Super Savers, I've been leaning on AI lately to help me create recipes using only what I already have in the house. A couple of weeks ago, I barely had enough money to buy a $0.98 bag of frozen vegetables, but my pantry still had a mix of frozen, canned, and shelf-stable goods. A few items were clearance finds I had grabbed without any real plan, so when it came time to turn them into meals, I turned to AI for backup.
My latest experiment was with Gemini, which gave me a recipe for butternut squash soup. The broth amount seemed suspiciously high, so I checked with someone more experienced to confirm my instincts. After that, I asked Gemini to reduce the broth, but even then, the soup turned out thinner than I like. I asked again for help thickening it, but the suggestions were a bit too involved for the exhausted cook I was that day. Thankfully, I’ve spent enough time in the kitchen to find my own simple workaround. The final result? Adequate. Not bad, not memorable—just fine.
Next up was Copilot, which gave me several ideas for making pigs in blankets. The version I chose called for way too much almond flour for just two hot dogs. I didn’t bother asking for a corrected recipe, because, truth be told, I don’t find Copilot quite as sharp as the other models. The end result wasn’t impressive, but it was edible—sometimes that’s all we really need.
Finally, I turned to ChatGPT for a sweet potato bread recipe. The first version it gave me used a seemingly large amount of almond flour. After a pretty convincing explanation, I decided to trust the process. The bread came out incredibly moist, which was a win, but it lacked flavor and the outside burned before the inside finished baking.
None of these creations would make Gordon Ramsay proud, but they did help me make the most of what I had without spending hours scrolling through random recipes online. In the end, each one got me through another meal, and sometimes that’s the real victory.
Comments
Post a Comment