Here's the scene: I'm wearing Meta's latest Ray-Ban glasses, which have a new feature called Live AI that can answer questions about the world around you. I'm preparing for a four-hour road trip to my in-laws' house over Christmas break. I'm pre-planning my breakfast for the next day because I'm 99.9% sure I won't have the brain cells to make an edible breakfast at 5am. I don't even know if I have anything to make a meal of and. I opened the refrigerator door and said, "Hey Meta, start Live AI." Suddenly, John Cena's voice rang in my ears, telling me that the Live AI session had begun.

"What breakfast can I make with the ingredients in the refrigerator?" I asked.

It was a sad sight inside, containing leftovers from a month-old Thanksgiving, a carton of eggs, soda, condiments, a tub of Greek yogurt and a large jug of maple syrup. Meta-AI-as-John-Cena replied that I could make "a variety of breakfast dishes," such as "scrambled eggs, omelets, or yogurt parfaits."

To be clear, there is no fresh fruit that can be used to make parfaits. There are two eggs in the egg carton. My spouse put an empty milk carton back in the refrigerator, which meant the scrambled eggs and omelets were gone too. My stomach growled, reminding me that I hadn't eaten lunch. I gave up on the breakfast idea and instead opened the refrigerator door and asked what kind of dinner I could make with the ingredients inside. Mainly a bunch of frozen pizza, assorted frozen veggies, and burgers. I was told, “Frozen meals, stir-fries and casseroles.”

I decided to order dinner. This will be a drive-in breakfast on the road.

This is the problem with Live AI. Many times, I don't know when to use it. When I do this, the answers I get are too obvious to be helpful.

The selling point of Live AI is that it lets you talk to an AI assistant like you would a friend. While it functions similarly to the glasses' multi-modal AI capabilities, you don't have to constantly prompt the AI. It (supposedly) knows when you're talking to it. You can also string multiple queries and follow-up questions together. If you notice something in a cooking class that looks a little off, you alert the instructor and they take a look at the mess in your pan and tell you what you did wrong and how to fix it. It's kind of like a version of that, but with an invisible artificial intelligence in your glasses. It sees what you see and can help you in real time.

This is a cool concept. But when it came time to use Live AI without guardrails, I was stumped. Whenever a question pops into my head, I automatically reach for my phone. This is what I have been trained to do for over a decade. The first and biggest obstacle to using Live AI is remembering that it's an option.

The second question is knowing when Live AI will appear more More useful than a quick Google search. Mehta suggested that I try scenes involving fashion and cooking. I already told you how my cooking problems go. So, I asked AI what color combination I should try with a set of colorful pastel press-in nails.

The AI ​​suggests that "pastel color combinations" will "complement pink nails well." I asked which books on the shelf I should read. The AI ​​reminded me that it had "no personal preferences or opinions," but that I should "read a book that interests (me) or a book that (I) have always wanted to read." Unsatisfied, I asked which book was the best Well received. It suggested I look it up online. I tried more scenarios and wondered: Why should I talk to an AI if all it does is reiterate the obvious and tell me to Google it myself?

My most useful experience with Live AI was when I asked it how to set up my home office. At first, I got another, less friendly answer - add art, plants, and rearrange the furniture to create a cozier atmosphere. I was very angry and asked it what type The artwork looks good. It told me again that "all kinds of artwork" might look good, "depending on (my) personal style." Have I considered adding posters, prints, or paintings that reflect my interests or hobbies? I wanted to scream, but instead I asked what style of poster would look good based on Now in the room. To which I got my first somewhat helpful answer: a colorful, playful poster with a fun design or cute character that would complement the stuffed animals in the room. I asked the artists to study it. It recommended Lisa Congdon, Camille Rose Garcia and Jen Corace for their "playful and whimsical style."

This is the biggest problem I always encounter with artificial intelligence: you have to know how to ask the right questions to get the answers you want.

I can avoid some grief if I tell Meta AI: "I want to hang art in my room. Given the current situation, which artists should I focus on?" This skill comes naturally to some people. My spouse is a master at promoting artificial intelligence. But for the rest of us, it's a skill that must be learned—and those of us new to artificial intelligence are rarely taught how to rewire our brains to make the best use of the technology.

After searching on Google for artists suggested by Meta AI, I went back to square one. I like their art but neither of them mine style. I told my best friend about this experience and she rolled her eyes and immediately Instagrammed me three artists. I like them all. She said in a reproachful voice that I should have asked her directly instead of bothering the robot. She said because unlike Meta AI, she actually understands me.

In addition to philosophical issues, there are other problems with real-time AI. It's hard to differentiate when you're talking to it and talking to other people in the room. At one point, he lied outright and said he had witnessed me feeding the cat, even though I was not feeding the cat. (My spouse said they fed the kittens, which confused me.) It also only works for 30 minutes before the battery dies. This means you have to be intentional about using it - a hard thing to do when there are few obvious use cases.

I have nothing against Live AI. Our overall vision is for us all to be like Tony Stark, wearing cool glasses with our own little Jarvis inside. The future feels inevitable when you hand-hold a controlled presentation and magic. It's only when you explore alone that the fantasy begins to crumble. When this happens, nine times out of ten, you'll be reaching for your phone.