In 60 seconds:
- OpenAI shipped GPT-5.5 (Wed) — agentic, hallucinations down 60%.
- Claude Opus 4.7 went GA (April 16), top of the coding benchmarks for ~48 hours.
- Claude Design launched (April 17) — describe → polished decks, prototypes, PDFs.
- SpaceX/xAI locked an option to buy Cursor for $60B.
- Claude Mythos stays gated — too capable in cybersecurity to release publicly.
I read way too much AI news and play with too many AI tools. Most of it is noise. Every Friday I’ll share what actually matters and more importantly, what it means for you as a leader, with more emojis than you can imagine 😂
This week was insane. Here’s what you need to know.
🤖 GPT-5.5 dropped (Wednesday)
OpenAI’s latest beast. Faster, smarter, built for agentic work. It can operate software, research the web, and write code on its own. Hallucinations down 60%. And better than all other models in all sorts of benchmarks.
👊🏻 So what: If your team isn’t using AI for coding and research yet, you’re not one generation behind. You’re two.
🧠 Claude Opus 4.7 went GA (April 16)
Anthropic’s answer. Leads on the hardest coding benchmarks. Better vision. Handles long, complex tasks with a consistency that feels genuinely new. But was overtaken within a few days by GPT-5.5.
👊🏻 So what: OpenAI vs Anthropic is neck and neck. No single winner. But the real question for execs isn’t “which model is best.” It’s “are we using ANY of them seriously?”
🎨 Claude Design launched (April 17)
This one’s a sleeper. Describe what you need, Claude builds it, refine in conversation, export to Canva, PDF, PPTX, or hand off to devs. Polished designs. Prototypes. Slides. No designer needed.
👊🏻 So what: Every exec who’s ever waited 2 weeks for a deck from the design team just got their time back. Been using it all week. It’s real.
🚀 SpaceX/xAI locks option to buy Cursor for $60B (Monday)
Not acquired yet, but SpaceX (merged with xAI in Feb) secured the right to buy the most popular AI coding tool on the planet. Microsoft was also looking. They lost.
👊🏻 So what: AI coding tools are now valued like defense contractors. That tells you everything about where software is headed.
🔒 Claude Mythos stays gated
Anthropic’s most powerful model. So good at cybersecurity that they won’t release it publicly. Gated access for defensive security only.
👊🏻 So what: We’ve entered the era where AI companies build things too powerful to ship. That’s a sentence worth sitting with.
🏃🏼♂️💨 The pace is relentless. Faster than a sub 40 min 10km. But the gap isn’t between people who read the news and people who don’t. It’s between people who take the time to USE these tools and people who just watch.
Bottom line for leaders: Pick one of these models (GPT-5.5, Claude Opus 4.7) and personally run it daily for 30 days. That’s the threshold where AI stops being a topic and becomes a tool. If you want a structured way to get there, that’s what Executive AI Fitness is for.
Which one hit you hardest this week? 👇