Slow Down to Speed Up with AI

💬 Testimonials

Thank you Trung so much for your amazing and insightful sharing yesterday. It was super helpful and I genuinely learned a lot. I truly hope to have further discussions!

— Hyeseong Ju, Aspiring AI and Fintech Product Manager

Once again thank you for the enthusiastic talk with the well prepared slides. Actually I wasnt part of the event yesterday but sat in to talk to my friend. however I was pleasantly surprised with your talk which coincidentally aligned with what I wanted to learn during this summer holiday.

— Dong En, SMU Merit Scholar

Thank you for the insight, I’ll be sure to explore new tools to build up intuition to be more AI native 👌

— William Hansel, Programmes Executive @ SMU AI

Three of us from Ascenda ran a workshop at SMU .Hack, on software architecture and working with AI. It was good to be back at SMU after more than a year. Wei Yi from 2025 workshop is now an intern with us. A year moves fast 😂

My part was a talk: Slow Down to Speed Up with AI. The idea is simple. AI made writing code almost free, so the slow thinking is the job now. What to build, and whether the code it shipped is right, is still on you. That part did not get less important. It got more.

I shared the mistake that taught me this. I vibe coded a product flow for weeks without reading the diff. It looked fine in the browser, so I kept shipping. When I finally read the code, I did not recognize it. Five components were rendering the same thing, and I spent days cleaning up what I thought I had saved.

So now I slow down in two places. I get clear on what I want before AI runs, and I read the diff before anyone reviews or deploys it. That is not extra process. It is the cheapest place to catch a mistake.

For students starting out, the good news is you do not need much. Learn 20% of a language and you can say 80% of what you need. AI is the same. Pick one tool. Run Explore, Plan, Code, Commit. Keep an AGENTS.md. Build skills for what you repeat.

One more thing I told them: delegate the doing, never delegate the learning. The slow phase is where you earn the judgment you will need later to steer AI.

Thank you SMU for having us, and I hope to be back.