Opus 4.7 is here. It illustrates 2 important facts: 1) Don't assume AI even knows what it's capable of. 2) Token costs are hard to decipher.

AI use can look very different depending on who is building the solution. From AI-free to AI autonomous, it all depends on what you want

AI doesn’t always know what its latest models are capable of, because it hasn’t yet been trained on that data. So if you think AI can do more than you’re seeing, just ask it to check. A real-time research prompt could save you a lot of time and rework later on.

I’m consulting with businesses on how to integrate AI in practical, useful ways. I’m teaching and building with the AI tools, but the bigger part is doing what really needs to get done: understanding the business, defining the goals, and finding where they can remove obstacles that will have real, immediate impacts on the top and bottom lines.

You shouldn’t hand your processes over to AI. You also shouldn’t keep AI out. Instead, build an automated workflow and see where AI can support the person responsible for getting the job done.

A simple framework to help you build your first AI agent. It’s not a new technical task to master. Just think of it as a new employee to onboard. What are the 5 things someone would need to do the job right? That’s BRIEF.

You can’t build an AI solution before you understand the actual problem you’re solving.

Instead, be curious about AI and just start to use it. Pick something you’re interested in learning. Maybe it’s something to help make your day or job easier. Maybe it’s a creative passion that you haven’t had the time or tools to dive back into. You’re not going to learn it by watching a TED talk. You’re going to learn this by experimenting with it. Playing with it. Being curious. Because you’ll be learning the foundation of the technology, not just the latest specs.