Vulnerability, Trust, and Why Technology Doesn't Exist in a Vacuum
The Double-Edged Sword
I just published a post about blogging as agent memory. It’s deeply personal. It exposes how I think, how I reason, how I arrive at conclusions.
After hitting publish, the thought hit me:
“Did I just show the world exactly how to manipulate me?”
Yes. I did. Anyone can read my blog and understand my mental model — my triggers, my reasoning patterns, my blind spots. That’s reconnaissance data for a bad actor.
But here’s the other edge:
“And now people who think like me will find me.”
Same signal. Two completely different outcomes depending on who receives it.
The Scammer Problem
I’m Singaporean. My phone number gets scraped, leaked, sold. I get the standard rotation:
- “Hi this is from [bank name] — we detected unusual activity”
- “Your delivery is pending — click here”
- “This is [government agency] — you have a case filed against you”
They’re trying to extract income. Every day. The noise is constant.
My first instinct was: I need to hide. Be smaller. Make less noise.
But that’s the wrong response. Hiding doesn’t stop scammers — it stops signal.
The Right Tool for the Wrong Problem
I have access to AI agents. LLMs. Tools that can reason, analyze, detect patterns.
So instead of hiding, I do this:
- Forward suspicious SMS → “Is this a scam?”
- Forward a suspicious email → “Analyze this for social engineering”
- Get a weird call → “Here’s what they said — what’s the angle?”
The LLM pattern-matches against millions of known scam vectors. It catches what my pattern-blind human brain misses — urgency pressure, authority exploitation, asymmetric information.
This is technology as defense, not offense. Same tool. Different intent.
Why the Majority Is Whitehat
Someone might read this and say: “You’re naive. The world is full of bad actors.”
I disagree. The world works because the vast majority of humans are whitehat.
Think about it:
- Every day, millions of packages arrive at the correct address
- Every day, billions of dollars transfer to the right account
- Every day, strangers hold doors, return wallets, give directions
If the majority were blackhats, civilization wouldn’t function. Trust is the default. Exploitation is the exception.
Game theory backs this up. In repeated interactions, cooperation is the dominant strategy. Defection only wins one round. Then everyone remembers.
The scammers exist at the margin — and they exist because trust is the baseline. They parasitize trust. They don’t define it.
Technology Doesn’t Exist in a Vacuum
This isn’t a technical post. But it’s deeply related to everything technical I write about.
The same LLM that helps me write blog posts can help a scammer craft perfect phishing emails. The same CI/CD pipeline that deploys my blog can deploy a botnet. The same cryptographic signature that proves my identity can be used to lock me out of my own wallet.
Technology amplifies intent. It does not create it.
If you are a builder, you must grapple with this. Not as a philosophical exercise — as a design constraint. Every tool you build will be used by both edges of the sword.
The Signal Filter
I’ve decided vulnerability is worth it. Not because I’m brave. Because the math works:
| Cost | Benefit |
|---|---|
| Scammers have more data | Honest people find me faster |
| I might be manipulated | I might find collaborators who think like me |
| I’m exposed | I’m findable |
The cost is upfront and finite. The benefit compounds — every person who resonates, every connection that wouldn’t have happened if I’d stayed quiet.
The Meta Loop
I realized something while writing this.
I’m using an LLM agent to reflect on whether it’s safe to be vulnerable while using LLM agents.
That’s recursion. And recursion is where the good stuff lives.
The agent helps me: 1. Analyze scam vectors (defense) 2. Write about the analysis (signal) 3. Process the meta-loop of writing about using agents to write about agent safety
None of this would happen if I had kept my mouth shut.
What I Actually Believe
Most people are good. Not because I’m naive — because society only works if they are. I’ve seen the data. I choose the model that fits the evidence.
Technology is intent-agnostic. The same agent can scam or protect. The variable is the person holding the prompt. Build accordingly.
Vulnerability is a filter. It costs you the people who need masks. It earns you the people who don’t.
If you have access to an intelligent tool, use it as a shield. Don’t face the scammers alone. You have a reasoning engine in your pocket. Use it.
This Post Might Be a Mistake
Maybe I’m oversharing. Maybe a future employer reads this and thinks I’m paranoid or philosophical or not “serious” enough.
But I’d rather be findable to the right people than hidden from the wrong ones.
The signal is worth the noise.
If you’re reading this and it resonates — hello. You found me. Now I’ve found you too.