Are your business plans AI-ready for 2025?

Hi Futurist,

As we approach 2025, it's time to refine our plans and ask ourselves: how ready are we for the AI-driven transformations ahead? Therefore, I am sending you insights, inspiration, and innovation straight to your inbox. Let’s dive into the depths of the digital future together and discover the waves of change shaping our industry.

💡 In this post, we're dipping in:

• 📣 Byte-Sized Breakthroughs: Discover why preparing your AI strategy for 2025 is critical. Explore the latest trends like custom AI copilots, productivity boosts, and the ROI of embracing AI-driven transformations.

• 🤖 Digital Toolbox: From Lindy’s seamless task automation to Bolt's in-browser coding and Buster’s plain-English data queries, these tools are set to supercharge your workflow.

• 🧐 In Case You Missed It: AI breakthroughs like Gemini’s personalized knowledgebase, Mistral’s enhanced Le Chat platform, and OpenAI’s advanced macOS integrations are changing the game across industries.

Do you have tips, feedback or ideas? Or just want to give your opinion? Feel free to share it at the bottom of this dip. That's what I'm looking for.

🎧 Listen on Spotify

No time to read? Listen to this episode of Digital Dips on Spotify and stay updated while you’re on the move. The link to the podcast is only available to subscribers. If you haven’t subscribe already, I recommend to do so.

📣 Byte-Sized Breakthroughs

Quick highlights of the latest technological developments.

Headstory: Are your business plans AI-ready for 2025?

The year is almost over. Plans are taking shape. Meetings, whiteboards, strategy decks. If you’re not planning already, now’s the time. But let me stop you there. Have you asked yourself the most critical question of all? How does AI fit into your plans? Not the AI of today, but the AI of six months from now. The AI that will define 2025. Think infinity memory. Think agents. Think copilots. Think beyond.

AI isn’t a trend; it’s a shift. It’s here, reshaping industries, workflows, and outcomes. Look around. In 2024 alone, generative AI usage skyrocketed from 55% to 75%. Companies aren’t just testing the waters anymore—they’re diving in. It’s saving sellers time, cutting doctors’ report-writing hours, and forecasting with stunning accuracy. The numbers are undeniable. But here’s the truth: this is only the beginning.

What’s next? Custom AI solutions that tackle your industry’s toughest challenges. Custom copilots tuned to your unique challenges. Agents that don’t just assist but solve. AI that’s smarter, faster, and tuned to your goals. Leaders embracing these changes are seeing ROI explode—$10.3 for every dollar invested. This isn’t hype. This is the new normal.

To thrive in 2025, you need to act now. AI must be part of your core strategy, not an afterthought. Productivity, innovation, customer experience—all amplified by AI’s potential. But building an AI-driven future is no small task. The hurdles are real. Skills shortages. The pressure to innovate. The need to embed AI into every workflow. Solutions exist. But they require foresight and investment.

Questions every leader should be asking right now:

Stop. Think. Reflect. As you prepare for 2025, these are the questions that should keep you up at night:

1. Are our 2025 plans AI-proof? You’ve built your roadmap. Have you woven AI into every level of your business strategy?

2. How are we investing in skills? AI tools are useless without the people who know how to wield them. Are your teams prepared?

3. How will AI drive value in your business? Beyond productivity, how can AI improve customer experiences, innovate products, or reduce costs?

4. Are you building or buying? Should you invest in custom AI solutions tailored to your business, or optimize off-the-shelf tools?

5. How agile is your organization? Can you implement AI fast enough to capitalize on its ROI — or will you be left behind by more nimble competitors?

Here’s the truth: the businesses thriving in 2025 will be the ones planning today. The organizations asking the right questions. The leaders embracing AI not as a tool, but as a cornerstone of strategy. This isn’t just an opportunity. It’s a necessity.

In the next editions of Digital Dips, I’ll break it all down. AI isn’t just about tech. It’s about transformation. How to spot AI opportunities. How to embed them into your organization. How to stay ahead in a world that’s changing faster than ever. Let’s make sure you’re ready before the year ends, and your plans for 2025 take off.

The AI slowdown: what comes next?

TL;DR

Across the AI industry, companies like OpenAI, Google, and Anthropic are grappling with diminishing returns from scaling large models. As performance plateaus, the focus is shifting to new strategies like inference-time compute and reasoning-based improvements. This paradigm shift, exemplified by OpenAI’s o1 model and echoed in other projects, has sparked widespread discussion over the past two weeks, becoming one of the hottest topics in AI development.

Read it yourself?

Sentiment

Initial reactions were mixed but leaned towards optimism. Many felt their concerns had been validated—that the pace of groundbreaking AI innovations has slowed over the past year. This news seemed to confirm that pre-training has hit a ceiling, and such leaps as seen from GPT-3.5 to 4 are now harder to achieve. However, after reflection, consensus emerged: while pre-training may no longer deliver exponential gains, inference-time compute and other techniques represent fresh avenues for substantial progress. There’s no wall, just a different path forward.

My thoughts

The scaling law isn’t a law in the scientific sense—it’s an observation, much like Moore’s Law. And just as Moore’s Law guided technological growth for decades, the scaling law for AI will likely hold for years to come. What excites me most is the shift to inference-time compute, a new paradigm that allows models to "think" longer during response generation. By investing computational resources here, OpenAI has unlocked an entirely new scaling potential, far beyond pre-training. Even if progress on large language models (LLMs) were to stop today (and it hasn’t), the possibilities for building on existing models could fuel innovation for at least a decade. Altman’s statement resonates deeply: “There is no wall.” I firmly believe that.

China’s Qwen and DeepSeek AI models redefine open-source innovation

TL;DR

China is making waves in AI with two groundbreaking open-source models: Alibaba’s Qwen2.5-Coder and DeepSeek-R1-Lite. Qwen matches the performance of leading closed-source models like GPT-4o and Claude 3.5, while DeepSeek impresses with reasoning capabilities comparable to OpenAI’s o1. This release has intensified discussions about China’s rapid progress in AI development

Read it yourself?

Sentiment

The reaction has been one of amazement. Many are almost in disbelief at how quickly China’s open-source models have risen to match or even surpass Western competitors in certain areas. The sheer quality and speed of development have left many in awe, with some suggesting this marks a turning point in the global AI race. For years, the narrative was that China lagged behind the West, but models like Qwen challenge that assumption.

My thoughts

Whenever I give talks about AI, people often ask, “How is China doing?” My answer was always the same: “They’re two years behind America, and it’s unlikely they’ll catch up anytime soon.” Well, here we are. The release of Qwen2.5-Coder proves how wrong that assumption was. The sheer number of tokens it’s trained on—23.5 trillion—is staggering, far surpassing many Western counterparts. The addition of open-source reasoning models performing on par with OpenAI’s o1 is even more impressive. Meanwhile, U.S. companies haven’t yet delivered a definitive response to this development. It’s no surprise that Sam Altman recently proposed a "blueprint for U.S. AI infrastructure" to the government, likening this AI race to the Manhattan Project. If the past week’s developments are any indication, the competition is heating up—and it’s no longer a one-sided race.

More byte-sized breakthroughs:

• Gemini adds knowledgebase for smarter Gems
  Gemini Advanced subscribers can now upload up to 10 files to enhance their Gems with more personalized insights. The new knowledgebase feature allows users to reference files or Google Drive documents while building custom Gems. With smarter responses and the ability to edit, view, or delete shared data, Gemini just became your most adaptable AI assistant yet.

• Mistral AI introduces Pixtral Large and enhanced Le Chat
  French company Mistral AI is stepping up in the AI race with new updates to its Le Chat platform. Users can now edit and create on the new Canvas feature, search the web, explore smarter document processing, and even generate images—all for free. Le Chat positions itself as a strong competitor to OpenAI’s ChatGPT and Anthropic's Claude. With these upgrades, Mistral AI is making powerful tools more accessible for everyday users.

• ChatGPT for macOS gets smarter with apps

  OpenAI’s ChatGPT app for macOS now integrates with coding tools like Xcode and VS Code, providing smarter, context-aware support. Plus and Team users can connect ChatGPT to their apps for streamlined workflows and enhanced productivity. From reviewing code to offering tailored suggestions, this feature simplifies your daily coding tasks. More app integrations are on the way, making desktop AI more useful than ever.

🤖 Digital Toolbox

Three (new) tools that you can use for your work now or later.

Lindy AI - Your AI assistant for seamless automation

Lindy.ai simplifies your workflow by integrating with over 3,000 apps, automating tasks from email management to scheduling. Customize your AI assistant to handle repetitive tasks, allowing you to focus on what truly matters. No coding required—set up in minutes and start saving hours each week.

Bolt - AI-powered web development in your browser

Bolt.new enables you to prompt, run, edit, and deploy full-stack web applications directly from your browser—no local setup required. Leverage AI to generate code structures based on natural language prompts, and deploy your applications with a single click. Skip the complexity and focus on building smarter and faster.

Buster - Query your data with plain English

Buster.so empowers you to explore company data and build reports effortlessly using AI. Ask questions in natural language and receive data-driven insights instantly. Customize and drill down into information without technical expertise, and create shareable, beautiful dashboards in no time.

🧐 In Case You Missed It

A roundup of recent key updates.

• Vidu-1.5 redefines video creation as the first multimodal model, integrating people, objects, and environments effortlessly.

• Gemini latest update delivers top coding performance, advanced reasoning, and improved visual understanding—now recognized as the best model.

• Gemini app launches on iPhone. Experience Gemini Live with 10 voices in 13 languages for interviews, brainstorming, and more.

• YouTube introduces AI-powered song remixing. Creators can now transform tracks with simple prompts for unique, custom audio content.

• YouTube debuts gifting for vertical livestreams. This new feature takes on TikTok, allowing fans to support creators directly during live streams.

• Freepik Tunes, Custom Characters and Video are here! Animate images, generate AI voices, and create royalty-free music with custom styles.

• ElevenLabs revamps Projects! Transform books into audiobooks, scripts into voiceovers, or text into rich, multi-speaker dialogue effortlessly.

• ElevenLabs introduces conversational AI agents. Create custom AI-driven agents with natural, dynamic dialogues.

• Screaming Frog SEO Spider 21 introduces Direct AI API integration! Analyse and optimise with AI-powered insights directly within your SEO workflows.

• HeyGen drops major updates! Lifelike avatars, text-to-video prompts, emotion-responsive expressions, plus a new API and iPhone app.

• Chinese researchers used GPT-4 to simulate voter behavior for the 2024 U.S. elections—and the results were strikingly close to the actual outcome.

• ByteDance introduces SeedEdit. An image model that uses text prompts for high-resolution edits, local changes, transformations, and style adjustments.

• Anthropic now supports Google Docs integration. Add content directly to chats and projects by pasting links or selecting recent documents.

• Amazon launches its Haul Store to compete with Temu and Shein. It offers discounted products across categories, making budget shopping easier than ever.

• Krea AI launches its new trainer! Now you can train your own image generation model, enhanced with upgraded quality and hundreds of community styles.

• The EU releases the first draft of the General-Purpose AI Code of Practice. It sets the stage for clearer objectives, measures, and compliance pathways.

• NEO is the first autonomous machine learning engineer. It’s a multi-agent system that automates the entire ML workflow.

• Coca-Cola taps AI for their new Christmas ad—marking a major milestone for generative AI in creative storytelling.

• Twenty is building an open-source Salesforce alternative. Designed for flexibility and customization, it gives businesses full control over their CRM.

• Spotify launches text-to-playlist in the U.S. Generate custom playlists from detailed prompts and even include tracks based on your listening history.

• Make introduces Make Grid! A new tool for building and visualizing workflows, offering unmatched clarity and flexibility for automating tasks with ease.

• OpenAI expands Advanced Voice Mode. Now available on the web for all paid users at ChatGPT(.)com.

• OpenAI 4o gets an upgrade. Enhanced creative writing and improved file-handling deliver more engaging, insightful, and thorough responses.

• OpenAI’s "Operator" launches in January. Similar to Claude’s "Computer Use," it performs tasks like coding or booking travel independently via a browser.

• Microsoft Teams to introduce voice cloning. Soon, attendees can create AI voice replicas for meetings, adding a new level of flexibility and personalization.

• Microsoft introduces Copilot Actions! New agents and tools empower teams to work smarter and faster with AI-driven assistance.

• GitHub unveils Copilot Workspace at Ignite! The first agentic IDE, Workspace features five AI agents—Brainstorm, Spec, Plan, Implement, and Build/Repair.

• Fireworks f1 is the first reasoning system to outperform GPT-4o and Claude 3.5 Sonnet on tough coding, chat, and math benchmarks.

• Suno launches v4. Create songs with sharper lyrics, dynamic structures, and enhanced audio. Explore remastering, creative lyrics, and custom cover art.

• Instagram to add algorithm reset. Soon, users can refresh their recommendation feed, giving them more control over what they see.

• Runway released Gen-3 Alpha updates: 20-second video-to-video outputs, vertical aspect ratios for Act-One, expand video, and advanced camera control.

• Stripe launches an AI Agent SDK. Now, GPT-4o and other LLMs can call Stripe’s APIs for payments, billing, and more.

• Perplexity Shopping is here. Research and buy products effortlessly, blending answers with seamless actions—making online shopping 10x easier and fun.

• FLUX.1 Tools by BlackForestLabs now live. Advanced inpainting, structural guidance, and image restyling features elevate text-to-image editing.

👋 Let's Connect!

This was it. Our twenty fourth digital dip together. It might seem like a lot, but remember; this wasn't even everything that happened in the past few weeks. This was just a fraction.

I know spotting the right AI opportunities can be daunting. The questions, the choices—it’s a lot to navigate. That’s why I’m here to help. In the next edition, I’ll break it all down, guiding you to find and harness the potential of AI for your business. In the meantime, feel free to reach out via email or connect with me on LinkedIn—I’d love to hear your thoughts and ideas.

Looking forward to what tomorrow brings! ▽

-Wesley