This story is a bit long, but sometimes the best journeys are.

The Breakup That Built a Blueprint

How do people get started with something new like The Goods? For me, it was a moment of clarity. It stemmed from a moment of silent competition with a business that didn’t even know they were included.

Two years ago, a "New York-style-ish" bagel shop opened nearby. Their bagels were incredible: fluffy with the best crusty crust. Addictive. Eventually, I realized I was essentially funding their business weekly. I decided I wanted to master the architecture of a real bagel at home so I could walk away for good.... or for a while at least.

The Benchmark (and the Rivalry)

I was determined I could do this on my own. My journey started by pulling two of the most reputable sources on the web: Sally’s Baking Addiction and Sophisticated Gourmet. I used ChatGPT to compare their bagel recipes with the objective of using one to make the fluffiest bagels. It recommended Sally’s as the superior base but suggested a "fluffier twist" by extending the rise time and cautioning against over-kneading. After discussing at length which ingredients would best hit my target, I asked for the instructions to be completely modified for a bread machine to handle the heavy lifting.

The 12-Round Audit: Trial, Error, and Flour Everywhere

Success wasn't linear. I went through (estimated but close to) 12 rounds of testing. By round 9, I was exhausted. I almost abandoned the bagel dream entirely. Stickiness was a constant battle, and it was a delicate balance using my knowledge and experience vs the AI’s expertise.

In troubleshooting stickiness, you have two avenues: dry ingredients or wet. For many rounds, I increased the dry. But at one point, I decided to reduce the water instead. It was a conundrum: reducing water lowers the hydration ratio, which usually means a less fluffy bagel. But I realized that by getting the hydration right in the machine, I didn't have to add extra flour during the shaping process. In the end, it actually created a better fluff.

Human in the Loop: When I Had to Course Correct ChatGPT

As a Google Workspace and Gemini expert, I know these models are powerful, but they aren't perfect. I had to step in repeatedly as the "Human in the Loop" to keep the project from being ruined:

The Hallucinated Measurements: ChatGPT suddenly pulled new measurement values out of thin air mid-conversation. I had to course-correct it to get my recipe back on track before I wasted another batch of flour.
The Bread Machine Myth: ChatGPT claimed the "Pizza Dough" setting had no kneading on the Breville bread machine. Having used my machine, I knew that was flat-out wrong. I had to call it out and force it to update the instructions.
The Temperature Debate: ChatGPT suggested a bake temp of 410°F instead of the standard 425°F. Instead of blindly following, I asked "Why?" This allowed me to understand the logic and make an informed decision for my specific oven.
Flour Auditing: I had both 00 flour and Bread flour in the house. I drove the conversation to determine the difference between "chew" and "softness," eventually deciding that a 00 and All-Purpose blend was the winning move for that specific shop-style texture.

The Gatekept Secrets (and One Happy Accident)

Tips people don't say but might show you, hidden in the details of the story, and uncovered because of a detailed eye in review.

The Brown Sugar Accident: I mistakenly put the brown sugar in the dough before the water. I didn't realize it until it was too late, but the resulting depth of flavor was so good that I made it a permanent part of the recipe.
The Orientation Secret: I watched a video of pros at work and noticed something they never said out loud: bagel orientation. To get the tallest rise, the side of the bagel that faces "up" during the proofing must be the side facing "up" on the cookie sheet.
The Steam Injection: To get that professional crunch, the bagel shop uses commercial steam injection. That not being an option for me, I remembered a tip from a sourdough book: place a hot pan in the bottom of the oven and pour water over it the moment the sourdough goes in. I figured why not try this for the bagels for the same objective.

The Lightbulb Moment: Bridging Food and GenAI

When I finally pulled that perfect bagel out of the oven, I realized I had developed my first from-scratch yeast recipe. I’ve adapted recipes before, but the tweaking here was so extensive that I can confidently say this recipe is mine.

But more importantly, I realized that most people don't know how to talk to these models to get what they want. By bridging my love of food with my experience as a GenAI trainer, I want to inspire you to use these tools to make your life easier. By showing you my real prompts and my "Human in the Loop" push-backs, you can learn the skills I use every day.

Welcome to The Goods—the product of a bagel experience, fueled by a shop that doesn't know me, and dedicated to unlocking GenAI for all of you.

Beyond the Bagel: Let's Build Something New

I didn’t just walk away from my local shop to save a few dollars; I did it so we could both learn how to build skills for the future. This is an opportunity for us to lean into trying new things, staying curious, and choosing the path of learning over the easy path of convenience.

Who cares if every recipe doesn’t work out perfectly on the first try? Who cares if I love a result that isn't for you? The point is to put yourself out there and try something because you actually care. When you lead with curiosity and hold the tools accountable, you might be surprised by the path awaiting you.

RecipeMedium