Welcome to Lead with AI, the only executive AI brief for busy leaders. Every edition, I deliver the latest AI updates through real-world insights and discussions from our community of 150+ forward-thinking executives.

Reader, what if every student—and every professional—had their own AI tutor or coach?

OpenAI’s Shipmas has been buzzing with feature-packed updates, and one that truly stood out to me? ChatGPT’s Voice Mode is now paired with Vision.

Last week, I ran a training session for an education group, and we had an electric discussion imagining this future. Sam Altman once said AI could make personalized, adaptive, and always available coaches a reality. Now, his AI can combine text, voice, and visual understanding, making that vision feel closer than ever.

I also spoke with Alex Osterwalder (inventor of the business model canvas), who sees the same potential in companies. L&D could transform dramatically as these capabilities evolve over the next few years.

The real question: How do we get there—and how should we prepare?

That aside, here’s my executive brief from a blockbuster week of AI launches, with news and inspiration to help you apply AI in your work, team, and organization:

• “Your AI Team” Updates: OpenAI’s Day 5 to 8 Updates, NotebookLM Upgrade & Plus Tier, Claude Latest Model, and Gemini New Agentic Tools.
• AI Implementation Case Study: Puma Boosts Product Imagery with Google Cloud AI
• The Executive AI Brief: Ilya Sutskever: “We’ve Hit Peak Data” - What’s Next for AI?
• Weekly Quick Hits: Replit’s Powerful New Tools, Telegram’s AI Cleanup, Midjourney’s Patchwork, and more.

And other interesting sharings from our Lead with AI PRO community.

“Your AI Team” Platform Updates

ChatGPT

OpenAI dropped five practical updates last week through its advent calendar:

• ​Canvas Expansion: All users can now access Canvas, a workspace for editing text, code, and graphics with GPT-4o. It runs Python, previews charts, and debugs errors seamlessly.
• ​ChatGPT + Siri: ChatGPT is now integrated into Apple’s Siri (via Apple Intelligence) for tasks like answering questions, analyzing images, and querying documents on iPhone, iPad, and Mac.
• ​Advanced Voice Mode with Vision: Plus, Team, and Pro users can combine voice, image, and video inputs on mobile. Show your surroundings, analyze photos, get recipes, and use Screenshare for troubleshooting.
• ​'Projects' Feature: Organize chats, files, and instructions with “Projects.” It remembers context, reduces repetitive prompts, and enables Canvas.
• ​'Search' Updates: Faster results, voice queries in Advanced Voice Mode, and integrated maps for location-based searches—now available globally, including free-tier accounts.

​

NotebookLM

Google upgraded NotebookLM with:

• Interactive Audio Overviews: Ask ‘AI hosts’ questions mid-listen, like having a real-time tutor guiding your content.
• Unified Interface: Manage sources, chat, and outputs in a sleek, three-panel layout.

Google also announced NotebookLM Plus, a premium tier launching in early 2025 via Google One AI Premium, offering 5x capacity, team notebooks, and enterprise privacy.

​

Claude

Anthropic’s fastest and most cost-effective model, Claude 3.5 Haiku, is now on web and mobile. It excels at coding, data analysis, and real-time tasks with a 200K-token capacity but still lacks live web browsing.

​

Gemini

Google unveiled major model updates:

• ​Deep Research: Automates multi-step web research for organized, source-linked reports. It’s available in Gemini Advanced and is suggested for competitor analysis, academic work, or strategic benchmarking.
• ​Gemini 2.0 Flash: An early preview of Google’s fastest, chat-optimized AI model that delivers twice the speed and advanced multimodal capabilities for analysts. (VentureBeat review)
• ​AI Code Agent "Jules," powered by Gemini 2.0, helps fix bugs in Python and JavaScript code on GitHub. The new tool from Google differentiates itself by actually planning and carrying out multi-step solutions, updating multiple files at once, and preparing the final changes for review.

Google also showcased two experimental AI agents for everyday tasks:

• ​Project Astra: A virtual assistant with real-time multimodal capabilities, short-term memory, and integration into Search, Lens, and Maps.
• ​Project Mariner: A browser-based tool for hands-free workflows with tasks like building shopping lists.

This is What the 1% of Leaders Do Differently

Elite leaders have at least 12 AI assistants. Imagine how much more you’d achieve if an army of AI employees worked alongside you 24/7.

Lead with AI Cohort 6 kicks off this January, perfectly timed to equip you with the AI expertise, tools, and workflows you need to stay ahead in 2025. (Enroll with the special offer here.)

Master the complete AI leadership landscape in just three weeks, from advanced LLM applications to implementing organization-wide AI strategies.

Our program uniquely covers every aspect of AI leadership, ensuring you're fully prepared to drive AI transformation in your organization.

Special Early Bird Offer: Save $100 When You Register in the Next 24 Hours 💰

Only in Cohort 6 do we have Sodexo Head of Future of Work, Henrik Karlesjog, do a Masterclass in building AI Teams.

With 15 on-demand daily lessons, an online peer study group, three live sessions, a certificate, lifetime access to all materials, and 12 months of access to the 120 executives in our Lead with AI PRO community (a $348 value) – there was never more value.

How Companies Implement AI

No matter your company size, if you deal with product photos, take a page from Puma's book.

Puma partnered with Google for a built-in AI tool to create dynamic, localized product images on their website. A single click generates AI-powered visuals, which can be easily customized for local markets—like featuring Mt. Fuji for Japan.

As a result, Puma saves time, delivers high-quality visuals at scale, and improves the shopping experience. Puma India, for example, saw a 10% increase in click-through rates.

​>> Watch how they do it here.​

The AI Executive Brief

“We’ve Hit Peak Data” - What’s Next for AI?

At NeurIPS 2024, OpenAI Co-Founder and Former Chief Scientist Ilya Sutskever made a bold claim: “We’ve achieved peak data, and there’ll be no more.” In short, the internet’s finite training data is running out, forcing a major shift in AI development.

Sutskever predicts AI will evolve in three key ways:

• Agentic AI: AI that acts independently, anticipating tasks, adapting to workflows, and solving problems without constant direction.
• Synthetic Data: Artificially generated data will enable more precise simulations and specialized training.
• True Reasoning: AI will develop logical, step-by-step thinking, minimizing errors and delivering more reliable results.

​>> Watch TheAIGRID’s breakdown of Sutskever’s predictions here.​

Quick Hits

• Replit shipped a new Agent for no-code software creation and an Assistant for seamless project refinements, making development faster, easier, and fully accessible. (Read and watch demos here)
• Telegram removed 15.4 million suspect groups and channels in 2024 using advanced AI moderation after unprecedented pressure following founder Pavel Durov’s arrest in France.
• Midjourney is launching Patchwork - think Figma, but for collaborative ‘worldbuilding.’ It enables real-time storytelling, character creation, and interactive canvases for up to 100 users. (Try it here with tutorials here)
• Billionaire entrepreneur Mark Cuban says AI will disrupt jobs with simple decisions but will require humans to guide it in roles that demand critical thinking.

From The Lead with AI Community

Every day, Lead with AI PRO members discuss practical ways to benefit from AI in their work and organizations. This week's highlights include:

• We had the first lively demo call for our Lead with AI PRO community with SocialSonic, Leonardo, and Asana AI.
• ChatGPT launched Santa Voice for this Christmas, and Joe turned it into Santa’s workshop on hybrid working. His recipe: record with Loom, extract audio, split voices in Audacity, upload to Hedra for Santa’s AI voice, and combine everything back—done in 20 minutes.
• I shared KeyReply CEO Peiru Teo's post on what they did to foster an AI-driven culture.
• Pawel made this 50-second explainer video in just a few hours using ChatGPT (for the script), Veed (for AI voiceover, subtitles, phone frame visuals, and background music), SORA (for customer interaction footage), plus existing customer videos and platform screenshots.

If you made it this far, reply and let me know what was most valuable to you. (For example, are you setting up your AI agents yet?)

Also, please forward this newsletter to a colleague and ask them to subscribe.

If you have any other questions or feedback, just reply or inbox me.

See you next week,

​Daan van Rossum​ Host, Lead with AI

Ready to Learn AI with Us?

Check out our self-study online courses and live workshops.

Join over 120 graduates from companies like Apple, Gartner, Amazon, L'Oreal, Toyota, and more:)