What Superhuman Problems Can You Solve with AI?

Friends,

Welcome back to The AI Optimist - your weekly dose of calm clarity in the face of big shifts and bigger narratives.

This week, two things stood out:

🎙️ OpenAI quietly rolled out built-in meeting transcription, consolidating another layer of software into their ecosystem. A signal for anyone building wrap-around tools: the platform is coming for your features.

🔐 The Model Context Protocol (MCP), which caused quite a stir over security concerns, has just had its security model upgraded (read spec). That means more secure, more flexible ways to let AI do real work - faster than many are ready for.

So, here’s what matters this week:

🔥 URGENT PRIORITIES

✅ No fires to fight
✅ No system shocks
✅ Stable ground to plan from

Take a breath. The world is moving fast, but you don’t need to.

🎯 STRATEGIC CONSIDERATIONS

🧠 It’s Time to Ask: What Superhuman Problems Can You Solve?

As we approach the end of H2, here’s a fresh lens to bring to your strategic reviews:

Courtney Hohne (ex-Google X) just launched a studio focused entirely on unowned, global-scale problems. Her framing is powerful.

👉 Read her LinkedIn post
👉 And this FT piece touches on the same topics explaining why productivity isn’t enough to drive the economic opportunity.

Why this matters:
AI’s true upside won’t come from efficiency alone - it will come from reinvention. That requires headspace, not just headcount. That’s why we focus so much on helping leaders build the capacity to think differently.

🚨 Reminder: The next cohort of AI Night School starts June 26th.
If you’re a newsletter subscriber and want to join, reach out—we have a 50% discount available to help you get ahead of the curve.

🤓 GEEK OUT

💻 Claude Code Can Decompile Itself: Here’s How it’s done
Read: Yes, Claude Code can decompile itself (Credit Ian Moss for finding this article and sharing with me)

Why this matters:
This is another signal that AI is accelerating tech evolution. The implications for cybersecurity are massive. If you sell software or use legacy systems, the barrier to deep access just dropped again.

📊 AI 2027 Tracker Is Now Live - and It's 100 for 100
Read the original AI 2027 blog to understand their predictions.

Explore the Tracker 

Why this matters:
If you’re trying to understand where the world is going, this isn’t futurism, it’s forecasting. 100 predictions tracked. 100 correct (so far). First mover advantage will be far more powerful than fast follower tactics in this next era.

🎨 WEEKEND PLAYGROUND
🗣️ Try Hulme.ai – Emotional Text to Voice

Head outside, type a story, and hear it come to life with emotion.
👉 Try Hulme.ai

Why this matters:
Voice is about to become a primary interface for work—and creative expression. Testing tools like Hulme is a fun way to explore that shift. Plus, it’s one of the few AI tasks you can do on a walk with a coffee.

☕ Also: Curious (or concerned) about AI’s environmental impact?
This piece is worth reading over your next iced drink.

📅 UPCOMING EVENTS

Here’s what we’re running to help you build practical AI fluency:

🆓 Free Webinar
How to Benchmark Your Organisation’s AI Readiness
→ With Hugo
→ A practical walkthrough for leaders planning their AI-first roadmap

💼 Premium Webinar
Build Your First AI Agent
→ With Ben
→ Hands-on session—you’ll leave with a working agent built

🎓 AI Night School: New Cohort Starts June 26
→ Core curriculum for leaders, founders & teams
→ If you’re interested, reply or reach out to Ben or Hugo directly

📢 Share Your Thoughts

If The AI Optimist is helping you think more clearly, send it to someone else navigating the shift.

If it’s not quite landing, reply and tell us what’s missing. We read every response.

Stay cool,

Hugo & Ben