Two years ago I was seriously considering moving to Austin for a job that put me at the forefront of all things AI. Everyone I knew in the industry seemed to be going. I'd visited nearly every year for SXSW, liked it very much, and had convinced myself that the energy of the place was exactly what I needed. I even went as far as pricing homes. But, in my head I wasn't completely sold.
That's around the time I developed the hype check prompt for ChatGPT, and I've used it countless times since on Gemini and Claude. When you run this prompt, what often comes back is...uncomfortable, to say the least. Here's how it works.
What a hype check actually is
In the case with a big move, ChatGPT pointed out factors that I hadn't thought of — that my visits to Austin had always been in early spring — arguably the prettiest time in Austin other than the fall. None of that was information I didn't have. It was information I'd been rearranging around the conclusion I'd already reached.
I didn't end up moving. A friend who made a similar move around the same time came back eighteen months later. The first thing she said was that she hadn't anticipated how much she'd underestimated what she was leaving.
Those unasked questions are exactly what makes this prompt work so well. A hype check flips the usual question. Instead of asking "will this work?" you imagine it's already failed — and then you work backwards to figure out why.
The concept comes from organizational psychology. It's been used for decades. The idea is simple: humans are bad at anticipating failure when they're emotionally invested in success. AI doesn't do any of that. It has no ego investment in your decision going well.
Here's the exact prompt I use: "Imagine it's 12 months from now and this decision turned out to be a complete disaster. I'm about to [decision]. What went wrong? Be specific, be brutal, and don't reassure me."
That last instruction matters more than it sounds. Without it, the AI hedges. It balances every risk with a reassurance. The "don't reassure me" line unlocks the version that's actually useful — the one that names the thing you've been quietly afraid to say out loud.
Three huge mistakes I avoided
The role I mentioned above was my clearest example. On paper: senior position, 30% salary bump, very well-known AI company. In my gut: something I couldn't quite name.
When I ran the hype check, ChatGPT came back with specifics. None of which the AI invented, In fact, they all were things I either knew or could have found out. The prompt just forced them onto the page where I had to look at them directly instead of letting them float at the edges of my thinking.
I turned down the offer and landed a job here at Tom's Guide instead, which has changed my life in all the best ways.
I used this prompt to avoid buying a house that seemed fantastic at the time but ended up being a disaster (and that I was lucky to avoid). I've also used this prompt with my husband's business and financial decisions.
Let me be clear: AI is not the decision maker, you are, but using a prompt like this can help you ask the right questions to make your decisions clearer.
The variation that's equally useful
Once you've run the failure version, try this follow-up: "Now imagine it's 12 months from now and this went better than I could have hoped. What made the difference? What did I do in the first 30 days that set it up to succeed?"
This is a positive hype check. What I like about it is the way it shifts from risk-scanning to execution-planning, and it tends to surface the specific early actions that matter. These are the things, at least for me, that are easy to skip when I'm excited about a new decision.
When I have big decisions like buying a new car or even considering which place to send my kids to camp, I run both versions. I typically do them back to back, for anything significant. The whole process takes less than ten minutes. The decisions it's shaped have saved me more than I can easily calculate.
Bottom line
This prompt is good, but it's not magic. It can't know what you haven't told it. The quality of the output scales directly with the specificity of what you put in. "I'm about to take a new job" produces generic output. "I'm about to take a senior role at a Series B startup where I'd be the first person in this function, reporting to a founder who has no previous experience managing a team at this level" produces something genuinely useful.
And, as I've mentioned, it's also not making decisions for you. What it's doing is surfacing the risks you've already half-registered but haven't made yourself confront. It's a mirror, not an oracle. The call is still yours.
Let me know what you think about this prompt and how you plan to use it in the comments below.
Follow Tom's Guide on Google News and add us as a preferred source to get our up-to-date news, analysis, and reviews in your feeds.
