How to win in the AI-powered Search era

Hi Futurist,

In the previous Dip, I mentioned after the release of Grok 3 that Claude’s new model was just days away, and GPT-4.5 was weeks away. That was two weeks ago. Little did I know. Both models have now been released. OpenAI plans to charge thousands of euros per month for its AI Agents. But that plan might already be heading for the bin—thanks to the next DeepSeek moment, this time within AI Agents. 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: AI-powered search is here, and clicks are fading. Google’s Gemini, OpenAI’s GPT Search, and Perplexity are rewriting the SEO playbook. Want to stay visible? It’s time to optimize for AI-driven answers, not just rankings.

• 🎙️ MarTech Maestros: AI researcher Andrej Karpathy shares how he uses ChatGPT—from casual chats to deep research, coding, and content creation. A must-watch if you want to level up your AI game with expert insights.

• 🧐 In Case You Missed It: From Atla’s Selene 1 and Ideogram 2a’s stunning graphics to OpenAI’s latest API updates and Amazon’s quantum computing ambitions, this roundup keeps you ahead with the newest AI tools and tech shaping the future.

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.

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.

Quick highlights of the latest technological developments.

Headstory: How to win in the AI-powered Search era

In the first Digital Dip of 2025, I predicted that the impact of AI on marketing, especially SEO, would shift from clicks to conversations. "SEO becomes Search Experience Optimization (SXO)," I wrote, pointing to Google's Gemini, OpenAI's GPT Search, Bing Copilot, and Perplexity. My message? Prepare, because the click era was fading.

Fast-forward two short months, and Google's latest moves have confirmed this transformation is well underway. Gemini 2.0, powering AI Overviews and Google's experimental AI Mode, now serves comprehensive, multimodal answers directly in search results. AI Overviews, which two months ago appeared in 25% of searches, now show up in 37%—a massive shift in how users interact with search results. No clicks required.

But Google's moves aren't isolated. Vision Match has launched in Google's Shopping, turning fashion inspiration into clickable purchases with AR beauty try-ons and virtual fitting rooms. Now your ideas become products instantly. This AI-powered leap from concept to purchase shrinks the gap between inspiration and transaction.

Yet, this progress isn't without friction. Chegg recently filed a lawsuit against Google, claiming AI-driven summaries devastate their web traffic and revenue. The implications are clear: clicks will become rarer, yet visibility more critical.

Here's the harsh truth: Many searches now end in answers, not clicks. Your SEO playbook needs a rewrite. Here's your new playbook:

1. Embrace AI intent & context
   AI doesn't just look at keywords. It decodes intent. Identify and address the true intent behind queries. Provide rich, detailed content that explains the what, why, and how, positioning your content as the definitive answer AI selects.

2. Structure to be seen
   Featured snippets are golden tickets. Craft concise, direct, structured answers. Lists, tables, bullet points—they’re AI's best friends. Schema markup isn't optional anymore. Mark your FAQs, your How-Tos, your reviews, and your articles clearly, consistently. Each schema type provides specific information to AI. AI loves clarity.

3. Authority beats everything
   AI respects authority more than any keyword or backlink could earn alone. Regularly publish authoritative, expert content. Make your credentials clear, your authors visible, your insights undeniable. Secure mentions from other authoritative sources. AI listens closely to who others listen to.

4. Technical optimization
   Ensure your website is technically optimized: prioritize speed, mobile responsiveness, and secure HTTPS connections. Experiment with emerging standards like the llms.txt file to guide AI crawlers more efficiently. A technically robust website ensures maximum visibility across AI-powered platforms.

5. Humanize your approach
   Here's a paradox: as AI advances, content needs more humanity, not less. Write for humans first, with clarity, simplicity, and ease. AI prioritizes content humans actually engage with. AI systems evaluate content based on E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness). Leverage AI tools to refine your language, ensuring it's accessible, friendly, and natural.

6. Multimodal SEO
   Optimize content not only for text but also for visual and voice search. Include detailed metadata, alt text, and schema markup for visual elements. Prepare content to answer follow-up questions to enhance conversational visibility within AI dialogues.

7. Local & Personalized Experiences
   Don't overlook local optimization. Ensure your business profiles, location data, and reviews are accurate and compelling. AI-driven local recommendations increasingly influence visibility. Personalized experiences, delivered seamlessly, will position your content favorably in personalized AI search results.

8. Track, adapt, repeat
   Measure success not just in clicks but visibility within AI overviews. Track featured snippet appearances, voice search hits, and AI-driven referrals. Use Google Analytics segments dedicated to AI tools. Adapt quickly, refine constantly. AI evolves rapidly—your strategy must too.

Here's the bottom line: AI search is the new reality. You can't out-keyword it. You must out-think it, out-structure it, out-authority it. Your SEO strategy can't be stuck in yesterday. It must live in tomorrow, today.

OpenAI introduces GPT-4.5

TL;DR

GPT-4.5 is OpenAI’s newest and most advanced chat model, focusing on improved creativity, emotional intelligence, and natural conversation. Unlike reasoning models, it’s not designed for hardcore logic or math problems. Instead, it shines in writing, brainstorming, and social interactions. Early users report that it "feels like talking to a thoughtful person," though it comes with a steep price tag.

Read it yourself?

Sentiment

The first reactions are mixed. Some users are thrilled, calling it a “clear improvement over GPT-4o,” while others expected more. One tester noted, “I had early access and tested it extensively. It’s the best non-reasoning model yet, but if you're using it for code, you're doing it wrong.” Another user ran it through personality tests and found it “more extroverted, open, and conscientious.” Others were amazed by its creativity: “I needed a mission statement for a project. Claude struggled, but GPT-4.5 nailed it instantly.”. It all depends on how you use it.

My thoughts

GPT-4.5 was never meant to compete with OpenAI’s top reasoning models, and that shows in benchmark results. But as a general-purpose AI for daily tasks, OpenAI has delivered exactly what they needed to. It’s more engaging, more emotionally intelligent, and better at creative work. Interestingly, there was a brief moment when GPT-4.5 overtook Grok as the top AI model across the board—only for Grok to reclaim its position a day later.

But the real takeaway is why OpenAI launched 4.5 at all. With a price tag of $75 for 1M input tokens and $150 for 1M output tokens, OpenAI never intended this model for widespread API use. If GPT-5 is going to be an all-in-one model, they need a version fine-tuned for everyday use. That’s what 4.5 is: they’re watching how users interact with it, figuring out what works, and using that data to shape GPT-5. The better OpenAI understands its strengths and weaknesses now, the stronger GPT-5 will be.

Anthropic introduces Claude 3.7 Sonnet

TL;DR

Claude 3.7 Sonnet is Anthropic’s smartest model yet and the first hybrid reasoning model on the market. It seamlessly switches between instant responses and step-by-step reasoning, giving users more control over speed and depth. Alongside the model, Anthropic introduced Claude Code, a command-line tool for agentic coding that lets developers delegate complex engineering tasks directly from their terminal. Early tests show that Claude 3.7 Sonnet outperforms competitors in coding, reasoning, and agentic tasks.

Read it yourself?

Sentiment

Developers are hyped. Claude 3.7 Sonnet is already topping software engineering and agentic benchmarks, even surpassing OpenAI’s o1 and o3-mini-high. Some call it “the best agentic model ever” and predict it will dominate LiveBench for coding. One tester noted, "Claude now handles full-stack updates better than any model before. It's insane." Another added, "The best coding model powering Cursor and Windsurf just got a whole lot better.". But it’s not just about raw performance. The integration of step-by-step thinking, improved web search, and enhanced coding workflows is what makes this release stand out.

My thoughts

Anthropic has packed the most exciting details at the very end. They’re not just building an LLM, they’re laying the groundwork for fully autonomous AI research agents. Their roadmap is ambitious—by 2025, they expect AI to independently conduct research for hours. By 2027, these agents could solve problems that would have taken scientists years. That makes me wonder: what else does Anthropic have behind the scenes? If they’re this confident about the future, they’re likely much further along than they’re letting on.

More byte-sized breakthroughs:

• The first AI agent that does it all—simultaneously

  Manus redefines multitasking, controlling 50 browsers at once to handle complex tasks in parallel. It edits files in real-time, deploys full websites, and generates detailed reports at record speed. Need research done? Manus outperforms competitors with lightning-fast analysis. It even manages financial transactions and purchasing autonomously. It’s shaking up the AI world. OpenAI, Claude, and others might have reason to worry.

• The future of data analysis with Gemini’s data science agent
  ​Google Colab has introduced the data science agent, powered by Gemini, to simplify data analysis. Users can now upload data and describe their analysis goals in plain language. The agent then generates a complete, functional notebook automatically. This feature eliminates tedious setup tasks, allowing data scientists to focus on insights rather than coding. Early users report faster workflows and smoother research.

• Search your company’s files and the web in one go with Perplexity
  ​Perplexity AI introduces Deep Research for Enterprise, connecting to your company’s Google Drive, OneDrive, and SharePoint. Now, businesses can search company files and the web at the same time. With enterprise-grade security, data stays protected. No more switching between platforms—find answers instantly. Available for Enterprise Pro users. A smarter way to explore and analyze information.

A must-see webinar, podcast, or article that’s too good to miss.

How an AI researcher uses ChatGPT

Large Language Models are evolving fast, but how do experts actually use them? Andrej Karpathy, one of the most respected AI researchers, takes you through his personal workflow. From basic interactions to deep research, coding, and even content generation—this video is a practical guide to getting the most out of AI. Clear, structured, and full of insights.

A roundup of updates that are too cheesy to ignore.

• Atla launches Selene 1, the most accurate LLM Judge, offering precise AI response evaluations and customizable criteria.

• Sesame unveils a preview of its expressive voice technology, moving closer to lifelike, all-day wearable voice companions.

• Ideogram introduces 2a, offering stunning graphics with text and photos in seconds and now available at half the cost.

• Luma AI’s DreamMachine's new video-to-audio feature now lets you add sound to your video creations with just a click or detailed prompts.

• Apple invests in US manufacturing with $500B commitment, unveiling a new AI server facility in Houston to support its AI endeavors.

• Apple's AI journey demands a swift turnaround, facing a potential fall behind as a fully conversational Siri AI isn't expected until 2027.

• Perplexity unveils Comet, an innovative browser for agentic search, available soon.

• Perplexity's new voice mode lets you ask any question and hear real-time answers instantly.

• China's quantum computer outpaces other supercomputers, executing tasks a quadrillion times faster than the best efforts of Google.

• Qwen launches their reasoning model Thinking (QwQ) in Qwen Chat, capable of math understanding, coding, agent, and more.

• Alibaba unveils Wan2.1, its open-source video generation suite, excels in complex motion dynamics & physics simulation & text rendering.

• Microsoft's Copilot now offers free, unlimited access to Voice (advanced voice mode) and Think Deeper (reasoning capabilities) for all users.

• Microsoft releases Phi-4 multimodal and Phi-4 mini, bringing state-of-the-art multimodal and text-based AI capabilities to developers worldwide.

• GitHub Copilot code review feature in is now publicly previewed, offering automated bug detection and performance improvement suggestions.

• Adobe launches the free Photoshop app for iPhone users, allowing seamless photo and graphic editing on the go.

• Convergence unveils Proxy Lite, a 3B parameter open source browser agent.

• Proxy unveils Deep Work, a browser mode designed to tackle complex tasks, self-correct, and execute multiple operations simultaneously

• Open AI’s Deep research is now rolling out to all ChatGPT Plus, Team, Edu, and Enterprise users.

• Open AI’s ChatGPT free users can daily preview the Advanced Voice powered by GPT-4o mini.

• OpenAI's latest focus on Safety and Alignment shapes the approach to artificial general intelligence, ensuring its benefits extend to all of humanity.

• OpenAI o1 and o3-mini are now available in the API for developers on all paid usage tiers.

• OpenAI’s ChatGPT for macOS now seamlessly integrates with IDEs, allowing Plus and Team users to directly edit code within macOS.

• OpenAI plans to charge up to $20,000 monthly for a PhD-research-agent, $10,000 for a software engineer agent and $2,000 for high-income knowledge worker agent.

• World debuts World Chat, a mini app enabling seamless chats and crypto transactions with real humans on the World Network.

• Google’s Gemini now allows users to interact with videos and on-screen content, transforming how questions are asked in real time.

• Google Gemini's new features elevate data visualization and insights directly in Google Sheets.

• Google introduces an AI-powered Gemini panel within Calendar to streamline schedule management through conversational prompts.

• Google’s Gemini Code Assist is now available for free, offering AI-powered coding assistance and code reviews in GitHub, powered by Gemini 2.0.

• Google’s Gen AI Toolbox for Databases enters public beta, offering developers agent-based generative AI tools to connect databases.

• Bolt introduces dynamic reasoning, a feature that lets the AI choose how hard to “think” when tackling complex coding problems.

• Elevenlabs launches Scribe, the world's most accurate Speech to Text model, transcribing in 99 languages with precision and ease.

• Elevenlabs is now empowering authors to create and distribute AI-generated audiobooks through its own Reader platform..

• Hume releases Octave, the first LLM built for text-to-speech. You can create any AI voice with just a simple prompt.

• Amazon launches Alexa+, operating as your AI assistant that connects LLMs, agentic capabilities and third-party services to your devices.

• Amazon is reportedly crafting its own AI 'reasoning' model, aiming to rival established players like OpenAI's o3-mini.

• Amazon launches Ocelot, its inaugural quantum computing chip, aligning with competitors Google and Microsoft in the quantum race.

• Wonder Dynamics' Motion Prediction, predicts an actor's movement even when obscured, bringing more realism to your animations.

• Meta is reportedly set to unveil a stand-alone app for Meta AI to rival OpenAI's ChatGPT and other AI chatbots.

• Instagram considers creating a separate app for Reels to increase the pressure on TikTok.

• Pika released 2.2, delivering 10s generations, crisp 1080p resolution, and innovative Pikaframes with flexible 1-10s key frame transitions.

• LangGraph introduces prebuilt agents for streamlined AI development in Python and JavaScript.

• Udio unveils a teaser of their latest v2 model with a captivating preview. Sound on!

• Synthesia introduces dynamic captions, seamlessly syncing with your script and offering customizable styles, available in all plans.

• Cloudflare introduces Agents that allows developers to build and deploy AI agents with persistent states and real-time interaction capabilities.

• Deutsche Telekom partners with Perplexity to launch a low-cost AI Phone under $1K, targeting the European market in 2026.

• Opera unveils Browser Operator, an AI agent that simplifies your online tasks by understanding natural language commands.

• You launched a research agent that reads and analyzes up to 400 sources, including company data, creating reports in under 5 minutes.

• Hedra Studio introduces AI video creation with Character-3, the first omnimodal model for content generation across image, text, and audio.

• Podcastle introduces Asyncflow v1.0, a text-to-speech model offering over 450 AI voices, now with an API for easy app integration.

• Mistral introduces its groundbreaking OCR API, setting a new standard in document understanding.

• Freepik's Colors unleashes vibrant creativity by generating AI images with your custom or pre-made, on-brand, color palettes.

This was it. Our thirty-first 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.

The impact of AI on SEO is just one of the many areas AI is influencing—or will influence. The key is not to react when it happens but to stay ahead of it. And that’s where I want to help you. Whether it’s on an operational, tactical, or strategic level. Reply to this email with your question, and I’ll help you out.

Looking forward to what tomorrow brings! ▽

-Wesley