Why I'm Optimistic
About AI 🤖
Especially for Software Engineers 💡
🎙️
Tarun Chawdhury
AI Optimist — with Caution ⚖️
Especially for Software Engineers 💡
Someone in my family couldn't step over a bathtub safely. I called contractors, asked friends — and almost everyone said the same thing: replace the whole tub, install a walk-in shower. Expensive, and supposedly the only option. So I asked an AI for another perspective.
Not because I trust it blindly — AI makes mistakes, but so do humans. Sometimes its real value is just another perspective.
It named something I'd never heard of — a tub cutout that lets you step safely into an existing tub, no full replacement.
A real product category with multiple providers. A full replacement became a small, much cheaper transformation.
💡 Maybe the future of AI isn't about replacing everything — maybe it's about connecting and adapting what already works.
That bathtub was the exact same choice every enterprise faces with AI today.
How do I make this safe — without tearing the whole thing out?
✓ Keep the tub. Cut an opening. Add a step.
How do decades of ERP, CRM & legacy apps join the AI world?
✓ Keep the systems. Add an adapter. Expose them to AI.
The value isn't trapped because these systems are old — it's trapped because they're isolated.
Two false choices everyone assumes — versus the smarter third path.
Standards may evolve — MCP today, A2A emerging — but the need to adapt, not replace, won't change. 🔗
When people say software engineering is disappearing, I see the opposite. Someone has to build, secure & govern that adapter layer — and the most valuable engineers won't build the models, they'll connect them to the real world.
That's only half the opportunity. Even when AI writes the code, humans still design the systems — and the principles are changing. Historically we built software for human users; now we build for humans and AI agents.
We designed applications primarily for human users.
Every new app thinks about interoperability from day one — how AI discovers it, understands it, and invokes it safely.
Not because AI is perfect. Not because it won't disrupt jobs — it will. But because every major technology shift creates new categories of work.
The bathtub taught me something simple: sometimes the smartest solution isn't replacement — it's adaptation. Connecting what already works to what's coming next. And we're only getting started.