Meta launched its new personal AI tool, Muse Spark, on April 8 following an aggressive talent acquisition strategy aimed at advancing its artificial intelligence capabilities. Positioned as a versatile assistant for health tracking and travel planning, the model was developed with input from a team of physicians to address common health inquiries. To evaluate its practical utility, a user tested the system by requesting a nutritional analysis of their lunch and dinner recipe suggestions based on available ingredients. For the lunch evaluation, the user uploaded a photo of a salmon bento box containing seared salmon, rice, egg, mixed greens, raw salmon, and fish roe. Muse Spark successfully identified the ingredients and provided a detailed caloric breakdown, estimating the total meal at 760 calories. The AI noted that the sauces and dressings were likely high in sodium, suggesting the user had limited room for additional sodium intake that day. It also highlighted that the meal was rich in micronutrients like Omega-3s but deficient in fiber, vitamin C, and calcium. Based on this analysis, the AI assigned the meal a score of 7.5 out of 10. However, the system included a disclaimer stating it is not a licensed nutritionist. A specific weakness emerged when the AI attempted to generate a labeled image of the food, failing to produce readable or accurate text labels despite multiple attempts. The second phase of the test involved a dinner challenge. The user provided a photo of remaining ingredients in their refrigerator and requested recipes that were easy to prepare and clean, noting the availability of various unshown condiments. Taking into account the nutritional gaps identified during lunch, Muse Spark suggested recipes designed to increase fiber and vitamin C intake while moderately managing carbohydrate consumption. The AI offered practical advice, such as rinsing canned tomatoes to reduce sodium content and avoiding soy sauce to stay within recommended daily limits. While the AI demonstrated strong reasoning in recipe generation, it made a notable error in identifying a package of freeze-dried strawberries, failing to recognize that they were coated in sugary yogurt, which the user deemed unsuitable for a smoothie. Despite this minor misstep, the suggestions provided sufficient inspiration and taught the user creative ways to utilize leftover items, such as a half papaya. Ultimately, the user chose to prepare a Japanese-style oyakodon dish using one of the AI's recommendations. The experience highlighted Muse Spark's potential as a helpful culinary and health assistant, though it still requires human oversight to avoid misinterpretations of ingredient details. The user concluded that while they appreciated the nutritional insights, they were unlikely to completely alter their habitual use of condiments like soy sauce.