AI Business News #9: 2025 Trends, Predictions and Investments

Welcome to the latest edition of the MadebyAgents newsletter. In this issue:

• DeepSeek’s V3 model

• Microsoft invests $80 billion in AI

• 2025 AI predictions and what to watch

• Tree of thoughts prompting explained

• Create a slide deck with AI

• Free AI resource to get better results from LLMs

Enjoy the 9 minutes read.

Latest News

🐳 DeepSeek-V3 Released: A Game-Changer in AI

DeepSeek-V3, a groundbreaking Chinese open-source AI model, was released on December 26, 2024. With 671 billion parameters and a focus on efficiency, it rivals top-tier models like GPT-4o and Claude-3.5-Sonnet at a fraction of the cost. Its release has sparked discussions about the future of AI development and affordability.

What's New:

DeepSeek-V3 introduces several technical innovations. It uses a Mixture-of-Experts (MoE) architecture, activating only 37 billion parameters per task, reducing computational costs. The model incorporates Multi-head Latent Attention (MLA) to cut memory usage by 93.3% and FP8 precision training, lowering hardware demands by 30%. These advancements enable faster inference speeds—60 tokens per second, three times faster than its predecessor.

In benchmarks, DeepSeek-V3 excels in math, coding, and long-text tasks, outperforming competitors like Llama-3.1-405B and Qwen2.5-72B. It also supports 128K context windows, making it ideal for complex applications.

Business Impact:

DeepSeek-V3 further democratizes AI access. At just $0.1 per million input tokens, it’s 70 times cheaper than GPT-4o. This affordability empowers small businesses and startups to integrate advanced AI without heavy investments.

For more details, visit DeepSeek’s official website or explore the GitHub repository.

📈 Microsoft on Track to Invest $80 Billion in AI Data Centers

Microsoft’s president Brad Smith outlines a vision for American AI leadership, emphasizing three key pillars: advancing AI technology and infrastructure, expanding AI skilling programs, and promoting AI exports globally. He highlights AI as a transformative force akin to electricity, with the potential to drive economic growth, create jobs, and strengthen U.S. competitiveness. Microsoft is investing heavily in AI infrastructure and skilling initiatives, aiming to position the U.S. as a global AI leader while countering China’s growing influence in the sector.

What’s New:

He underscores the rapid evolution of AI as a General-Purpose Technology (GPT) with the potential to revolutionize industries. Microsoft is investing $80 billion in AI-enabled data centers, with over half of that investment in the U.S. The company is also launching large-scale skilling programs, aiming to train 2.5 million Americans in AI by 2025. Additionally, Microsoft is expanding its global AI infrastructure, partnering with international allies to counter China’s AI advancements. These efforts reflect a broader push to integrate AI into education, workforce development, and global markets.

Business Impact:

For small businesses, startups, and technical professionals, AI presents both opportunities and challenges. AI can level the playing field, enabling smaller companies to compete with larger enterprises by automating tasks and enhancing productivity. However, businesses must adapt by investing in AI skilling and integrating AI tools into their operations. Microsoft’s initiatives, such as AI training programs and partnerships with community colleges, aim to make AI accessible to all. For the U.S. economy, AI promises to drive innovation, create jobs, and boost global competitiveness, but success depends on collaboration between the private sector, government, and educational institutions.

AI Investments - What to Watch¹

🔮 AI Outlook 2025: Key Predictions

2025 will be a decisive year for artificial intelligence. AI will become ubiquitous, influencing almost every aspect of life. Here are the most important developments you should keep an eye on:

• AI Goes Mainstream: AI advertising reaches its peak before a backlash sets in. Companies must adapt to AI but also emphasize human values.

• "AI-Free" as a Premium: Products made by humans gain value. Authenticity and human imperfection become selling points.

• Faster Prototyping: AI accelerates the development of software prototypes. Ideas can be tested quickly. This lowers the cost of software development.

• Personalization is Crucial: AI applications are personalized through a better understanding of human psychology. Understand your customers.

• AI Agents Take Over: AI agents will perform tasks independently. Orchestrating these agents becomes critical. Middle management will transform into agent supervisors.

• Voice AI Gains Importance: Voice AI will be the next big innovation after smartphones. Voice-controlled interaction will gain importance.

• Content and Video Generation: The creation of content and videos will be transformed by AI. Authenticity will become more important. Verification will switch from blue checks to "human-generated" badges.

• Software Development Changes: AI writes code, so design and product understanding become more important. Product managers and designers gain importance.

• AI-Driven Companies Emerge: AI permeates supply chains and product design. Adaptation is crucial for business success.

• New Forms of Companies: Individuals or small teams can generate high revenues with AI agents. The definition of "company" will change. Investments shift towards AI agent networks.

• AI Acts in the Background: Instead of chatting with AI, agents will work for us. AI supports decisions in the background.

• Focus on Humans: AI systems should be oriented towards pro-social values. Collaboration and understanding become more important.

What Does This Mean For You?

For small business owners, startup founders, and tech professionals, this means:

• Expand AI Knowledge: Stay up-to-date with the latest developments.

• Maintain Flexibility: Be ready to adapt to new AI technologies.

• Understand Customers: Understand how AI can improve customer experiences.

• Act Ethically: Develop AI responsibly.

• Use Creativity: Find ways that AI can inspire your creativity.

• Develop the Right Skills: Focus on orchestrating AI.

🇦🇷 Nuclear Energy and AI - Argentina Leading South America

Argentina is seeing a resurgence in nuclear technology and aims to become an AI development hub. Cheap energy is crucial for this.

Why Cheap Energy is Important for AI:

AI requires massive computing power, demanding significant energy. Nuclear energy offers a clean, safe, and scalable solution. Argentina views nuclear energy as vital to support its AI goals. The global demand for AI energy is expected to drive a nuclear energy resurgence.

The Role of Uranium:

Argentina possesses vast uranium reserves. These reserves are essential for fueling nuclear reactors. Argentina plans to mine and export uranium, meeting domestic demand and strengthening its international position.

Argentina in a Nutshell:

Argentina was once among the five richest countries. The country faces economic challenges and high inflation. It aims to re-establish itself as a technological leader, with current reforms intended to improve the situation.

New Nuclear Reactor Technology:

Argentina is focusing on small modular reactors (SMRs). SMRs are more flexible and cost-effective than traditional reactors. Argentina is developing this technology independently and plans to use it domestically and for export. An Argentine engineer has developed a self-contained microreactor, the N1, capable of powering a city. The N1 reactor can operate for 20 years without refueling, and is safely located underground.

Argentina's nuclear plan includes building an SMR at the Atucha site. The plan has support from the International Atomic Energy Agency (IAEA). The N1 reactor is a 17 MW pressurized water reactor (PWR) designed to supply power to 50,000 people. It uses an air-cooling condenser system which means it doesn't need natural water sources. The N1 reactor is designed to be produced in gigafactories, similar to the auto industry.

Use Case of the Week

🌳 Tree of Thoughts Prompting

Tree of Thoughts (ToT) is a novel method to better solve complex tasks with Large Language Models. ToT enhances the reasoning abilities of LLMs by allowing them to explore different solution paths.

How it works:

ToT treats a problem as a search within a tree structure. Each step in the tree is a "thought" or partial answer. The model evaluates each of these "thoughts" independently. It selects the most promising thoughts to continue the search. It can pursue multiple paths simultaneously and return to previous steps if necessary. It's like a team of experts working together on a complex task and learning from each other.

Difference from Chain of Thought (CoT):

Chain of Thought (CoT) encourages LLMs to explain their thought process step by step. This often leads to better results. However, CoT follows only a single line of reasoning. ToT can pursue multiple paths concurrently and is more flexible, as it can backtrack and examine alternatives when needed. ToT is thus better suited for more complex problems than CoT.

Example Prompt (Business Scenario):

This example prompt shows how ToT prompting can be used to get an LLM to consider different strategies, much like a team of experts in a company. This "expert discussion" can lead to a better outcome.

Tested AI Tools

💼 Revolutionizing Business Presentations with AI

Gamma is an AI-powered tool designed to simplify the creation of professional presentations, documents, and web pages. It uses advanced AI to transform your ideas into visually stunning and engaging content in minutes.

Why Use Gamma?

• Speed: Create presentations in under a minute.

• Ease: No design skills needed. AI handles formatting and layout.

• Customization: Tailor templates to match your brand.

• Collaboration: Work with teams in real-time.

• Engagement: Add interactive elements like videos, charts, and quizzes

Step-by-Step Guide to Creating Captivating Business Presentations

1. Sign Up & Log In: Visit Gamma.app, create an account, and log in.

2. Create Project: Click “Create New AI” and select “Presentation” as your project type.

3. Choose Template: Pick a template or let the AI suggest one based on your topic.

4. Input Content: Enter your topic, key ideas, and additional details (e.g., bullet points, data, or images). The AI will generate an outline and slides.

5. Customize Slides: Edit text, add visuals, adjust layouts, and apply themes/colors.

6. Add Multimedia: Insert videos, GIFs, charts, or interactive elements like polls/quizzes.

7. Collaborate: Invite team members to review, edit, and comment in real-time.

8. Export & Share: Export as PDF, PNG, or shareable link. Present live or send to stakeholders.

MadebyAgents Updates

📋 Get Better Results from ChatGPT

Just created an AI resource for you.

A 4-page LLM prompt engineering cheat sheet.

Use it to enhance your prompts and get better results from ChatGPT or other LLMs. We included the latest techniques, many templates, and examples.

That’s It for This Week

✨ Before You Go:

We’d love to hear what you think. Please share your opinion.

See you next time!

Tobias from MadebyAgents.com