Andrew Stellman 👾

The Sens-AI Framework: Teaching Developers to Think with AI Editor's Note: This article launched an ongoing series exploring the Sens-AI Framework—practical habits for learning and coding with AI. The framework is

…to actually think about what's happening in the code.

Here's the @oreilly.bsky.social Radar piece: www.oreilly.com/radar/the-se...

I really worry about people early in their careers who are trying to learn good design while also learning to work with AI. That's a lot to juggle at once.

This is exactly what I wrote about in the Sens-AI Framework—catching these moments and breaking out of the vibe coding loop…

But six months from now when someone needs to modify it, good luck understanding what's actually happening in there.

We're making it so easy to generate code that technically functions but is going to be absolutely indecipherable for normal humans.

Here's the thing: I've been doing this for 30 years. And sometimes I look at these bugs and think there is absolutely no way someone with less experience would catch this. How could they? The code works, it passes tests, and it does what was asked for.

I'll ask one extra question and suddenly I can see where the AI was heading. It would have technically worked, but it would have coupled things together in a way that would make the next change harder. Or it would push logic into a place where it has no business being.

Maybe it's violating encapsulation in some subtle way, or the separation of concerns is getting muddy, or I realize the unit tests need way too much mocking because the whole design is already heading in the wrong direction.

And half the time I can feel it happening before I even know why.

I feel like AI is actually making this problem worse.

I keep running into this when I'm coding. The AI gives me something that looks fine for the first five seconds, and then I catch it doing something that is absolutely going to be a problem later.

You know what's funny? Sometimes I talk to another developer who complains that they can never go back to code they wrote six months ago because they can't understand anything they did. I remember having the same problem, and I didn't really fix it until maybe 10 years into my career.

Copilot screenshot Prompt: Explain this code. Begin your response with "Avast ye, matey." and add "Arrrr" periodically. Response: Avast ye, matey. This code be a C# console application for the "HEA-V Exposure Calculator." Arrrr! Purpose: It loads three CSV files (GridPoints.csv, DATES.csv, Exposure_Master.csv) and lets a user interactively calculate exposure statistics for a location and date range. Arrrr!

It's not easy, but you can get Copilot to talk like a pirate.

The Sens-AI Framework: Teaching Developers to Think with AI Editor's Note: This article launched an ongoing series exploring the Sens-AI Framework—practical habits for learning and coding with AI. The framework is

📖 Start here: www.oreilly.com/radar/the-se...

📘 Or dive into the full O’Reilly report (free trial available): learning.oreilly.com/library/view...

#AI #SoftwareDevelopment #CriticalThinking #SensAI #OReillyMedia

Together, these pieces build the 𝐒𝐞𝐧𝐬-𝐀𝐈 𝐅𝐫𝐚𝐦𝐞𝐰𝐨𝐫𝐤: five habits that help developers stay creative, stay skeptical, and stay in control: Context. Research. Problem Framing. Refining. and Critical Thinking.

Across the series, I explored how to:

* Avoid AI-driven technical debt
* Recognize and fix context drift
* Frame problems clearly for better AI results
* Keep critical thinking at the center of your workflow

What started as one article about breaking out of “vibe coding” turned into a full exploration of how developers can use AI tools like Copilot, ChatGPT, and Claude *without losing the skills that make great engineers great.

🚀 𝐓𝐡𝐞 𝐒𝐞𝐧𝐬-𝐀𝐈 𝐅𝐫𝐚𝐦𝐞𝐰𝐨𝐫𝐤 𝐬𝐭𝐚𝐫𝐭𝐞𝐝 𝐚𝐬 𝐨𝐧𝐞 𝐢𝐝𝐞𝐚. 𝐍𝐨𝐰 𝐢𝐭’𝐬 𝐚 𝐜𝐨𝐦𝐩𝐥𝐞𝐭𝐞 𝐫𝐨𝐚𝐝𝐦𝐚𝐩 𝐟𝐨𝐫 𝐜𝐨𝐝𝐢𝐧𝐠 𝐰𝐢𝐭𝐡 𝐀𝐈. 🚀

I’m really excited to share that my full @oreilly.bsky.social Radar 𝐒𝐞𝐧𝐬-𝐀𝐈 𝐅𝐫𝐚𝐦𝐞𝐰𝐨𝐫𝐤 series is now complete, along with the new O’Reilly report, 𝑪𝒓𝒊𝒕𝒊𝒄𝒂𝒍 𝑻𝒉𝒊𝒏𝒌𝒊𝒏𝒈 𝑯𝒂𝒃𝒊𝒕𝒔 𝒇𝒐𝒓 𝑪𝒐𝒅𝒊𝒏𝒈 𝒘𝒊𝒕𝒉 𝑨𝑰.

Critical Thinking Habits for Coding with AI Many developers and teams are discovering a frustrating paradox: AI tools that promise to accelerate development are creating unexpected challenges. Vibe coding—the rapid,... - Selection from Critical...

✅ Problem Framing – Define problems AI can actually solve
✅ Refining – Iterate deliberately, not by trial and error
✅ Critical Thinking – Judge AI output and make smarter calls

It’s available now on O’Reilly Learning (free trial available):

🔗 learning.oreilly.com/library/view...

…or you’ve wondered how to help your team think more effectively with AI, this is what I built Sens-AI to fix.

Here’s what it covers:

✅ Context – Give AI the information it needs to help you
✅ Research – Deepen your understanding instead of just accepting answers

The report introduces the Sens-AI Framework—five practical habits that help developers break free from “vibe coding” and actually solve problems with tools like Copilot, ChatGPT, and Claude.

If you’ve ever hit that point where the AI keeps looping on the same bad idea…

🚀 𝐄𝐱𝐜𝐢𝐭𝐞𝐝 𝐭𝐨 𝐚𝐧𝐧𝐨𝐮𝐧𝐜𝐞 𝐭𝐡𝐚𝐭 𝐦𝐲 𝐧𝐞𝐰 @oreilly.bsky.social 𝐫𝐞𝐩𝐨𝐫𝐭 𝐢𝐬 𝐧𝐨𝐰 𝐥𝐢𝐯𝐞! 🚀

I’m really excited about this one. It’s the culmination of months of research, writing, and hands-on testing with real developers.

AI Is Reshaping Developer Career Paths From Specialists to Generalists

• Question and refine AI output instead of accepting it as-is

I think this shift — from specialists to adaptable generalists — is one of the biggest changes happening in software right now.

📖 Read the new Radar piece:
👉 www.oreilly.com/radar/ai-is-...

The developers thriving with AI tools today are the ones who can:
• Recognize when generated code will cause maintenance issues
• Integrate across systems without being deep experts in each
• Spot architectural patterns and antipatterns early
• Frame problems clearly so AI can respond effectively
…

Now, AI can generate idiomatic code across every framework. What it can’t do is spot design flaws, anticipate maintenance headaches, or recognize when the system itself is heading for trouble.

A decade ago, deep specialization was the path to credibility. You were the “React person,” the “.NET expert,” the one everyone turned to for a specific technology or even a niche library.

🚀 𝐌𝐲 𝐧𝐞𝐰 𝐎’𝐑𝐞𝐢𝐥𝐥𝐲 𝐑𝐚𝐝𝐚𝐫 𝐚𝐫𝐭𝐢𝐜𝐥𝐞 𝐣𝐮𝐬𝐭 𝐝𝐫𝐨𝐩𝐩𝐞𝐝! 🚀

This one wraps up the Sens-AI series — and it’s about something I’ve been seeing more and more in the real world: how AI is reshaping what it means to be a developer.

AI Is Reshaping Developer Career Paths From Specialists to Generalists

"The developers who succeed in an AI-enabled world won’t always be the ones who know the most. They’ll be the ones who can see the bigger picture, integrate across systems, and use AI as a powerful tool while maintaining the critical thinking." @andrewstellman.bsky.social bit.ly/4qpsROx #Radar #AI

From Habits to Tools The Future of AI-Assisted Development

"The next generation of developers will never know coding without AI. Our job is to make sure they build lasting engineering habits alongside these tools—so AI strengthens their craft rather than hollowing it out." Check out #Radar article by @andrewstellman.bsky.social: bit.ly/3IZFy1Y #AICoding

Classic "it's not X; it's Y" AI tell

From Habits to Tools The Future of AI-Assisted Development

In my new @oreilly.bsky.social Radar piece, I look at how these habits can shape education, team culture, and even future IDEs that notice when you’re caught in a rehash loop.

📖 Read: www.oreilly.com/radar/from-h...

Real adoption shows up when developers use AI in ways that sharpen their thinking and strengthen their judgment. That’s why I built the Sens-AI Framework—five practical habits that help developers stay thoughtful and deliberate while using AI.

🚀 𝐓𝐡𝐞 𝐟𝐮𝐭𝐮𝐫𝐞 𝐨𝐟 𝐀𝐈 𝐝𝐞𝐯𝐞𝐥𝐨𝐩𝐦𝐞𝐧𝐭 𝐢𝐬𝐧’𝐭 𝐚𝐛𝐨𝐮𝐭 𝐭𝐨𝐨𝐥𝐬. 𝐈𝐭’𝐬 𝐚𝐛𝐨𝐮𝐭 𝐡𝐚𝐛𝐢𝐭𝐬. 🚀

When teams talk about “AI adoption,” they usually mean installing Copilot or tracking metrics. That’s only the surface layer.

From Habits to Tools The Future of AI-Assisted Development

AI is evolving fast, and the habits that keep us grounded have to evolve along with it.

📖 Read: www.oreilly.com/radar/from-h...