From Symbian to AI: Why "Judgment" is the only skill that scales
I’ve spent two decades navigating the tectonic shifts of technology. My journey started by coding for the first generation of feature phones (Nokia/Symbian/MTK). Later, I founded a smart pet hardware startup, which I eventually led to a successful acquisition.
But I didn't stop at building. I’ve also sat on the other side of the table—investing in the AI sector and traditional education/training institutions. These experiences gave me a macro view of how technology either creates real value or just creates noise.
The "Partial AI" Trap (Observation from 30+ Interviews): Recently, I’ve been back in the builder’s seat. After interviewing over 30 developers for my new projects, I’ve noticed a disturbing trend. Most senior devs are stuck in a "Copy-Paste" loop. They use AI to generate a single function, but they refuse to let the AI see the "Full Truth" of the entire repository.
They are treating a "Co-Architect" like a "Dictionary." In my view, this is a defensive mechanism—a refusal to adapt to a new paradigm because it threatens the traditional "manual control" they’ve mastered for years.
The "OpenClaw" Hype vs. Reality: From my perspective as an investor and founder, the influencer-led hype around tools like OpenClaw is irresponsible. Viral threads promise magic, but they conveniently ignore the massive token burn, the brutal fine-tuning required, and the systemic risks that can drain a startup's runway in days.
AI is a leverage tool, not a "set and forget" engine.
The "Vibe Coding" Shift: This is why my team is moving away from line-by-line management to "Vibe Coding." We focus on system logic and product value, letting the AI handle the heavy lifting with full context. This philosophy has allowed us to ship:
ObjectRemover.video: Solving complex video masking with SAM 3.
TokenLens.ai: Data-driven intelligence for viral growth.
My Question to the Community: With my background in hardware, education investment, and now AI, I see a massive gap between "Coding" and "Building."
Are we raising a generation of "Prompt Copy-Pasters," or are we finally entering the era of the "System Architect"? How are you ensuring AI is actually driving ROI in your workflow?
Replies