I’ve been following Humane since its inception. This company, founded by ex-Apple veterans, intrigued me, but I must admit, I harbored skepticism from the start. And with every update they released, my doubts only intensified. However, I love technology and I always want to try out new things. As a regular ChatGPT user, I feel squarely within their target demographic.
So, when I got back home from work yesterday and saw that the announcement video has been dropped, I immediately started watching (the 10 minute length helped, of course). Despite my growing skepticism, which only deepened after a second viewing, I noted several key points.
1. Form factor and weight
The device’s weight is noticeable, stretching even a thick sweater in the demo. I wonder about its effect on a regular T-shirt. There is an accessory for delicate fabric, but it would automatically compromise battery life. It seems more suited for short-term use rather than all-day wear.
2. Voice recognition
This being a 10 minute introduction video, one would hope for a packed and checked video, but in the very beginning this quote stood out for me.
“There are no wake words. So it’s not always listening or always recording. In fact it doesn’t do anything until you engage with it and your engagement comes through your voice, touch, gesture or the Laser Ink display.”
If it doesn’t listen to you, how would engaging with your voice help it do something? It seems like a poorly worded phrase, but they’ve had a lot of time to prepare.
3. Humane network, connected by T-Mobile
It seems like they’re launching as an MVNO on top of T-Mobile infrastructure, which explains the initial US-only release. But scaling up internationally seems challenging, especially considering data usage for photos and videos. How will roaming work? Will you be charged extra or it won’t work outside the US?
With this model, they’ll need to go country by country and make deals with the operators? Good luck getting it anywhere other than US and a couple of major countries.
And last and the most important issue on this topic – it has a dedicated phone number. The different one from your phone. So in the year 2024 they want you to ditch your iPhone and use only the ai pin. This doesn’t sound promising.
4. Ultra-Wide camera
In the video you can see how the pin shakes on the jacket. The image stabilisation must be insane to take useful photos most of the time or you need to hold the device still when you take a photo.
Although, later they show that camera is just bad. I know the best camera is the one that is with you, but god this is a bad photo and you are better off with iPhone (duh).
5. Trust Light
An interesting feature, but it raises privacy concerns for those around the user. Unlike the indicator lights on MacBooks or iPhones, which alert the user, this light offers little control to bystanders.
6. Laser Ink Display
Its real-life performance remains to be seen, but in the video the laser display doesn’t look very legible. Especially on the homescreen – if I really look, I can get what is written under the time, but I can’t make out the word above and that’s while on pause, how would it work in real life, on the go? What about really bright light?
Also gestures looked really interesting. He was moving his hand veeeery slowly. Again, there is hope that it’s to make a point, so we understand, but it doesn’t look promising or comfortable.
7. AI performance
X users already discussed some mistakes, but still. There are just a couple of questions and AI got wrong some of them. In the demo. Demo that is showing the device for the very first time. Pre-recorded demo.
First one is the eclipse question. The answer the ai pin gave is plain wrong. Not in the details, totally wrong. The mentioned eclipse won’t be visible anywhere near Australia. It will happen mostly above America, that’s why some are calling it the Great North American Eclipse.
AI lies to you and very confidently at that. Google Search really deteriorated recently, but you’ll get a good result if you ask it the same question and if there are multiple conflicting answers – you’ll see all of them on the screen. Here you just receive the wrong answer and no additional context. They also didn’t care to check the answer for the first introduction of the product. Where is the attention to details?
Apart from the cringy prompt “make it more exciting”, the message is just grammatically incorrect.
The presenter asks for the gate code he received from Andrew before, which looked cool and useful, but then added: “Before I’d have to scroll through a long text chain to the last time I went to Andrews house”. I’m sorry, I know you are excited about the product, but did you hear about the new and amazing technology called SEARCH?!
Wrong information about the eclipse probably won’t kill anyone, but the later example is a bit frightening. Allegedly it can help you with your diet, but it got it wrong, again.
The weirdest thing is that GPT-4 provides correct information if you just show the screenshot from the video (the same is true for the eclipse question):
After the eclipse and almonds example, how can I be sure it bought the correct book? “This”, it didn’t even say the name of the book to confirm. The presenter then said “great, buy it” and didn’t receive the answer. The most logical explanation – it can’t do that yet, so why show it? If it can – where did it buy it from? There were no book stores on the partnerships slide, as far as I know.
8. Web portal
One of the issues could be a bit of a nitpick. When he asked for the total protein amount he had today, the pin answered – 22 grams. But below we can see that the protein is full for the day and the amount is 56 grams. And a second slide shows that the information is taken from the presentation (there is a book, a call and a song from the examples before).
Why does it show the same information 2 times? Just to fill the screen? Or Yanirs call and note about Notes are so important?
There is a lot of potential in AI. As I said in the beginning I use ChatGPT daily and I love it. It helps a lot throughout the day. But nor we, nor ChatGPT is ready for such a device. The mission of Humane is “To build for the world not as it exists today, but as it could be tomorrow” and it could be true, but that tomorrow is still long way ahead.
I’m sure they’ll devices to the early adopters and enthusiasts, but I think it won’t make a dent. This is not it. This is not the next smartphone.
I would probably love it as a part of Apple ecosystem (or Android, if that’s your cup of tea). You have all the information there – notes, reminders, camera, photos, mail, texts, calls. Everything. And here is a new company asking you to ditch all of that and use their pin. Impossible.
