We need to talk about o1 pro (just released!).
Also, CIO CISO X Germany confirmed, Gemini video applications, Elevenlabs great announcement, and DEMO. 🚀

Last night, I flew from Munich to Seoul, a 36-hour trek door-to-door, soon to host a workshop on building custom AI agents.

When I landed in Incheon Airport, I learned about what might be the biggest AI news yet in 2024. OpenAI had announced a “12-day tour” of daily releases—some big, some small—and now they dropped the bomb: o1 and o1 Pro.

We already know o1’s reasoning skills match top-tier PhD benchmarks in physics, math, coding, and beyond. o1 Pro pushes even further. The subscription runs $200/month (earlier rumors suggested $2,000/month).

Is it worth it? Absolutely. Its benchmark results are off the charts. Above PhD level. What do we even call that?

But, real-world capabilities are what matter.

o1 Pro makes building apps, including games, trivial. Structure your idea, and o1 Pro codes it instantly.

I used to imagine coding a game step-by-step with my kids over weeks, nurturing their logic skills and long-term motivation. Now, with ready-made early versions, that dynamic will shift. Maybe we won’t dive as deep into the code, but the trade-off might still be positive.

Coding by screenshot replication? Done.

o1 Pro converts screenshots of programs directly into working code.

Last example I’d like to show, o1 Pro’s math skills are extraordinary. We’ve never had anything quite like it.

Here’s o1 Pro solving IMO 2006 Q3—the hardest problem that year—in 6 minutes and 48 seconds.

For context: in 2006, among ~500 top under-19 math prodigies worldwide, only 28 solved it fully. They had 4.5 hours.

P.S. Obviously this is a very summarized solution.

(Source)

I’ll integrate o1 Pro into my coding workflow as much as possible.

I’m already using Cursor, Bolt, ChatGPT and others. Let’s see how to unlock o1 Pro’s full potential, especially in 2025, to deliver powerful results at scale.

I’ll share most of it with you—and all of it with my premium subscribers. 🤎

I would love to see you there. 

I’ll cover:

💡 Unlocking the full potential of generative AI across industries

💡 Leveraging AI agents and automation to drive innovation

💡 Strategies for creating measurable impact 

📅 January 30, 2024

📍 Sofitel Munich Bayerpost, Germany

Free for my community to join, RSVP here:

🎫 CIO Track: https://eu1.hubs.ly/H0f74VS0

🎫 CISO Track: https://eu1.hubs.ly/H0f74Wn0

AI has sparked a wave of new programs (everyone can code fast), many built without ingenuity or lasting value—often to prove a point.

That’s why it’s refreshing to see someone create a beneficial program with a sustainable, valuable idea capable of becoming a real company.

With it, you can now upload your documents and build a conversational agent that can answer questions against that document base. You can also provide a system prompt to speak in a specific language. That is very powerful.

Step 1 - Bolt.new

I’m using Bolt.new, an AI tool to build, edit, and deploy full-stack apps—no coding required.

With clear prompts, you can iteratively refine and enhance your tool to near full functionality.

Step 2 - Cursor AI (Agent) - Make everything functional and adapt it to your needs

In step 2, I use Cursor AI Agent, the most potent co-development agent.

It is a code editor built on VS Code. Its AI agent automates tasks like code completion, terminal commands, and full app development—plan, execute, debug.

See how I adjust the webpage. Learning to work with this agent will make you unstoppable.

Yours,
Martin 🙇