Two Prompting Tips That Matter

It’s been 3.5 months since I stepped away from institutional investing and committed fully to AI-driven investing.

In that brief span, something fundamental shifted.

AI adoption in investing didn’t edge forward quietly.

It roared, like fire through dry grass.

Everywhere, investors are experimenting. Builders are building.

And yet, this is only the prologue.

The deeper potential remains hidden, waiting to be unearthed.

---

What I’ve Built (So Far)

I’ve pushed LLMs to their limits - questioning, testing, improvising.

From those experiments came a multi-agent, AI-native investment system.

More than 20 modules already, with more still being forged.

A next-generation AI investing system isn’t a dressed-up ChatGPT with a RAG database behind it, nor a fancy summarizer.

It isn’t just about serving investors more information faster.

A true AI investing system should make an investor much better and make much better decisions.

In my own setting, it must pitch me.

And it must win against me.

And in the short term, it has.

It scored where I hesitated.

It spotted winners I second-guessed.

It flagged losers I was too romantic to let go of.

This isn’t a toy. It’s a cognitive partner that evolves.

---

Two Prompt Questions to Address

Prompts are everywhere.

Everyone is talking, sharing, and showcasing.

But the overwhelming tide reduces to two core questions:

---

Question 1 — Are Prompt Templates Necessary?

On LinkedIn and X, you’ll see “Top Analyst Prompts” bundled neatly, or demos showing how to spin up earnings recaps in seconds.

The underlying claim: collect enough prompt recipes and you’ll master AI.

Do investors actually need to store prompt templates?

No.