I’ve been using AI for a while now and I continue to learn better ways to use AI and prompt engineering. My latest use is something I picked off the Internet that really scaled things up.
Essentially, I came across a five stage method for interacting with AI and gave it a go on a topic that has been troubling me for some time: the never ending rise in the stock market.
But first, let’s take a look at the five stage process – this came from a Reddit post and I apologize for not saving the link or name it came from as it was while I was traveling and juggling a ton of things:
THE PROCESS
Proceed through the following five stages one by one. After presenting your findings for a stage, ask for my feedback or input before moving to the next.
Stage 1: Gather and Scrutinize Evidence
Identify the core facts and data. Question everything.
- Where did this info come from?
- Who funded it?
- Is the sample size legit?
- Is this data still relevant?
- Where is the conflicting data?
Stage 2: Identify and Challenge Assumptions
Uncover the hidden beliefs that form the foundation of the argument.
- What are we assuming is true?
- What are my own hidden biases here?
- Would this hold true everywhere?
- What if we’re wrong? What’s the opposite?
Stage 3: Explore Diverse Perspectives
Break out of your own bubble.
- Who disagrees with this and why?
- How would someone from a different background see this?
- Who wins and who loses in this situation?
- Who did we not ask?
Stage 4: Generate Alternatives
Think outside the box.
- What’s another way to approach this?
- What’s the polar opposite of the current solution?
- Can we combine different ideas?
- What haven’t we tried?
Stage 5: Map and Evaluate Implications
Think ahead. Every solution creates new problems.
- What are the 1st, 2nd, and 3rd-order consequences?
- Who is helped and who is harmed?
- What new problems might this create?
FINAL SYNTHESIS
After all stages, provide a comprehensive summary that includes the most credible evidence, core assumptions, diverse perspectives, and a final recommendation that weighs the alternatives and their implications.
My Problem Statement: Best hedging strategy against an overvalued stock market.
Because this was an interactive conversation for each stage of the process, I won’t post the very long detailed conversation but instead jump to the summary it posted. I’d like to add that AI changed my mind about some of the assumptions I had been making about how/why/where to hedge and the outcome was far better than I anticipated or suspected.
WARNING: DO NOT DO ANY TRADING OR SPECULATION BASED ON THE INFO BELOW, THIS IS AN EXPERIMENTAL DISCUSSION WITH AI AND THIS IS FOR EDUCATIONAL AND ENTERTAINMENT PURPOSES ONLY!
Note: I decided to XXXX out the specific date ranges because I don’t want anyone making foolish trading decisions.
FINAL SYNTHESIS
🎯 Goal
To hedge against a potential 30%+ drawdown in U.S. equities over the next 1–2 years, using a multi-instrument, contrarian strategy targeting the most sensitive and likely crash vectors: tech valuations, corporate credit stress, and volatility spikes.
✅ Most Credible Evidence
- Historical data shows 30%+ drawdowns occur once every 10–15 years — we are within a plausible window based on elevated valuations and macro stress.
- Leading indicators (inverted yield curve, credit market divergence) suggest late-cycle conditions.
- SPY has strong momentum, but the real fragility may lie in QQQ (tech beta), HYG (junk credit), and volatility suppression.
🧩 Core Assumptions
- Timing is unpredictable, so protection should be long-dated (1–2 years).
- The next correction will likely begin in either high-beta tech, stressed credit markets, or via a volatility event.
- A layered hedge offers broader protection than a single bet (e.g., SPY puts).
🌐 Diverse Perspectives We Considered
- Passive investors dismiss hedging; Taleb-style thinkers argue it’s essential.
- Contrarian macro investors argue that SPY is the wrong thing to hedge — the stress starts elsewhere.
- Quant and behavioral perspectives highlight the need to control costs and avoid timing error.
🧪 Final Hedge Structure (Based on $100,000 Budget)
| Hedge Instrument | Allocation | Example Trade (as of mid-2025)* | Purpose |
|---|---|---|---|
| QQQ LEAPS Puts | $40,000 | Buy XXXX 320 Puts (~25% OTM) | Hedge tech-led equity unwind |
| HYG Puts | $25,000 | Buy XXXX 70 Puts (~15% OTM) | Hedge credit stress |
| VIX Calls | $20,000 | Buy XXXX 35 Calls (cheap tail) | Hedge volatility spikes |
| Cash Buffer / Flex | $15,000 | Reserved for rolling or tactical add | Optional / Rebalancing fuel |
*Note: Strike prices and premiums may vary — these are illustrative. You’ll want to check real-time pricing and IV before execution.
📉 Risk Profile
- Max Loss: $100,000 (premium paid)
- Potential Gain: Several multiples of invested amount in a severe correction (~30–50%+ market decline)
- Break-Even: Likely triggered if QQQ, HYG, or VIX move sharply in <18 months
🔁 Management Guidelines
- Quarterly check-ins: Re-evaluate implied volatility and price decay.
- Avoid knee-jerk exits: Let long-dated options run unless thesis invalidates.
- Profit harvesting: If hedge doubles or triples in value, consider harvesting and rolling to new strikes.
I’m blown away and concerned about how AI is rapidly improving over time and the right prompt engineering is taking things to a whole new level. I recently read that AI is better at creating legal motions better than most junior attorneys now, perhaps the best lawyer in the country will be a prompt engineer.
Share The Wealth
What do you think? Should I hedge using AI’s strategy or use my own? Let me know in the comments below.