In this week’s newsletter, I wrote an essay about how I too, can feel FOMO about how others claim to use AI agents in their lives. But we shouldn’t! I also recap what I wrote about vibe coding’s “Three Ps”, a new AI lifehack, rogue autonomous agents from OpenAI, the backlash against artificial-looking movies, AI nutrition labels, and more.

Photo of the week

A 16-year-old dog, July 2026, Hong Kong.

Weekly Recap

The Three Ps to remember about vibe coding – Problem-solving, play and patience.

I discovered another simple but useful AI lifehack on my phone.

No, don’t tell me to calm down about rogue autonomous agents! We’re ten years away from T-800!

Kevin Kelly wrote a great essay that can stir you to ask some difficult questions about yourself.

It’s good to see the backlash against artificial-looking movies.

A useful infographic from Ethan Mollick educating everyone about chatbots and agentic systems.

“Nutrition labels for Gen AI chatbot apps” is probably a good idea but difficult to implement in a simple way.

LinkedIn’s latest collaborative feature doesn’t seem to work. But we had fun with it anyway.

A compilation of my photos from our recently concluded NTU journalism expedition in Hong Kong, known as Going Overseas For Advanced Reporting. Here’s me being a tailor’s model.

Sunday Essay

Whenever other AI enthusiasts post online about all the AI agents that they are using to run their lives, I will feel as perplexed as you. “Am I missing out on something here?”

For example, some people wax lyrical about how great it is to have an AI Chief of Staff to read through all their emails, messages, schedules, reports, et cetera, et cetera. and the AI will neatly packages all the go-dos in one long summary message.

Ok, I guess CEOs of big firms are pretty busy. I do have an Copilot Workflows agent that reads my NTU emails and updates me daily about my immediate action items, and that’s about all the automation that I currently have.

I try to think through my daily routine to see if I can AI-automate my activities, and again, nada. My routine involves morning exercise, doing groceries, grinding coffee, Bible study, learning new things, engaging with people, teaching hard and soft skills to young people and adults, being asked to solve random problems, walking many steps, and appreciating my wife’s cooking.

Of course, within each activity, I might use Gen AI.

Before I start my jogs, I might use ElevenReader to convert a website or PDF into an audiobook. During my Bible study, I might do some research on Bible history with ChatGPT to understand the verses better. When I learn new things, I use Google AI summaries to accelerate my research. For example, yesterday, I read in a book that the “saw palmetto” plant extract was good for aging men’s prostate issues, but Google said that there was no scientific evidence.

And just this morning, I was buying groceries. I took a photo of Korean muscat grapes at the supermarket and asked ChatGPT: “How many calories?” The answer frightened me. I decided not to buy the grapes because I would not be able to resist eating the whole bunch of 310 calories at one go.

Now that said, I do need to be actively involved with AI automation and AI agents because people want to learn them in my AI workshops. These AI solutions are most useful for people who have to make strategic decisions using a lot of data and transactions. AI is not perfect at web scraping, but it can pull in a good deal of market data for bosses, marketers, data analysts and other data-driven roles.

However, if your job requires you to constantly engage with people directly (eg. teaching in person, customer satisfaction, financial advice, etc), you will realise that it’s less about agents or automation, and more integrating AI capabilities to improve your workflow.

What I’m trying to say here is – don’t feel that you have to use AI like other people. To each their own. Here is my way of applying AI in my life:

1. Start with the real problem you want to solve.
2. Work backwards and understand the steps involved (the workflow).
3. Ask where AI can help you, and then immediately try it out. Don’t procrastinate!
4. Then decide if AI was of any help. Do you need to refine AI’s involvement? Automate it? Create an app? Or just do it the old way?

The caveat here is that in order to do the third step, you need to have good working knowledge of the different AI apps and capabilities out there. This takes time and curiosity on your part, and it helps to read AI newsletters (see my reading list here) to discover what AI tools are out there.

In closing, don’t have FOMO. Just keep experimenting with AI to see how it can actually work in your life. Sometimes yes, sometimes no.

Thanks for reading!

26-Jul-2026