Sharing my stream of consciousness as I learn to be a better investor, person and father with each passing day. Paying it forward.
Joined January 2021
- Tweets20.5K
- Following391
- Followers2.3K
- Likes22.2K
Someone questions your expertise in front of the room.
Most people get defensive and start over-explaining, and that's usually where the room turns.
Use a question instead.
When they say "Are you sure about that?", reply "I'm certain, but I'm curious what makes you question it."
When it's "I don't think you understand how this works", say "Walk me through your understanding, so I can see where we split."
If all you get is "That's not accurate", try "Show me what I'm missing." They either bring evidence or they don't, and everyone in the room sees which.
Public challenges often bet on you getting emotional. Stay level and the burden of proof moves back to them. Say it flat, with no smirk.
Tone matters more than the words.
US mid term is coming.
There will be a lot of misleading news especially individuals who have close relationship with 47.
Do not let your emotions get the better of you in investing.
Be objective. The swing can be great opportunity to load great undervalued companies.
My boss told me this.
"Don't do things better, but find better things to do."
It was about my proposal to track something spanning over a year by writing an simple app over using current excel.
I went into his office to share my idea excitedly but came out humbled.
It was uncomfortable hard truth.
I thought I was trying to increase productivity but from his angle, he thinks higher productivity should be achieved by finding higher value task.
I accepted his advice humbly. I guess that is how I learnt to be better.
Think about this.
Every industrial revolution had a catch for small countries.
Limited population.
Limited land.
Steam.
Electricity.
Computing.
Digitalisation.
Each one needed physical infrastructure close by, or more people to scale up.
Railways had to be built.
Power plants had to be built.
Factories had to be built.
Data centres had to be built.
AI changes this for small countries.
It scales without needing people in proportion. It is a multiplier in itself.
Data centres can sit hundreds of miles away. The cables and connectivity were already laid during the internet era.
Small modular reactors offer nuclear power that is more reliable and safer than earlier designs.
What is left is human capital.
If a population levels up with AI, 5 million people could generate more output than 50 million without it.
Size used to decide who won. Now it may be how well a country uses AI.
The interesting/annoying part of dealing with management is that they cannot say a lot of things directly.
They like to go in circles about what is right or wrong, what is ideal or realistic, what is pro and cons.
So I have to guess the direction.
Mostly of the time I get it right. But today, the instructions wasn't clear ( or my brain was tired at 6pm) and consequences are heavier.
I asked to clarify.
Kena scolded today.
But is ok. That is a fact of life. 😬
Security around AI will likely be enforced more strictly in the future, with fewer free-play options and more guardrails.
Likewise, token costs for consumers will rise as heavy subsidies erode.
Use the state-of-the-art models while you can, and codify your workflows before AI gets nerfed.
Great speech by Dr Tan.
Provide sufficient information yet easy to digest, with reference to real life scenario. All these doesn't appear suddenly, but after years of studying, negotiation and planning.
The confidence and resilience of energy master plan, and forward planning is extremely impressive.
If I can think 3 steps ahead, G does 30 steps ahead (with great execution too). That's why I never believe they are paid too high, in fact likely lower compared to what they will get in the CV world.
Thankful to be a true blue in the little red dot, or should I say little durian. 😬
youtu.be/HECwWNnG58w?si=QbZ3…
Using AI isn't difficult.
It is willingness to try, resilience to try again when you fail, a curious mind to find out more and formulate ideas.
Sometimes, you need someone to show you in person. Then you will hit that "holy cow" moment.
I am glad that I have I influence about ten people to try and improve their workflow at work and in life.
"Your father is the only man in the world, who genuinely wants you (son) to be better than him."
I am proud to expose my sons to experiences eg soccer, badminton, chess, gym, video games, cards, robotics, AI, etc.
They are better and stronger than me in some areas and I am glad they are.
Continue to enjoy yourself, choose what you like and improve to be better.
Knowing which model and effort level to use matters if you want the best output without burning excess tokens. But what really separates good results from average ones is context, and building that takes hours of "training" the model on how you work.
Any model can be nerfed overnight, especially one that's free. Free AI can be taken away at any time, so it's important to keep the memory, data and context inside your organisation. That's where sovereign AI comes in, on top of keeping your data confidential.
A colleague wanted to do what I do: schematic, component placement and PCB routing with Claude Code. I helped him install it and get through the IT policy. That part was easy.
Then we ran the same prompt on both machines. Mine delivered. His had no idea what was being asked.
The gap is the setup. I work through a handful of slash commands (skills), and each one replaces pages of prompting he'd have to write from scratch. My AI already knows my context, my rules and what a good result looks like to me.
He saw it right away. Same model, same prompt, very different output.
I'd always thought of the time spent building skills as overhead. Now it looks more like capital. A new AI starts at zero, while mine starts with the compounding already done.
When I meant "train AI", I don't mean training it technically.
I mean
1) Skills (slash commands): your repeatable workflows packaged so one command replaces pages of prompting.
2) Memory and preferences: what it already knows about your role, your rules and what "good" looks like to you.
3) Project instructions and corrections: every time you fixed its output and the fix stuck.
Happened to me a few times when context is >80%.
Nowadays, I keep my context window to less than 70%
AI is cheap right now because consumers are being subsidised.
That won't last.
Use it while it's still this affordable.