This is Optimizera weekly newsletter sent every Friday from a senior Verge reviewer Victoria Song that analyzes and discusses the latest phones, smartwatches, apps and other gadgets that promise to change your life. Optimizer arrives in our subscribers’ inboxes at 10 a.m. ET. Report yourself Optimizer Here. We leave for another two weeks and return on November 7.
As I wrote last week, I’m quickly running out of body parts to do my job. Part of being human is knowing when to ask for assist, so a few months ago I asked Senior Editor Sean Hollister – another intelligent glasses geek – to assist me test Halo Glass, an always-listening AI companion that lives inside the glasses.
Halo is the brainchild of two former Harvard students who made headlines last year after donning a pair of Ray-Ban Metas to attack strangers in real time. In August, AnhPhu Nguyen and Caine Ardayfio announced they created a pair of always-on AI glasses that could listen, record, transcribe, and then organically provide you with answers to questions relevant to your conversations in real time. It’s sort of a combination of Cluely, another AI startup that aims to assist you “cheat everything,” and Bee, an AI wearable that claims to act as your second memory. Instead of a pin or wristband, you can discreetly view answers with intelligent glasses.
Of course I wanted to test them.
Sean and I spoke with Ardayfio, who told us that while Halo will eventually build its own hardware, for now we will be among the first to experience their app running on the platform Even Realities G1 glasses. You may not have heard of Even Realities, but it was one of the more impressive intelligent glasses makers at CES. All we would have to do is try out the prototype, compare notes, and then report our experiences. Simple, right?
The attraction for both of us was having a second memory. We are busy people, sometimes forgetful. Wouldn’t life – and our work – be a little easier if we stopped forgetting that one thing we said to our co-workers, bosses and spouses? Wouldn’t it be easier to interview sources if, when an esoteric term was used, the definition could appear in real time without having to interrupt the flow of conversation?
It certainly sounds good, but always-on AI wearables pose a host of ethical conundrums. Since this is a completely recent product category, the ethics of it all took us a bit by surprise. To start with, Sean lives in California, a state that legally requires both parties to consent to recording a conversation. Is he committing a crime if he wears these glasses and doesn’t reveal to everyone around him that he’s recording? And Sean’s wife has a job that requires confidentiality. The always-on recording device could threaten her livelihood if Sean forgets to turn it off while she is working and he is nearby. As a result, Sean I couldn’t actually test these glasses at home. Meanwhile, my spouse is fed up with AI wearables that are always listening after I reviewed Bee and recorded one of our fights. (To test Friend, I had to wear it outside the house). Our solution was to each put on a pair of Halo-enabled G1 glasses and hop on a video call to test it out with each other.
Theoretically, Halo works like this: in the app, you see a live transcription of the conversation going on around you. Every now and then a factoid pops up about something it refers to. For example, perhaps you are talking about animals found in Australia and someone asks which one is the most hazardous. This reply will be sent to your glasses and you can look like a smartass while you talk. Once the call is over, you’ll see a low summary of the call and some actionable items, similar to what you do at the end of a meeting.
In practice, our call was ridiculous.
It started with a 20-minute troubleshooting session, including multiple firmware updates and disconnections. I’ll spare you the details, other than that, because it’s just… the most awkward way to interact with AI imaginable: the G1 glasses require you to look up to bring up the display. You can adjust the required angle – a wise choice as the default value is 40 degrees. It’s kind of like just tilting your head back and looking at the ceiling. We’ve both adjusted to temperatures around 15 degrees, but it’s still a comically obvious trigger.
Wonky’s prototype hardware is forgivable because you’re exploring the idea. And the thought that AI glasses can make you look smarter without knowing the person you’re talking to makes me uncomfortable.
I talked to Sean about my concerns. We wondered if intelligent glasses really assist people stay present in the moment. We wondered, can you really be yourself when you know you’re being recorded? What level of disclosure is ethical? How can you protect the privacy of your loved ones who may not be as interested in this technology as you are?
It was a riveting conversation, apart from the AI interruptions. At this point, one of us would have to tilt our heads back to see what notification had appeared. Imagine Sean and I, after 30 minutes of talking, throwing our heads back like demented sea lions barking at a pier.
Sometimes he would throw in useless tidbits. For example, he showed me the definition of the word “domesticated” when I used it correctly. I felt slightly offended that the AI perhaps thought I didn’t know the meaning of the word in context. When I mentioned Cluely, the Halo AI gave facts about it instead Unconscious“1995 coming-of-age comedy directed by Amy Heckerling.” Typical artificial intelligence.
The worst part was when Halo showed a message explaining that cell phones first appeared in the 1970s and 1980s. Sean had to say something about the phones for me to take it. I relayed this fact to Sean. He then told me that his glasses showed the same notification. Artificial intelligence alerted me again that the first phones appeared in the 1970s and 1980s. We are stuck in an AI-powered ouroboros hell. We nodded even more.
Several times Halo AI provided helpful facts. The definition of “nits” came up when we talked about intelligent glasses displays. This defined “doomerism” as Sean and I spiraled, wondering about the impact of constant recording on the lives of those around us.
But ultimately, using Halo was more of a distraction than a assist. All the while, about 10 percent of your brainpower was spent wondering when your assistant would cut in and when you’d hang up. Re-reading the transcript of our conversation on the app, there were so many dropped threads that I wish we had delved into them without all the distractions.
Sean told me that his interest in Halo stemmed from a very human desire to “remember better.” I bet anyone with a to-do list would do this. I had the same feeling when testing Bee’s wearable AI. Yet, This the conversation — where the AI spit out the same facts to each of us at the same time — simply reminded me of Microsoft’s Clippy. Always there, pestering you with tidbits that weren’t very useful and interrupting your train of thought as soon as you started.
For now, I think I’ll take my imperfect mix of analog sticky notes and to-do lists. I settle for potentially looking stupid in conversation by asking, “Excuse me, what does that mean?” It’s not thrilling, but I’d rather not nod my head the next time I need an answer.