifts AI development away from simply making bigger models toward creating more efficient self-improvement mechanisms. The internal 'judge' allows the model to critique and refine its own outputs, potentially leading to better results without the computational costs of larger models.

Key Takeaway

DeepSeek-GRM uses an internal critique system to improve its reasoning, The model claims to outperform larger competitors while using fewer resources.

Source: Link

YouTube Expands AI-Generated Content Detection

Summary

YouTube is extending its test of AI-based likeness detection to prevent misuse of synthetic content. The system, built on its existing Content ID platform, flags unauthorized AI-generated replicas of creators' faces or voices, allowing them to request takedowns.

Why It’s Interesting!

As AI-generated deepfakes become more convincing, platforms are developing countermeasures to protect individuals' likeness rights. YouTube's approach leverages its existing copyright infrastructure to address this emerging challenge, potentially setting a standard for how other platforms might handle synthetic content.

Key Takeaway

YouTube's system detects AI-generated replicas of real people's faces and voices, Creators can request takedowns of unauthorized synthetic content featuring their likeness.

Source: Link

AI Tools Of The Moment

Midjourney V7 

New image generation model with 10x faster 'Draft Mode' and improved texture quality.

Link

Firebase Studio

Google's open-source tool that builds complete web applications from simple prompts.

Link

Nova Reel

Amazon's video generation model that creates 2-minute multi-shot videos.

Link

Canva AI

Transforms text prompts into polished designs adapted to your style and branding.

Link

Airtable Assistant

Build apps, manage databases, and research web data through natural conversation.

Link

Expert Prompt Of The Week

CONTEXT

With companies like Shopify requiring employees to show why AI can't handle tasks before hiring, crafting effective prompts is becoming an essential skill.

PROMPT

I need a detailed strategy to [describe specific business process]. For this task: 1) Analyze current workflow inefficiencies, 2) Recommend AI tools that could automate key steps, 3) Outline implementation steps with estimated time savings, 4) Identify which aspects still require human oversight and why, 5) Suggest metrics to evaluate success. Format your response with clear headings and bullet points for each section.

EXAMPLE USE CASE

Use this prompt when evaluating business processes that might be candidates for AI automation. It forces you to think critically about both automation potential and necessary human involvement, creating a balanced approach that maps directly to cost-benefit analysis for decision makers.

Trending AI Topics For The Week

Rumored OpenAI Acquisition of Jony Ive's Hardware Startup

Summary

OpenAI is reportedly in talks to purchase io Products, the AI hardware startup led by former Apple design chief Jony Ive and backed by OpenAI CEO Sam Altman.

Why It’s Important

This potential acquisition signals OpenAI's ambitions beyond software into consumer AI hardware, possibly creating an AI-first device that could challenge traditional smartphones and computers with a new interaction paradigm.

ChatGPT Gets Long-Term Memory

Summary

OpenAI has added persistent memory to ChatGPT, allowing it to remember details from past conversations, even those from months ago.

Why It’s Important

This feature transforms ChatGPT from a stateless tool into a more personalized assistant that builds on shared history, though it raises important privacy considerations that users will need to manage carefully.

Stanford AI Index Shows Chinese Models Nearly Matching US Performance

Summary

Stanford's latest AI report reveals that Chinese frontier AI models have nearly caught up to US models on major performance benchmarks, despite receiving far less investment.

Why It’s Important

The narrowing performance gap suggests the global AI race is intensifying, with implications for technological leadership, national security, and economic competitiveness.